Acta Optica Sinica
Co-Editors-in-Chief
Qihuang Gong
2025
Volume: 45 Issue 16
38 Article(s)
Jie Fu, Hu Wang, Pengxiang Wang, Hao Dong, Xingkun Liu, Shanming Li, Wenqing Zhang, Zijun Gao, Yin Hang, and Yuchuan Shao

ObjectiveX-rays have extensive applications across multiple fields, including medical imaging, security screening, container inspection, defect detection, pollution detection, and quality control in the material industry. Halide perovskite single crystals have emerged as promising materials due to their high X-ray absorption coefficient, substantial carrier mobility-lifetime product, tunable bandgap, high resistivity, and cost-effective synthesis methods. These perovskite single-crystal materials demonstrate successful implementation in X-ray detectors and show potential as next-generation ionizing radiation semiconductor materials. Cesium lead bromide (CsPbBr3) single crystals present a viable option for large-scale commercial X-ray detection applications. However, ion migration negatively impacts the long-term stability and optoelectronic properties of three-dimensional perovskite CsPbBr3 single crystal X-ray detectors. PEABr, a typical component of long-chain organic cations in two-dimensional perovskite, forms two-dimensional perovskite benzylamine lead bromide (PEA2PbBr4). The phenylethylammonium ion (PEA+) organic layer between the octahedral layers of lead bromide exhibits high resistivity and hydrophobicity, resulting in enhanced stability and high ion activation energy. Consequently, low-dimensional perovskite materials demonstrate superior ion migration inhibition compared to three-dimensional perovskite. Additionally, bromo perovskite materials grown in halogen-rich environments, such as hydrobromid (HBr), exhibit improved crystal quality. This study investigated the growth of two-dimensional perovskite microcrystals on CsPbBr3 single crystal surfaces using PEA2PbBr4 solutions with varying HBr concentrations. The research examined surface morphology changes under different passivation conditions and analyzed the performance variations of CsPbBr3 single crystal X-ray detectors before and after passivation. The study compared detector performance for different treatment methods and demonstrated the effectiveness of passivation in enhancing X-ray detection capabilities, potentially advancing the commercialization of perovskite single crystal X-ray detectors.MethodsCsPbBr3 single crystal ingots were prepared using the Bridgman method, followed by cutting and polishing to obtain CsPbBr3 single crystals. The crystals underwent treatment in HBr solutions containing varying concentrations of two-dimensional perovskite PEABr can form two-dimensional perovskite benzylamine lead bromide (PEA2PbBr4) for microcrystal growth. Following passivation, titanium (Ti) metal electrodes were deposited on the top and bottom surfaces to fabricate X-ray detectors. Characterization methods included X-ray diffraction (XRD) for phase analysis, photoluminescence (PL) measurements using a 405 nm wavelength laser, and scanning electron microscope (SEM) analysis for surface morphology examination. X-ray irradiation utilized an Amptek Mini X2 X-ray tube (Ag target) at 30 kV voltage, with dose control through tube current adjustment. Voltage-current measurements were conducted using a Keithley 2634B source meter within a 3 mm thick lead-shielded box. The experimental setup incorporated a programmable motion platform positioning an aluminum (Al) plate between the X-ray tube and detector, enabling two-dimensional image generation through photocurrent variations corresponding to X-ray dose changes through aluminum gaps.Results and DiscussionsXRD analysis and PL spectra confirm the formation of two-dimensional perovskite PEA2PbBr4 microcrystals on the CsPbBr3 single crystal surface (Fig. 2). The CsPbBr3 single crystal surface, passivated with a 0.01 mol/L PEA2PbBr4 acid solution, exhibits minimal surface scratches from cutting and polishing, without large-scale two-dimensional microcrystal formation due to excess concentration (Fig. 4). Following deposition and device fabrication, the CsPbBr3 single crystal detector maintains a low dark current of approximately 10 nA and a hysteresis width below 1 nA under a 200 V bias voltage (Fig. 6). Under varying X-ray dose exposures, the CsPbBr3 single crystal detector maintains stable dark current and demonstrates superior resolution at low X-ray dose rates (Fig. 7). Data analysis reveals that the CsPbBr3 single crystal detector exhibits high sensitivity and low X-ray detection threshold (Fig. 8). Comparative analysis with CsPbBr3 single crystal detectors treated through alternative methods indicates that the acid solution-based two-dimensional perovskite passivation method substantially improves X-ray detection performance (Table 1). In imaging mode, the CsPbBr3 single crystal detector produces distinct mask pattern images at a low dose rate of 200 nGy·s-1 (Fig. 10).ConclusionsThis research successfully demonstrates the passivation of CsPbBr3 single crystals with microcrystals using PEA2PbBr4 in HBr solution, effectively suppressing ion migration within the single crystal detectors. The findings indicate that CsPbBr3 single crystal detectors passivated at 0.01 mol/L concentration demonstrate optimal passivation performance. The detector achieves superior sensitivity of 9.55×104 μC·Gy-1·cm-2, and a minimal dose level of 13.8 nGy·s-1. Additionally, the detector enables X-ray imaging with 1 mm2 spatial resolution at a low X-ray dose rate of 200 nGy·s-1. These results establish that growing two-dimensional perovskite microcrystals on three-dimensional perovskite single crystals through passivation significantly enhances optoelectronic detector performance, offering valuable passivation strategies and design guidance for three-dimensional perovskite detectors.

Aug. 18, 2025
  • Vol. 45 Issue 16 1604001 (2025)
  • Shiyao Liu, Kai Liu, Yongqing Huang, Xiaofeng Duan, Qi Wang, and Shiwei Cai

    ObjectiveThe rapid advancement of large-scale artificial intelligence (AI) models has driven exponential growth in the demand for computing power and data capacity. To date, China has established the world’s largest and most advanced fiber-optic and mobile communication networks. The data center industry in China has maintained an average annual growth rate of 30% over the past five years. As the backbone of global telecommunications infrastructure, optical communication systems now face unprecedented challenges in handling massive data transmission. Transmission rates have advanced from 45 Mbit/s in the first generation to 400 Gbit/s in the fifth generation, with a target capacity of 10 Tbit/s. To meet the evolving demands of high-speed optical communication systems, there is an urgent need to develop next-generation high-speed photodetectors that meet new technical standards. These devices must achieve a response bandwidth exceeding 57 GHz in single-wavelength 100 Gbit/s optical channels while maintaining a responsivity above 0.5 A/W. Based on this, we propose an avalanche uni-traveling carrier photodetector with an electric field regulation layer (ER-AUTC-PD). This innovative structure significantly enhances photodetector responsivity while preserving high-speed response characteristics.MethodsIn this paper, we focus on the device structure design and performance optimization of the proposed ER-AUTC-PD, employing the Silvaco ATLAS semiconductor device simulation tool (TCAD) for modeling. An intrinsic InP multiplication layer is introduced between the cliff layer and the absorption layer to achieve photocurrent avalanche gain via impact ionization. In addition, a p-InP electric field regulation layer is inserted between the collection layer and the cliff layer. By co-optimizing the doping concentrations of both the cliff layer and the electric field regulation layer, the high-electric-field region is localized in the InP multiplication layer to enhance avalanche gain, while the electric field within the collection layer is tuned to approach the field intensity corresponding to the electron peak drift velocity, thus improving high-speed response characteristics. Subsequently, the thickness of the collection layer is optimized. The increased electron transport speed partially compensates for the delay caused by the longer transit distance in the thicker collection layer, helping to reduce parasitic capacitance and further enhancing high-speed performance. Moreover, the interrelationships and trade-offs among photocurrent gain, bias voltage, electron transport properties, and parasitic capacitance are comprehensively analyzed to achieve balanced performance. The final ER-AUTC-PD structure is thus designed.Results and DiscussionsThe proposed ER-AUTC-PD demonstrates improved gain-bandwidth product performance compared to traditional APDs, achieving a gain-bandwidth product of up to 360 GHz [Fig. 6(a)]. The device also exhibits excellent responsivity and high-speed optical response. Under a bias voltage of -10.1 V, a 14 μm-diameter device achieves a 3 dB bandwidth of 63 GHz, a responsivity of 0.501 A/W, and a multiplication gain (M) of 2.34. Under a -10.4 V bias, a 10 μm-diameter device reaches a 3 dB bandwidth of 75 GHz with a responsivity of 0.505 A/W and M=2.44 [Fig. 6(b)]. With an incident optical power of 15.4 μW and a responsivity of 0.5 A/W, the optimized ER-AUTC-PD achieves a 3 dB bandwidth of 63 GHz, significantly outperforming the 43 GHz bandwidth of a pre-optimized MUTC-PD, clearly demonstrating the benefits of structural optimization [Fig. 7(b)]. Furthermore, comparison with several reported APDs (Table 2) shows that the ER-AUTC-PD offers superior performance in terms of low power consumption, high responsivity, and high bandwidth.ConclusionsIn this paper, we propose an avalanche uni-traveling carrier photodetector with an electric field regulation layer. By introducing a multiplication layer between the absorption layer and the cliff layer, photo-generated electrons undergo impact ionization under a strong electric field to achieve avalanche gain. A p-InP electric field regulation layer is also added between the collection layer and the cliff layer. Through coordinated optimization of the doping concentrations in the cliff and electric field regulation layers, the electric field is precisely distributed to localize the high-field-intensity region within the multiplication layer and maintain a field strength corresponding to the electron peak drift velocity in the collection layer. This configuration enables both enhanced responsivity and improved 3 dB bandwidth. Furthermore, the trade-offs among photocurrent gain, bias voltage, electron transport behavior, and parasitic capacitance are analyzed to reach an optimal balance. The optimized 14 μm-diameter device achieves a 3 dB bandwidth of 63 GHz under a -10.1 V bias, with a responsivity of 0.501 A/W and an M of 2.34. The 10 μm-diameter version reaches a 3 dB bandwidth of 75 GHz at -10.4 V, with 0.505 A/W responsivity and an M of 2.44. The ER-AUTC-PD shows clear performance advantages in two key areas: first, it achieves a higher gain-bandwidth product than traditional APDs; second, it significantly reduces power consumption by operating at a lower avalanche reverse bias voltage. These results demonstrate that the ER-AUTC-PD provides excellent performance in low power consumption, high responsivity, and high bandwidth, meeting the requirements for single-wavelength 100 Gbit/s optical communication systems.

    Aug. 07, 2025
  • Vol. 45 Issue 16 1604002 (2025)
  • Ning Shen, Darui Zhang, and Changhe Zhou

    ObjectiveIt is well established that light beams gradually diverge during propagation due to diffraction effects, imposing limitations on their practical applications. Extensive research has been conducted to develop methods for generating non-diffraction beams. While significant progress has been achieved, substantial challenges remain. For example, propagation distance and efficiency are highly sensitive to beam parameters. Precise control and stable generation of high-quality non-diffraction beams in experimental settings thus remain a focal point of current research. Recent advancements in micro-nano optical devices have enabled optical field manipulation through structural parameter tuning. For instance, metasurface-based approaches have been used to engineer non-diffraction properties; however, these complex metasurface structures often pose significant fabrication challenges. In this paper, we study the diffraction field generated by the picometer misplacement combined grating (PMCG) with the property of non-diffraction. The PMCG is prepared by the double holographic exposure technique. By adjusting the angle between the two beams and the parameters of the exposure period, the period difference between the two exposures is set at a picometer scale. This new method of generating a non-diffraction field by adjusting the holographic exposure period is expected to be applied in the field of three-dimensional measurement and precision measurement.MethodsThe PMCG is a diffraction optical element generated by two exposures of a holographic interference field with a controllable period at the picometer scale. In the holographic exposure field, an angle θ exists between two interfering beams. By adjusting θ to θ+Δθ, the period of the exposure field changes from d to d+Δd, resulting in a period difference (Δd) between the two exposures. A scanning reference grating (SRG) system is employed for precise measurement of the interference field, enabling long-travel online measurement of the lithographic interference field. By measuring the period of the interference light field, the difference between the two exposure periods is accurately controlled at the picometer scale. The holographic interference field is used to expose a quartz substrate coated with photoresist. The first exposure records an interference field with a period of d. After the expected period change of the holographic interference field is determined using the SRG system, a second holographic exposure is performed, recording the superposed interference field from both exposures on the quartz substrate.Results and DiscussionsIn this study, the PMCG with dimensions of 30 mm×50 mm×10 mm is fabricated. The first holographic exposure has a period of 4096 nm, and the period difference Δd between the second and first exposures is 400 pm. The PMCG is characterized using an atomic force microscope (AFM). As shown in Fig. 7, variations in exposure intensity at different positions are observed, with adjacent grating periods exhibiting cumulative or subtractive states. These results confirm that the PMCG accurately records the superposed interference field from the two holographic exposures. Physical analysis of the diffraction field generated by the PMCG reveals that two interfering beams within the field maintain a constant field structure regardless of propagation distance, which is the mechanism underlying its non-diffraction property. A laser collimating and beam-expanding system is constructed to analyze the non-diffraction field produced by the PMCG, as shown in Fig. 9. During field propagation, the diffraction field structure remains essentially unchanged, with a divergence angle of only 0.001 rad. This experimental result verifies that the diffraction field exhibits propagation invariance, confirming the non-diffraction characteristic of the PMCG.ConclusionsIn this study, a novel non-diffraction optical element is proposed, whose diffraction field exhibits propagation invariance. The PMCG is fabricated using the double holographic exposure technique. By measuring the period of the exposure interference field with the SRG system, the periods of the two exposures are set as d and d+Δd, achieving period differences between the two exposures at the picometer scale. Through analysis of the physical principles governing the superposed interference fields from the two holographic exposures, the underlying mechanism for generating the diffraction field by the PMCG is identified. The mutual interference of two coherent beams within the diffraction field ensures that the diffraction field structure remains unchanged regardless of propagation distance. To validate the theoretical analysis, a laser collimating and beam-expanding system is constructed to characterize the propagation properties of the picometer optical comb in free space. By examining the diffracted field intensity patterns at different propagation distances, the non-diffraction characteristic of the PMCG is experimentally confirmed. This work introduces a new paradigm for non-diffraction optical elements, paving the way for potential applications in optical manipulation, optical imaging, optical measurement, and other precision-related fields.

    Aug. 07, 2025
  • Vol. 45 Issue 16 1605001 (2025)
  • Tao Zhang, Xiaolong Tao, Boyi Liu, Liu Yang, Fengguang Luo, Yuebin Li, Yongming Hu, and Yi Cai

    ObjectiveNonlinear compensation in coherent optical fiber communication represents a crucial challenge for capacity enhancement. Digital backpropagation (DBP) is extensively utilized for nonlinear mitigation owing to its robust theoretical interpretability. Previous research has enhanced the DBP algorithm through various methods, including algorithmic structure refinement and nonlinear operator optimization, each supported by theoretical foundations. However, the performance improvements achieved through nonlinear parameter optimization lack comprehensive theoretical analysis. This paper addresses this research gap by examining the error sources in nonlinear parameter optimization. Based on error analysis, the theoretical model is reformulated to implement nonlinear compensation. The results indicate that the proposed theoretical framework provides substantial validation for the observed performance enhancements.MethodsThe investigation begins with an analysis of theoretical errors in the DBP algorithm’s compensation process, examining the ratio between self-phase modulation (SPM) and self-steepening effects before and after sampling through propagation equation derivation and nonlinear operator calculations. The analysis reveals significant inherent errors in this process. The temporal scale variation during sampling causes substantial changes in the SPM-to-self-steepening ratio due to the time-dependent nature of the self-steepening effect. Based on this observation, self-steepening suppression is incorporated into the compensation framework. Through theoretical re-derivation, the nonlinear operator is reformulated, with particle swarm optimization (PSO) applied to enhance compensation accuracy.Results and DiscussionsThe investigation included reproduction of the adaptive DBP algorithm for optimized nonlinear compensation. While the adaptive DBP algorithm represents blind estimation, the extended DBP functions as its interpretable counterpart. Experimental results demonstrate that for quadrature phase shift keying (QPSK) modulation with four-step DBP compensation, the extended nonlinear operator delivers an average gain of 0.96 dB compared to fixed operators, and 1.87 dB improvement over dispersion compensation alone. For 16QAM (64-order quadrature amplitude modulation), the respective enhancements are 0.65 dB and 5.485 dB, while 64QAM exhibits gains of 0.22 dB and 2.06 dB. Notably, the adaptive nonlinear operator achieves similar performance to the extended operator across all three modulation formats. Further tests under varying step sizes and transmission distances consistently demonstrate comparable performance characteristics.ConclusionsThis study examines the theoretical mechanism for parameter optimization of the DBP algorithm in nonlinearity compensation, beginning with the nonlinear Schr?dinger equation. The analysis reveals that performance degradation in parameter optimization stems from the reduction of the nonlinear component in the theoretical derivation. To address this limitation, an extended nonlinear operator is derived from the nonlinear Schr?dinger equation, incorporating both power and amplitude terms. This extended nonlinear operator is implemented alongside the particle swarm algorithm for backend nonlinear compensation and evaluated against DBP and adaptive DBP algorithms. The theoretical mechanism established for the parameter optimization issue in the DBP algorithm aligns with previous theoretical analysis and simulation results.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1606001 (2025)
  • Bo Chen, Weile Zhai, Wuying Wang, Fangjing Shi, Yan Rong, and Yongsheng Gao

    ObjectiveFuture wireless communication, radar, and electronic warfare systems are inevitably evolving towards higher frequency bands, wider bandwidths, and larger dynamic ranges. Achieving high-quality reception of ultrawideband signals has become one of the current research hotspots. Most of the reported microwave photonic channelized receiver schemes are based on the optical frequency comb (OFC) reception principle. However, due to the limitations of OFC generation techniques, it is challenging to significantly enhance the operating frequency range and maximum instantaneous bandwidth of the receiver. Moreover, full-duplex systems can double the spectral efficiency, but they inevitably generate self-interference (SI) signals, which degrade the communication quality. Current research on microwave photonic channelized reception primarily focuses on channelization schemes based on OFCs, especially those employing dual OFCs. This requirement increases system complexity and makes it challenging to expand the number of subchannels due to the limited number of comb lines. Although Kerr OFCs generated by integrating microring resonators (MRRs) can provide a large number of comb lines, their line spacing is not tunable and is susceptible to instability caused by factors such as temperature.MethodsIn response to the challenge of generating ideal OFCs in existing channelized receiver schemes based on OFCs, a method for generating 6-line or 8-line OFCs is proposed. The generated combs exhibites low modulation index, high flatness, and flexible tuning of the comb line spacing. To address the issue of low channelization efficiency of a single comb line, a method utilizing an acousto-optic frequency shifter (AOFS) to shift the OFC is proposed, which enhances the channelization efficiency of a single comb line by a factor of three. To tackle the inevitable image interference in superheterodyne receiver architectures, a method for broadband signal image suppression in the digital domain using an in-phase/quadrature (I/Q) balance compensation algorithm is proposed. For the SI problem in full-duplex transceiver systems, a method of introducing a reference signal with the same amplitude but opposite phase as the self-interference signal to directly suppress self-interference in the all-optical domain is proposed.Results and DiscussionsWe carry out experiments to verify the proposed scheme. Fig. 4 shows the experimental results of the 6-line OFC generation, indicating good flatness and out-of-band suppression ratio of the OFC. Fig. 5 presents the experimental results of image suppression for a 300 MHz wideband signal. When the broadband signals of channel 2 and channel 5 are image signals to each other, the image suppression ratio is not less than 33 dB. The same applies to channels 7 and 10, as well as channels 15 and 18. Compared with the Hartley-based image-rejection mixer, our proposed algorithm shows remarkable advantages. Fig. 6 shows the self-interference cancellation experimental results. When the signal bandwidth is 100 MHz, the self-interference depth can exceed 32 dB. It should be noted that due to experimental limitations, it is not feasible to directly generate a broadband signal with a bandwidth of 36 GHz. Therefore, in our experiments, we employ a series of 16-QAM signals with continuous spectra and a bandwidth of 300 MHz each. These signals are used to verify the channelized reception and image suppression capabilities within the 2?38 GHz frequency band through spectral coverage. For the verification of self-interference cancellation, a broadband signal with a bandwidth of 100 MHz is used as the self-interference signal. It is worth mentioning that both the image suppression and self-interference cancellation performance are expected to degrade as the signal bandwidth increases.ConclusionsCompared to other reported microwave photonic channelization schemes based on dual OFCs, the proposed scheme can achieve instantaneous and complete reception of any wideband signal in the 2?38 GHz range using just a single 6-line OFC. It has obvious advantages in system complexity, cost, operating bandwidth, and maximum instantaneous bandwidth and shows great potential for future applications in radar, electronic warfare, wireless communication, and other fields.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1606002 (2025)
  • Guomei Zhu, Zhaofeng Liu, and Liren Zheng

    ObjectiveThe study of solitons in nonlinear optical lattices has attracted significant theoretical and experimental interest in recent years. Nonlinear optical lattices, characterized by localized nonlinear intensities and spatially periodic modulation, demonstrate distinctive photonic properties. In particular, parity-time (PT) symmetric nonlinear optical lattices support stable discrete solitons, achievable through careful modulation of nonlinear gain and loss in specifically designed waveguides. Building upon these advances, this research investigates the integration of nonlinear lattices with linear Hermitian lattices to examine their effects on soliton optical transmission properties in nonlinear-Hermitian Moiré photonic crystal lattices under non-Hermitian conditions. Through the development of a nonlinear Schr?dinger equation for hybrid linear-nonlinear Moiré lattices and numerical simulations, this study examines bandgap soliton excitation and stability regimes. The research specifically addresses how nonlinear lattice parameter modifications, including depth, gain-loss degree, and lattice period, affect soliton stability domains, presenting findings through power-profile analysis of stable solitons. This investigation provides essential understanding for the design of photonic devices that control nonlinear soliton dynamics in non-Hermitian lattice systems.MethodsThe nonlinear Schr?dinger equation (NLSE) governs optical soliton transmission. Building upon this equation, we establish a physical model for soliton propagation in non-Hermitian Moiré photonic crystals. Through the application of the squared-operator iteration method (SOM) and its modified version (MSOM), precise soliton solutions are obtained within the bandgap structure. The stability analysis involves introducing small perturbations to the exact solutions and linearizing the NLSE, converting the stability analysis into an eigenvalue problem. The associated eigenoperator is solved using the Fourier-collocation method, where eigenmatrices are constructed from Fourier basis components to determine the linear stability spectrum of solitons. The eigenvalue spectrum analysis provides understanding of soliton transmission stability characteristics. Based on these stability findings, the split-step Fourier (SSF) method simulates soliton propagation. This method alternates between linear propagation and nonlinear interaction processing, facilitating systematic investigation of soliton transmission properties and inter-soliton interactions within the lattice. The band structure of one-dimensional non-Hermitian Moiré photonic crystals is computed using the plane-wave expansion (PWE) method. This calculation specifically addresses the linear band structure, excluding nonlinear effects. The results demonstrate a significant correlation between bandgap characteristics and stable soliton transmission thresholds.Results and DiscussionsThis investigation models the nonlinear coefficient in the nonlinear Schr?dinger equation as a periodic function, establishing a linear lattice system with identical periodicity to the nonlinear Moiré lattice. With T=3 as an example (Fig. 2), increasing nonlinear lattice depth diminishes soliton power and narrows the stability region. This phenomenon stems from the interaction between self-focusing effects and gain-loss modulation, where the imaginary component of the gain-loss coefficient predominates, compelling the system to reduce soliton power. At a lattice period of 6 (Fig. 4), bandgap solitons maintain stable propagation in the semi-infinite band, independent of lattice depth variations. However, at period of 2 (Fig. 5), the soliton-lattice coupling equilibrium destabilizes due to enhanced localized gain-loss modulation, preventing stable bandgap solitons. The analysis indicates that lattice period must correspond to the soliton coherence length scale for stable propagation through periodic nonlinear modulation. The research examines soliton stability under periodic mismatch between linear (period TL=3) and nonlinear lattices. At nonlinear lattice period of 6, the system approximates constant-coefficient behavior, facilitating stable bandgap soliton propagation with stability regions unaffected by lattice depth (Fig. 7). Increased gain-loss coefficient expands instability zones while maintaining power curves (Fig. 9). Conversely, at period of 2, intensified spatial modulation constrains stability regions (Fig. 8). Substantial gain-loss modulation (gain-loss coefficient ω1≥0.30, Fig. 10) or sub-threshold periods disrupt soliton-lattice coupling balance, destabilizing solitons. The findings underscore lattice period matching: stable solitons require nonlinear periods above a critical threshold for quasi-constant effects, while shorter periods induce destabilizing localized nonlinear modulation.ConclusionsThis research establishes a composite lattice system incorporating linear and nonlinear components, governed by the nonlinear Schr?dinger equation. Through comprehensive analysis, the study examines soliton excitation and stability characteristics across various parameters, including lattice periodicity, nonlinear lattice depth, and gain-loss coefficient. The findings demonstrate that increased nonlinear lattice depth induces exponential power attenuation in solitons and substantially reduces the stable transmission window. Conversely, gain-loss coefficient variations minimally affect soliton power and stability domain. The analysis confirms that nonlinear lattice depth primarily determines soliton properties, with regulatory efficiency exceeding that of gain-loss modulation by two orders of magnitude.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1606003 (2025)
  • Ke Dong, Dan Chen, and Mingjun Wang

    ObjectiveOptical camera communication (OCC) serves as a significant complement to existing radio frequency (RF) communication methods in the realm of outdoor vehicle-to-everything (V2X) communication due to its superior anti-interference capability, security, and relatively low deployment cost. However, the inherent limitations of image sensor imaging mechanisms and manufacturing processes make OCC technology exhibit a considerable gap in communication rate and reliability compared to traditional visible light communication (VLC) techniques. Existing research has explored higher-order modulation schemes and resource multiplexing to enhance the communication performance of OCC. Nevertheless, current higher-order data modulation schemes struggle to simultaneously achieve simplicity, flicker-free operation for human eyes, and flexible scalability. Furthermore, the exposure effect of optical cameras is a critical factor affecting the reliability of OCC, with its impact being more pronounced for higher-order modulation schemes. Consequently, this adverse effect must be effectively suppressed to ensure reliable communication. We proposed a higher-order modulation scheme with binary amplitude suitable for OCC and constructed a two-dimensional constellation space in the frequency and phase domains, enabling simple implementation of non-flickering and flexibly scalable higher-order data modulation. Additionally, a demodulation algorithm based on frequency domain equalization is presented to mitigate inter-symbol interference and frequency-selective attenuation caused by the camera exposure effect.MethodsFirst, an exposure model for a rolling shutter camera was established, which characterized the exposure effect as a low-pass filter and described its distortion of optical pulses as inter-symbol interference and frequency-selective attenuation. A joint phase-frequency shift keying modulation method was presented by leveraging the constant amplitude envelope and evenly distributed characteristics of both frequency shift keying (FSK) and phase shift keying (PSK). This method constructed a two-dimensional constellation space in the frequency and phase domains to achieve flexible higher-order modulation. The waveform of each modulated symbol consists of square waves with different frequencies and phase offsets, ensuring ease of implementation. The data frame structure was designed, comprising Q data symbols, q header symbols, and q tail symbols to facilitate synchronization at the receiver and eliminate inter-symbol interference. Based on this frame structure, a corresponding data demodulation algorithm was provided with the core steps therein, including a frequency domain equalizer and a symbol demodulator. The former calculated the optimal equalization filter coefficients under the minimum mean square error (MMSE) criterion to compensate for the frequency-selective attenuation experienced by the received signal and enhance data demodulation performance. The latter estimated the constellation point corresponding to the transmitted symbol from the pixel value sequence after frequency equalization, followed by constellation de-mapping to obtain the payload bits. Finally, the influence of imaging parameters of the rolling shutter camera, such as exposure time and readout time, on the system transmission performance was briefly analyzed.Results and DiscussionsThe symbol error rate (SER) performance of the proposed joint phase-frequency shift keying modulation system was evaluated using Monte Carlo simulations. The influence of different camera exposure time parameters on the average SER of the system was compared. The results demonstrate that a longer camera exposure leads to more significant inter-symbol interference and frequency-selective attenuation, thereby considerably affecting demodulation reliability. However, the introduction of a cyclic prefix and a frequency domain equalization algorithm effectively mitigates this adverse effect. The average SER performance of the system is compared when employing constellation diagrams with different frequency and phase modulation orders, indicating that the system reliability gradually decreases as the number of constellation points increases. Moreover, this influence is more pronounced for an increase in the frequency modulation order compared to the phase modulation order. Furthermore, the average SER performance of the joint phase-frequency shift keying modulation system is compared when using image sensors with different readout time. Although using an image sensor with a shorter readout time yields better SER performance, it comes at the cost of reduced pixel utilization efficiency. Finally, the SER performance of four higher-order modulation schemes with binary amplitude is compared under the same exposure time, modulation order, and symbol rate. The simulation results show that the proposed joint phase-frequency shift keying modulation scheme outperforms the other higher-order modulation schemes.ConclusionsWe propose a joint phase-frequency shift keying modulation scheme for OCC along with a demodulation algorithm based on frequency domain equalization. This modulation scheme utilizes binary-amplitude square waves as carriers to achieve flicker-free and scalable higher-order modulation. Compared to existing higher-order modulation techniques, it offers lower implementation complexity and better illumination compatibility. At the receiver, by integrating frequency domain equalization into the demodulator structure, inter-symbol interference and frequency-selective attenuation caused by the image sensor exposure effect are effectively suppressed. Through numerical simulations, we compare the SER performance of the proposed scheme under different exposure time parameters and elucidate the negative influence of the exposure effect on data transmission reliability, while also confirming the effectiveness and necessity of employing a frequency domain equalization algorithm in the demodulator. Based on the low-pass filtering characteristic of the exposure effect, we explore and validate the significant influence of the constellation design, particularly the selection of frequency combinations, on the system's SER. By relating the readout time parameter of the rolling shutter image sensor to the equivalent sampling frequency in an oversampling modulation scheme, the trade-off between communication effectiveness and reliability under imaging resolution constraints is verified. Furthermore, under the same modulation order and symbol rate, the proposed modulation exhibits superior reliability compared to traditional modulation schemes such as PSK, FSK, and pulse position modulation (PPM).

    Aug. 08, 2025
  • Vol. 45 Issue 16 1606004 (2025)
  • Jianyong Yu, Xiaoli Hu, Qian Wang, Ling Qin, Fengying Wang, and Xinchao Kou

    ObjectiveVisible light positioning (VLP) has demonstrated promising outcomes in indoor environments. However, underground mines present significant challenges compared to ideal indoor conditions, substantially increasing positioning complexity. Traditional indoor positioning methodologies no longer adequately address underground mine positioning requirements. Consequently, developing a reliable and efficient positioning system for underground mines has become a critical challenge. In underground mining operations, the unpredictable nature of workers' head movements and variations in receiver height significantly impact visible light positioning performance. Thus, three-dimensional (3D) positioning presents a more suitable approach for personnel tracking in underground environments. During operations, head tilting introduces directional reception uncertainties, causing random fluctuations in received signal strength (RSS). This substantially reduces the accuracy of fingerprinting methods based on vertical received power and potentially leads to positioning failures. Moreover, reflections from irregular mine walls affect the optical power received by the photodetector (PD), resulting in unstable received power signals near walls and increased positioning errors. To improve positioning accuracy in complex underground environments, this study implements an inclination correction for vertical received optical power to accommodate tilted reception conditions (fingerprint correction). Additionally, we introduce a deep learning network that combines Transformer with MobileNetV3 (ViTs-MNV3) to process datasets from complex underground mine environments. Using point cloud technology, we simulate authentic mine environments to conduct comprehensive research on visible light positioning under these challenging conditions.MethodsCurrently, commonly used ranging-based positioning methods include RSS, time of arrival (TOA), time difference of arrival (TDOA), and angle of arrival (AOA). Among these approaches, fingerprinting-based positioning methods utilizing RSS are extensively researched due to their simple hardware requirements, operational simplicity, and capacity to achieve high positioning accuracy when integrated with deep learning networks for fingerprint database training. This study initially employs a conventional method to vertically collect optical power data for establishing a fingerprint database. However, receiver tilting causes the received optical power to deviate from the pre-established vertical fingerprint database, resulting in mismatches and decreased positioning accuracy. To address this challenge, we introduce a variable gain factor to convert the vertical fingerprint database into an inclined power database, thereby adapting to received power variations caused by receiver tilting and enhancing positioning accuracy. Furthermore, we propose a Transformer-enhanced MobileNetV3 network (ViTs-MNV3) for visible light positioning. This network utilizes the Transformer's multi-head attention mechanism to identify relationships between different input sequence components and learn global features, demonstrating enhanced performance. To increase simulation authenticity, we employ a depth camera to gather real-world point cloud data of mine walls and incorporate point cloud technology to construct a simulated underground mine environment that accurately reflects real-world conditions.Results and DiscussionsIn real-world environments, the fingerprint database collected under vertical reception conditions is used to train the proposed network. A comparative study is conducted between the proposed ViTs-MNV3 network and four other models: MobileNetV3, Transformer neural network, backpropagation (BP) neural network, and long short-term memory (LSTM) network. The results demonstrate that the proposed network outperforms the other four networks in terms of positioning accuracy in real-world scenarios, proving its feasibility and superior robustness (Table 2 and Fig. 6). In an ideal simulated environment, the proposed network achieves a root mean square error (RMSE) of only 2.36 cm and an average positioning error of 3.2 cm in 3D positioning (Fig. 7). However, in the simulation space incorporating point cloud data of walls, the influence of wall reflections became significant, increasing the RMSE to 5.42 cm at a reception height of 1.2 m (Fig. 9). When the receiver tilted, the received power no longer matched the vertical fingerprint database, leading to substantial positioning errors (Fig. 10). To validate the practical applicability of the proposed algorithm, real-world positioning experiments are conducted. Optical power data are collected at four different heights under vertical reception conditions, and after deep network training, the final 3D RMSE reaches 10.95 cm, with an average error of 8.36 cm. The results achieved centimeter-level positioning accuracy, with 90% of errors below 25 cm and a low proportion of large errors (Fig. 12). Finally, a variable gain factor is applied to correct the vertical fingerprint database in real-world environments, transforming it into an inclined fingerprint database to match the received power under tilted reception conditions (Fig. 13). The transformed power closely resembles the actual inclined power measurements, and the RMSE is reduced to 21.43 cm (Fig. 14). However, compared to Fig. 12, this approach results in higher positioning errors due to the additional errors introduced by wall reflections during power transformation.ConclusionsThis research presents a three-dimensional visible light positioning system designed for complex underground mine environments through the integration of 3D point cloud technology. The investigation examines multiple influential factors, including wall reflections, receiver height variations, and receiver tilting, each affecting positioning accuracy distinctly. Receiver tilting emerges as the most critical factor, resulting in an average positioning error of 2.17 m within a 4.0 m×3.0 m×2.5 m simulated environment, significantly impacting personnel positioning capabilities in underground mines. To mitigate this challenge, the study implements a variable gain factor to modify the vertical fingerprint database, converting it into an inclined fingerprint database that corresponds to received power under tilted reception conditions. This methodology substantially enhances positioning accuracy and facilitates the expansion of fingerprint databases for various tilt angles without requiring extensive manual data collection, thus optimizing resource efficiency. The positioning algorithm incorporates the strengths of the Transformer neural network and MobileNetV3 network, forming an innovative hybrid model, ViTs-MNV3, which exhibits exceptional performance in positioning applications. The algorithm achieves a RMSE of 2.36 cm in ideal simulated conditions and 5.42 cm in simulated environments with wall reflections. Experimental validation in real-world conditions demonstrates a RMSE of 10.95 cm under vertical reception conditions. Following power correction for tilted reception, although the RMSE increased to 21.43 cm, the corrected power values demonstrate substantial improvement in positioning accuracy compared to uncorrected tilted reception scenarios.

    Aug. 08, 2025
  • Vol. 45 Issue 16 1606005 (2025)
  • Jialin You, Tao Yang, and Bingjie Zhang

    ObjectiveUnder singularity conditions, the S-curve slope of timing recovery algorithms based on signal power approaches zero, hindering the timing error detector’s ability to extract timing errors accurately and establish an effective timing synchronization loop. Additionally, impairments from inter-symbol interference (ISI), residual chromatic dispersion (CD), differential group delay (DGD), and sample clock offset increase the convergence cost of establishing the timing synchronization loop. Previous research indicates that timing recovery algorithms suitable for faster-than-Nyquist (FTN) systems struggle to establish effective timing synchronization loops under singularity conditions. This paper addresses these challenges by proposing an anti-singularity and low convergence scheme for polarization demultiplexing and timing recovery based on orthogonal training sequences, analyzing the underlying principles of clock recovery algorithm limitations under singularity conditions.MethodsThis paper presents a joint scheme for polarization demultiplexing and timing recovery utilizing orthogonal training sequences, designed for low convergence cost and anti-singularity performance. The scheme calculates the channel matrix representing polarization crosstalk using orthogonal training sequences to mitigate singularity effects on the timing recovery algorithm. An adaptive equalization algorithm integrated within the timing loop compensates for inter-symbol crosstalk, residual chromatic dispersion, and differential group delay impairments before timing error calculation, enabling accurate determination of timing error adjustment direction. This approach reduces the convergence cost for establishing timing synchronization. The loop filter module generates control words based on timing errors, while the control unit calculates basic pointers and fractional intervals to guide the interpolation filter’s timing recovery process. Through continuous iteration of the timing feedback loop, the fractional interval stabilizes into periodic changes, and the root mean square of timing error detection converges to zero, indicating synchronization.Results and DiscussionsTo validate the proposed scheme, a triple-carrier 64 Gbaud polarization-multiplexed 16-ary quadrature amplitude modulation (PM-16QAM) FTN-WDM wavelength-division multiplexing (WDM) simulation platform was established in this paper. The simulation results demonstrate that the proposed algorithm could achieve timing recovery under singularity conditions [Fig. 5(f)]. Under 800 km transmission distance and 0.90 and 0.85 acceleration factor, the proposed scheme could reduce the convergencecost at least 39%/41%, respectively, comparedto conventional scheme (Fig. 6). Furthermore, under the same conditions, the proposed scheme improves the OSNR tolerance by 0.4 dB and 0.6 dB (@64 Gbaud PM-16QAM FTN, BBER=2×10-2) compare with the conventional scheme (Fig. 7). Additionally, offline experimental results for a triple-carrier 40 Gbaud PM-16QAM FTN system reveal that under 0.90 and 0.85 acceleration factor, the proposed scheme could reduce the convergence cost required forestablish timing synchronization loop by at least 37%/38%, respectively, compared to the conventional scheme (Fig. 10). Moreover, under 800 km transmission and 0.90 and 0.85 acceleration factor, the proposed scheme achieves OSNR tolerance improvements of 0.50 dB and 0.75 dB (@40 Gbaud PM-16QAM FTN BBER=2×10-2) compare with the conventional scheme (Fig. 11).The proposed scheme demonstrates reduced computational complexity, approximately 28% lower than the conventional scheme (which includes adaptive equalization, polarization demultiplexing algorithm, and squared Gardner phase detector timing recovery algorithm), while requiring only 2.9% training sequences under identical acceleration factor, DGD, and azimuth conditions (Table 1).ConclusionsThis research presents an anti-singular and low convergence cost scheme for polarization demultiplexing and timing recovery based on orthogonal training sequence (TS). The proposed approach effectively achieves polarization demultiplexing under singularity conditions using orthogonal TS. The integration of adaptive equalization before timing error detection compensates for DGD, ISI, and residual CD impairments, significantly reducing convergence cost. The scheme requires approximately 2.9% TS overhead while achieving 28% reduction in computational complexity compared to conventional methods. Simulation results for 64 Gbaud PM-16QAM FTN-WDM demonstrate convergence cost reductions of 39%/41% at 0.90 and 0.85 acceleration factors over 800 km transmission, with OSNR tolerance improvements of 0.4 dB and 0.6 dB (@64 Gbaud PM-16QAM FTN BBER=2×10-2). Offline experiments with triple-carrier 40 Gbaud PM-16QAM FTN-WDM show OSNR tolerance improvements of 0.50 dB and 0.75 dB (@40 Gbaud PM-16QAM FTN BBER=2×10-2) at 800 km transmission with 0.90 and 0.85 acceleration factors. These results confirm the proposed scheme’s effectiveness in achieving low convergence cost and anti-singular polarization demultiplexing and timing recovery in FTN systems.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1606006 (2025)
  • Ying Wu, Jianjun Yu, Jianyu Long, Qingyi Zhang, Ying Wang, Yifan Chen, Wen Zhou, and Kaihui Wang

    ObjectiveDriven by the global data surge, free-space optical (FSO) communication at 1550 nm has become essential for 6G backhaul and inter-satellite links due to its anti-electromagnetic interference capabilities. However, atmospheric turbulence and beam misalignment reduce single-mode fiber efficiency and limit practical application. To address this, a hybrid architecture combining few-mode/multimode fiber (FMF/MMF) with FSO is proposed. Mode-division multiplexing (MDM) at the receiver decomposes signals into spatial modes, which are reconstructed using modular diversity reception and adaptive multiple input multiple output-digital signal processing (MIMO-DSP), effectively suppressing turbulence and misalignment. This system improves robustness through mode/polarization diversity, reducing inter-mode crosstalk via optimized DSP, and using cost-effective 1550 nm components. Despite challenges in receiver-side signal synthesis—such as hardware complexity in coherent schemes and high cost in programmable interrupt controller (PIC) solutions—a novel FMF-FSO structure is introduced. It integrates time-division multiplexing reception and a least mean squares/ constant modulus algorithm (MIMO-LMS/CMA)-based maximal ratio combining (MRC) algorithm. With six spatial modes and polarization multiplexing, 12-channel 32 GBaud quadrature phase shift keying (QPSK) transmission is achieved. We hope this architecture maintains resilience against hybrid channel damage, and its compatibility with the existing single-mode infrastructure can improve the application of adaptive fiber-FSO systems in future optical networks.MethodsTo address the hardware redundancy of multiple-receiver structures in traditional FMF-FSO systems, we propose a time-domain diversity (TDM)-based receiver architecture (Fig. 1) for comprising three modules: 1) A signal generator outputs a rectangular wave with a period T=80 μs, triggering an acousto-optic modulator (AOM) to produce a pulse width ΔT=8 μs. 2) A mode demultiplexer decomposes the input beam into six orthogonal modes, which are truncated by the AOM and injected with gradient delays (ΔL=2 km, total delay Δτ=6.66 μs) into standard single-mode fiber (SSMF). 3) An optical coupler (OC) incoherently combines these time-divided signals into a serial output. This setup allows multi-mode signal reception via a single channel, reducing hardware complexity by 83.3% compared to traditional multi-channel setups. The coherent receiver uses MIMO-LMS/CMA and MRC algorithms to suppress inter-channel crosstalk and enhance signal-to-noise ratio (SNR) without increasing bandwidth, maximizing spectral efficiency. In transmission, dual-polarized quadrature amplitude modulation (QAM) signals are generated from pseudo-random binary sequence (PRBS), shaped with a roll-off factor α=0.01, and modulated at 1550 nm using in-phase/quadrature (I/Q) modulators. Signals are polarization-multiplexed, amplified, and combined via a mode multiplexer into 12 channels over FMF. A tunable variable optical attenuator (VOA) simulates FSO turbulence and misalignment by axial fiber displacement (0?15 μm). After space-mode separation, the TDM receiver uses gradient-delayed SSMF to recombine signals, which are then coherently detected and processed digitally. Advanced DSP compensates for timing offsets and I/Q imbalance using generalized serial orthogonalization process (GSOP) and the joint MIMO-LMS/CMA-MRC algorithm, which mitigates polarization mode dispersion (PMD), mode coupling, differential group delay (DGD), and mode-dependent loss (MDL)—demonstrating the TDM receiver’s feasibility in FMF-FSO hybrid systems.Results and DiscussionsThe bit error rate (BER) performance of a 6-mode 32 GBaud polarization division multiplexing (PDM)-QPSK system using the MIMO-LMS/CMA-MRC algorithm is evaluated across a received optical power range from -6 dBm to 8 dBm. The input power is measured at the mode demultiplexer, and the best BER among all 12 channels is selected as control group to reduce the influence of polarization and mode coupling. Experimental results show that the BER consistently stayed below the 7% hard-decision FEC threshold (3.8×10-3), surpassing traditional DSP schemes which meet this threshold only when the received power reaches -2 dBm (Fig. 4). Even at -6 dBm, the system achieves a BER of 3.3×10-3, demonstrating high sensitivity. At powers ≥2 dBm, near error-free transmission is achieved (BER<7.6×10-6). Notably, the noise suppression gain of the algorithm increases with channel attenuation, with an average 4 dB improvement in sensitivity. Meanwhile, the constellation diagrams at 0 dBm further confirmed the effectiveness of the algorithm due to linear compensation and optimal combining. To evaluate the impact of modal delay, a tunable delay line (0?10 km) is introduced to decorrelate the six modes. In conventional DSP, BER exceeds the hard-decision FEC threshold due to modal delay. However, the proposed algorithm suppresses BER to 3.2×10-3 at ≥-2 dBm. At 8 dBm, BER increased due to receiver saturation. Even with delay lines, the algorithm reduces signal error, validating its robustness (Fig. 5). A short-distance FSO experiment using a graded index (GRIN) lens and adaptive receiver further confirmed its effectiveness. Compared to a benchmark BER of 1.3×10-3 in traditional FSO systems, the FMF-FSO system with the proposed algorithm achieves near error-free performance, showing three orders of magnitude improvement and strong resistance to atmospheric fading and spatial coherence degradation (Fig. 6).ConclusionsWe propose a novel few-mode fiber-free space optical (FMF-FSO) transmission system, integrating TDM receiver with improved MIMO-LMS/CMA-MRC equalization algorithm. The architecture employs PDM and MDM to simultaneously deliver 32 GBaud QPSK signals across 6 linearly polarized modes (LP 01, LP 02, LP 11a/b, LP 21a/b), establishing a 12-channel multiplexing framework. At the receiver DSP process, a hybrid MIMO-LMS/CMA-MRC equalization algorithm achieves an average improvement of 4 dB in power sensitivity by adaptively optimizing weight matrices, effectively mitigating atmospheric disturbances and alignment fluctuations in FSO transmission. Experimental validation demonstrates a BER of 3.3×10-3 with received optical power of -6 dBm, compliant with 7% hard-decision FEC thresholds. The system’s compatibility with existing single-mode infrastructure and demonstrated resilience to hybrid channel impairments position it as a viable candidate for next-generation long-haul optical networks requiring adaptive fiber-FSO convergence.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1606007 (2025)
  • Hailong Yang, Bing Lu, Weigang Hou, Wei Huang, Ye Zheng, Pengxing Guo, Bin Li, and Lei Guo

    ObjectiveWith the rapid advancement of information technology, the volume of data transmitted via optical fibers has grown exponentially. However, illegal eavesdroppers can exploit methods such as bending fibers and tapping into adjacent-channel crosstalk to cause slight leakage of optical signals and steal information. Ensuring the security of data transmission in optical fiber networks has thus become a critical issue that demands urgent resolution. Optical covert communication technology conceals information within the optical noise of a public transmission channel. Among noise sources, amplified spontaneous emission (ASE) light, with its broad bandwidth and high noise intensity, serves as an excellent carrier for hiding transmitted signals. Meanwhile, the randomness, sensitivity to initial conditions, and unpredictability of chaotic systems provide a novel approach to enhancing physical-layer information security in optical covert communication. Based on this, we propose an optical covert communication system employing chaotic digital encryption with an ASE light source, featuring a key space of up to 1087. Simulation results demonstrate that over a 50 km transmission distance, the covert channel achieves error-free reception at an optical received power of -10 dBm, while the public channel does so at -19 dBm. Moreover, the covert channel introduces only a 0.5 dB power loss to the public channel. Even if an eavesdropper is aware of the covert channel’s existence, the intercepted information exhibits a bit error rate (BER) as high as 0.5 without the encryption key. This scheme effectively ensures communication security and covertness while imposing minimal impact on the public channel.MethodsThe structure of the covert communication system is illustrated in Fig. 3, primarily consisting of a covert channel transmitter, receiver, and a public channel. The sender, Alice, and the authorized user, Bob, pre-share the same encryption key. At the transmitter, the covert signal, processed by chaotic digital encryption, undergoes power adjustment via a variable optical attenuator (VOA) to ensure its power is 12.8 dB lower than that of the public channel, thereby achieving signal concealment. The covert signal is then coupled into the public channel using an optical coupler (OC 3) and transmitted over 50 km of single-mode fiber (SMF). At the receiver, Bob employs dispersion-compensating fiber (DCF) to precisely counteract the chromatic dispersion introduced during transmission. The covert signal is then extracted from the public channel using an optical filter, and finally, the original covert signal is recovered by matching the dispersion value with the chaotic digital decryption process.Results and DiscussionsThe simulation results demonstrate that when the covert channel’s power is 12.8 dB lower than that of the public channel, the covert signal becomes completely hidden in the frequency domain [Fig. 7(a)], preventing eavesdroppers from detecting its presence through spectral observation. Without compensating for the dispersion introduced by the covert channel’s encryption or isolating the two branches of the covert signal, the covert channel exhibits noise-like characteristics in the time domain [Fig. 8(a)]. Even if an eavesdropper demodulates the public channel and examines its eye diagram, the comparison of eye patterns reveals no detectable trace of the covert channel (Fig. 9). Furthermore, the covert channel introduces only a 0.5 dB power penalty to the public channel [Fig. 10(a)]. When the received optical power of the covert channel reaches -10 dBm, error-free reception is achieved, and the impact of chaotic digital encryption on transmission performance is negligible [Fig. 10(b)].ConclusionsFacing these rapidly growing demands for information transmission security and covertness driven by the continuous evolution of information technology, we propose an optical covert communication system based on the integration of chaotic digital encryption and ASE light. Owing to its inherent randomness, sensitivity to initial conditions, and unpredictability, the chaotic system serves as an ideal choice for enhancing information confidentiality. By applying chaotic digital encryption to covert signals and then modulating them onto ASE light for covert transmission, this scheme provides dual-layer security protection concealment and confidentiality thereby strengthening the overall system security. Simulation results demonstrate that by adjusting the power of the chaotic digitally encrypted signal carried by ASE light to be 12.8 dB lower than that of the public channel, the covert signal can be effectively hidden in both frequency and time domains, while introducing only a 0.5 dB power penalty to the public channel. Furthermore, the proposed scheme achieves a key space of up to 1087, enabling robust resistance against brute-force attacks by eavesdroppers even when they are aware of the covert channel’s existence. This provides a solid safeguard for information security.

    Aug. 19, 2025
  • Vol. 45 Issue 16 1606008 (2025)
  • Ming Zhang, Qianshi Zhang, Yifei Xuan, Weijuan Cheng, Lixing Shi, Yuanhan Wang, Hailin Wen, and Ying Du

    ObjectiveIn recent years, the demand for high-precision sensing technologies has grown substantially across critical sectors such as industrial monitoring and medical applications. Within these domains, shape sensing emerges as a pivotal challenge that forms the foundation for device and process monitoring. While fiber Bragg grating (FBG) sensors demonstrate significant potential for shape sensing applications, they are confronted with three primary error sources: manufacturing inconsistencies, insufficient demodulation accuracy, and spectral drift. These errors collectively compromise reconstruction accuracy, with spectral drift being particularly exacerbated by material creep and prolonged sensor operation. To address this challenge, we propose a novel spectral drift correction method employing a hybrid bi-directional long short-term memory-long short-term memory (BiLSTM-LSTM) neural network architecture. The developed approach aims to enhance both the accuracy and operational stability of FBG-based shape sensing systems through effective drift compensation and precision optimization in shape reconstruction.MethodsThe proposed correction method strategically integrates BiLSTM and LSTM networks through a sequential processing architecture. The BiLSTM layer extracts bidirectional temporal features from FBG sensor data, establishing comprehensive temporal dependencies through its dual-directional processing capability. These features are subsequently refined by the LSTM layer to enhance spectral-curvature relationship modeling efficiency. Training data consists of 2690 samples with curvature variation of 0.05?0.15 m-1 obtained from a dedicated FBG sensing platform. Preprocessing incorporates Gaussian smoothing, controlled noise injection, and spectral drift simulation to emulate operational disturbances. Critical parameters including network units, dropout rate, and learning rate are optimized via particle swarm optimization (PSO).Results and DiscussionsResults demonstrate significant performance enhancement in FBG shape sensing through the proposed BiLSTM-LSTM architecture. Curvature reconstruction errors decreased from [15.7609,14.4787,20.5649,18.2307] mm to [0.1391,2.5222,1.7354,1.2951] mm across various configurations, achieving 44% average accuracy improvement over benchmark models: Frenet, convolutional neural network (CNN), LSTM-SelfAttention. As evidenced in Figs. 8 and 9, the correlation between predicted and actual curvature values exhibits exceptional consistency with a mean R2 of 0.9918. Residual analysis in Fig. 8(b) confirms normally distributed errors (mean is about 0), validating the model’s precision and stability. Comparative evaluations highlight the model’s dual advantages: surpassing Frenet-based methods in baseline accuracy while outperforming CNN and LSTM-SelfAttention architectures in error minimization: mean absolute error/total error (MAE/TE) reduction. By resisting long-term spectral drift, this capability enables sustained high-precision reconstruction. Such reliability is critical for structural health inspection and industrial monitoring applications.ConclusionsWe present a hybrid BiLSTM-LSTM network addressing FBG spectral drift through temporal dependency modeling and spectral-curvature nonlinear mapping. Experimental validation shows 44% accuracy improvement over traditional methods (e.g., Frenet framework) and other deep learning models. This study paves the way for more accurate and reliable FBG shape sensing systems, with potential applications in structural health inspection and industrial pipeline inspection.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1606009 (2025)
  • Jingjin Zhang, Jilin Zheng, Hua Zhou, Tao Pu, Jin Li, Yang Liu, and Xiaolong Zhao

    ObjectiveThe global electromagnetic spectrum has become increasingly congested due to widespread deployment of radar systems, satellite communications, wireless radio frequency transmission, and next-generation dynamic communication networks (e.g., 5G/6G). Frequency-hopping (FH) signals, characterized by dynamically varying carrier frequencies, present significant advantages in anti-jamming, anti-interception, and spectral efficiency optimization, establishing them as an optimal solution for addressing spectrum scarcity. Contemporary FH communications require ultra-wide bandwidth at high-frequency bands and rapid hopping speeds to enhance anti-jamming capabilities. Traditional radio frequency (RF) frontend technologies face inherent limitations due to the “electronic bottleneck”, constraining their operational frequency range, bandwidth, and hopping speed. Current systems typically achieve hopping rates at millisecond levels, with signal bandwidths restricted to several GHz and operational bands limited to the centimeter-wave regime. These limitations impede the development of next-generation ultra-wideband fast FH communication systems. This paper introduces a photonics-based FH signal generation and channelized reception architecture to address these challenges. The proposed solution utilizes photonics-assisted microwave technology, capitalizing on the inherent advantages of optical-domain processing, including broad bandwidth, high-speed operation, low power consumption, and immunity to electromagnetic interference. This approach effectively overcomes the constraints of conventional electronic methods, thereby advancing FH systems in high-density spectral environments.MethodsThis paper proposes a photonics-based FH signal generation and channelized reception architecture, employing symmetrical structures and identical operating principles for both transmitter and receiver. The methodology encompasses three primary phases. Initially, dual optical frequency combs (OFCs) with distinct free spectral ranges (FSRs) are generated using Mach-Zehnder modulators (MZMs), functioning as broadband FH local oscillator (LO) sources and signal reception LO sources, respectively. The system then achieves parallel separation of FH channels and signal reception channels through the integration of dual OFCs with wavelength demultiplexers at both transmitter and receiver ends, enabling ultra-wideband FH bandwidth and channelized reception. High-speed optical switches control synchronous switching of FH and reception channels, facilitating rapid FH signal generation and synchronized reception. Additionally, carrier-suppressed single-sideband modulation (CS-SSB) is implemented for both optical carrier and LO signals, enabling small-range frequency shifting and reconfigurable hopping channel allocation. The architecture’s feasibility is validated through theoretical analysis and comprehensive functional verification via simulation experiments on the Optisystem 15.0 platform.Results and DiscussionsThe system’s functionality is validated through a simulation platform established for both transmitter and receiver components, following the architecture in Fig. 1 and parameter settings in Table 1. The bandwidth and quality assessment of transmitted FH signals reveals seven hopping channels spanning 35?65 GHz, achieving a 30 GHz operational bandwidth with 5 GHz channel spacing, when configured with a 0 GHz LO branch frequency shift and 5 GHz single-tone input signal (Fig. 5). All channels demonstrate a carrier-to-noise ratio (CNR) exceeding 50 dB with power fluctuations of approximately 2 dB. Channel switching simulation, configured with a 100 ns optical switch transition time and 1 μs dwell time per channel, yields a measured switching time of 101.2 ns (Fig. 6), aligning with the optical switch specifications. Receiver performance evaluation utilizing a 4QAM signal (5 GHz carrier frequency, 2.048 Gbit/s data rate) demonstrates transmitter output channels with amplitude fluctuations of ~1.93 dB and error vector magnitude (EVM) values below 4.0%. Post-dehopping signals exhibit amplitude fluctuations of ~3.70 dB and EVM values <8.5%, indicating approximately 5 percentage points EVM degradation. Transmission performance analysis through bit error rate (BER) curves shows saturation below -30 dBm and above -15 dBm received power, achieving a BER of 1.9×10-3 at -21.52 dBm, remaining below the 7% forward error correction (FEC) threshold.ConclusionsThis paper presents a photonic HF signal generation and reception system designed for high-speed broadband signal transmission and channelized reception. The system architecture and operating principles of frequency hopping/de-hopping are theoretically analyzed and validated through simulation experiments. The experimental results demonstrate several significant advantages. First, the system generates ultra-wideband (35?65 GHz with 30 GHz bandwidth), high-frequency (millimeter-wave), and rapid FH (100 ns) signals. Second, it demonstrates robust broadband signal transmission and reception capabilities, effectively processing 5 GHz-carrier 4QAM signals at 2.048 Gbit/s. The EVM at the transmitter and receiver remains below 4.0% and 8.5%, respectively. At a received power of -21.52 dBm, the system achieves a BER of 1.9×10-3, below the 7% FEC threshold. Moreover, an electrically controlled precision frequency tuning mechanism enables arbitrary frequency switching throughout the operational band, ensuring exceptional adaptability. The architecture demonstrates remarkable reconfigurability. Modification of OFC generation parameters enables dynamic reconfiguration of hopping channel allocation, while expansion of the OFC spectral width effectively extends the hopping coverage range. Compared to conventional approaches, this system successfully generates millimeter-wave ultra-wideband rapid FH signals while achieving channelized reception. Additionally, it supports high-rate transmission of broadband signals, integrating ultra-wide bandwidth, rapid hopping agility, and enhanced data throughput. The system’s inherent flexibility and reconfigurability make it suitable for various application scenarios in complex electromagnetic environments. These features establish it as a promising solution for anti-jamming secure communications, electronic warfare systems, and next-generation radar architectures.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1606010 (2025)
  • Ziyue Zou, Yifan Sun, and Lingjun Kong

    ObjectiveHolography, an innovative optical technique for recording and reconstructing object information through wavefront modulation, was first introduced by physicist Dennis Gabor in 1948. As optical science has advanced, researchers have expanded beyond traditional interferometric methods to develop various holographic approaches, including digital holography and computational holography. Currently, multiple optical degrees of freedom—including polarization, wavelength, and temporal modes—have been utilized to enhance holographic systems. Among these, orbital angular momentum (OAM)-based holography emerges as particularly significant due to its inherent advantages. OAM, distinguished by theoretically infinite mode indices (corresponding to OAM quantum numbers), presents exceptional potential for expanding holographic information capacity through encoding data in spatially structured phase profiles. Traditional OAM holography has demonstrated benefits including high-dimensional multiplexing and enhanced channel capacity in holographic encoding. Recent developments extend OAM holography to entangled photon pairs, achieving superior precision and stability compared to classical implementations. However, the dependence on quantum entangled light sources, requiring complex experimental conditions, restricts practical scalability. Concurrently, classical optical systems that simulate quantum correlations have emerged as a viable alternative. This research presents an OAM holography scheme utilizing classical correlated light. Through comprehensive theoretical modeling and numerical simulations, we confirm its viability and demonstrate performance comparable to quantum entanglement-based methods. By connecting classical optics with quantum-inspired approaches, this work expands the theoretical framework of optical correlations while offering a cost-effective path to advance holographic technologies for practical applications.MethodsThis study investigates holographic techniques based on OAM using classically correlated light. First, we design a holographic scheme leveraging the properties of classical correlations and OAM. Subsequently, theoretical derivations are conducted to validate the feasibility of the proposed approach. Numerical simulations are then performed to model light field propagation, interference, and holographic reconstruction, thereby illustrating the theoretical performance of the scheme. Finally, a comparative analysis with quantum OAM holography is undertaken to highlight the advantages and potential applications of our classical approach. To ensure the practicality of the scheme, numerical simulations of the classically correlated light source are specifically implemented to verify its compliance with the incoherence condition—a prerequisite for exploiting statistical correlations in this study. Additionally, given the formal resemblance between the generated classical correlated light and quantum entangled sources, we adopt correlation detection methods analogous to quantum entanglement verification. Our results demonstrate that, under specific projection measurements, the classical correlated light exhibits properties akin to quantum entanglement. This finding further justifies its mathematical representation as a classical correlated state, reinforcing the validity of our theoretical framework.Results and DiscussionsWe conduct numerical simulations of the proposed OAM-based classically correlated light holography process. The imaging results (Fig. 5) demonstrate excellent agreement with theoretical predictions. Notably, the theoretically unbounded OAM quantum numbers enable our holographic scheme to achieve OAM-multiplexed holography for enhanced channel capacity. Compared to conventional OAM-based holography, our protocol exhibits inherent robustness against classical stray light interference (Fig. 7), as holographic information is encoded within the correlations of classical light fields. Furthermore, leveraging structural similarities between classical correlations and quantum entanglement, we demonstrate the feasibility of implementing high-security holographic encryption protocols (Fig. 8). Significantly, while the proposed classical correlated light holography shares comparable imaging processes, outcomes, and inherent advantages with quantum entangled OAM holography, it provides substantially improved experimental feasibility and cost-effectiveness.ConclusionsThis research presents a classical correlated light holography scheme based on OAM, with its theoretical viability thoroughly validated through theoretical derivation and numerical simulations. The proposed scheme demonstrates formal, characteristic, and imaging equivalency to quantum entanglement holography utilizing OAM. Significantly, it addresses practical limitations inherent in quantum-entangled light systems—including stringent preparation requirements, complex modulation protocols, transmission instability, and reduced operational robustness—thereby offering enhanced feasibility for real-world implementations. Compared with conventional classical holography, the OAM-based classical correlation scheme demonstrates superior robustness by maintaining holographic information integrity and accuracy under classical stray light interference, while achieving heightened security for holographic encryption relative to standard OAM-carrying classical light. Furthermore, the intrinsic properties of OAM and classical correlations suggest promising applications in cutting-edge domains such as high-channel-capacity holographic systems. These findings advance holography development while expanding its technical horizons. However, current research remains primarily theoretical, with experimental validation and practical performance evaluation requiring further investigation. Potential implementation challenges—including precision optical alignment, interferometer stability, and environmental noise suppression—demand thorough exploration to assess system adaptability and reliability in operational scenarios. Future work will focus on experimental optimization, comprehensive property characterization, and deeper investigation of the scheme’s physical principles and technical advantages, thereby establishing a robust foundation for expanding its application spectrum.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1609001 (2025)
  • Mingbai Li, Lun Jiang, Xu Zhang, and Ziru Niu

    ObjectiveNeutral wind plays a critical role in the dynamics of the upper atmosphere, and accurate measurements of thermospheric wind fields are essential for the comprehensive understanding and modeling of the ionosphere-thermosphere (IT) system. The Doppler asymmetric spatial heterodyne (DASH) interferometer employs a limb-viewing observational technique, similar to the occultation method, to measure atmospheric wind fields at a fixed viewing angle along the Earth’s tangent direction. In practical applications, raw interferograms are affected by both background and instrumental noise, which reduces fringe contrast. In the frequency domain, this degradation is reflected in weakened dominant frequency components and a low spectral signal-to-noise ratio (SNR), potentially resulting in biased frequency estimation or main-lobe energy leakage. Enhancing fringe contrast is therefore essential for amplifying the dominant frequency peak and improving the accuracy of phase retrieval. Consequently, the development of an effective contrast enhancement strategy is vital for enhancing wind speed measurement performance, particularly under low-signal conditions where it holds substantial practical value.MethodsDuring the inversion process of low-contrast interferograms, the spatial frequency of interference fringes is highly sensitive to preprocessing due to the low SNR, which can easily cause phase discontinuities and significantly reduce the accuracy of wind field inversion. To address this challenge, we present a contrast-adaptive enhancement algorithm for spaceborne DASH interferograms. As illustrated in Fig. 4, the algorithm first analyzes the image grayscale histogram and adaptively determines the clipping ratio by combining the local standard deviation, histogram dynamic range, and entropy. This ratio, along with the cumulative distribution function, defines the clipping region, effectively mitigating the influence of outlier pixels on contrast enhancement. The clipped grayscale range is then linearly mapped to the full dynamic range, resulting in a substantial improvement in interferogram contrast. The enhanced image is then processed through interferogram preprocessing and wind field inversion, enabling stepwise analysis of atmospheric wind speed distributions at different altitudes.Results and DiscussionsDespite ongoing advancements, existing interferogram contrast enhancement techniques remain limited in effectiveness. Conventional approaches such as linear stretching and adaptive histogram equalization (AHE) offer only moderate performance under challenging conditions. In this paper, simulated DASH interferograms with varying levels of contrast and superimposed random noise are processed using four methods: the proposed adaptive threshold regulation algorithm, frequency-domain enhancement, contrast-limited adaptive histogram equalization (CLAHE), and standard AHE. As shown in Fig. 6, the first curve represents the output of the proposed method. Under low-contrast conditions, this method exhibits superior noise suppression compared to the other methods, effectively enhancing fringe visibility while minimizing phase fluctuations. As illustrated in Fig. 7, results averaged across multiple trials reveal that when interferogram contrast exceeds 0.4, additional enhancement leads to a decline in inversion accuracy. This is primarily due to the distortion of fringe structure, which introduces spurious frequency components and shifts the dominant frequency in the Fourier domain, thereby impairing the precision of phase extraction. Therefore, for interferograms with inherently high contrast, it is advisable to bypass contrast enhancement and directly apply frequency-domain analysis to preserve the integrity of phase information. To further assess the relationship between fringe contrast and wind retrieval uncertainty, an indoor wind simulation experiment is conducted using the DASH optical system. As depicted in Fig. 9, when the interferogram contrast falls below 0.2, the signal becomes weak, and fringes are highly susceptible to noise, resulting in substantial errors in retrieved wind speeds. Following contrast enhancement, the retrieved wind speeds ranged from 22 to 28 m/s, with corresponding errors between 2.6~8.6 m/s. The primary contributors to these errors include uncertainties in the rotational speed of the wind-generating disc and deviations in the optical path. In addition, wind field retrievals are conducted on three sets of satellite limb-viewing data acquired on June 16. As shown in Fig. 11(a), the inversion results from the unenhanced interferogram closely follow the trend predicted by the horizontal wind model 2014 (HWM 14), yielding a root-mean-square (RMS) error of 9.2 m/s. After applying the proposed contrast enhancement [Figs. 11(b) and (c)], the SNR of the interferograms improved by more than a factor of 3, and the accuracy of the retrieved wind fields increased markedly. These RMS errors decreased from 643.55 m/s and 541.71 m/s to 17.99 m/s and 19.62 m/s, respectively, corresponding to error reductions of 97.2% and 96.4%, with an average reduction of 96.8%.ConclusionsThrough simulation and experimental analyses, we confirm that low-contrast interferograms adversely affect wind speed retrieval accuracy. Additionally, significant variations exist among different contrast enhancement algorithms in their ability to recover interferogram details. These results demonstrate that the contrast enhancement method based on adaptive threshold regulation effectively improves phase retrieval accuracy for low-contrast interference fringes. However, for high-contrast fringes, reconstructing the grayscale histogram can lead to a loss of phase information, and thus, no further contrast enhancement is applied following the optimized enhancement strategy. Based on experimental results, the error is controlled within 10 m/s in both simulation and indoor experiments, while the variance between satellite-based data and the HWM 14 remains within 20 m/s, reflecting a small yet statistically significant difference. Part of this discrepancy can be attributed to instrumental errors, such as noise interference and zero-wind calibration. The remaining differences may arise from atmospheric wind fluctuations or mismatches between the observation geometry and the HWM 14’s projection.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1610001 (2025)
  • Changchun Jiang, Shengjie Liu, Mulin Yao, Junwei Li, and Hao Xian

    ObjectiveThe exponential proliferation of space debris in near-Earth orbits presents substantial risks to operational spacecraft and crew safety. Current space surveillance systems, primarily functioning at night, encounter significant limitations in detecting small, low-brightness debris against intense daylight backgrounds. The development of low-SNR dim target detection technologies under strong skyglow is essential for maintaining space asset security and sustainable orbital environments. Existing approaches are categorized into hardware optimization and algorithmic development. While spectral filtering is relatively mature, it remains sensitive to filter window configurations; polarimetric filtering faces limitations due to spatiotemporal variations in sky polarization; detector performance enhancement remains challenging; spatial-frequency filtering shows parameter sensitivity; visual saliency detection encounters difficulties with adaptive thresholding for varying target sizes; low-rank sparse decomposition methods face localization inaccuracies and computational constraints; deep learning approaches are limited by insufficient training data. This study introduces a velocity estimation and correlation decision-based detection framework specifically designed for low-signal-to-noise ratio (SNR) moving targets under strong skyglow. Numerical simulations and experimental validations demonstrate enhanced detection capabilities, define operational parameters, and verify engineering feasibility, providing a practical solution for dim target monitoring.MethodsTo address signal correlation degradation caused by inter-frame displacement in low-SNR moving target detection, this study presents a velocity-iteration-estimation aberration modulation correlation method (VAMCM). The method initially applies motion compensation to image frame sequences through affine transformation, utilizing the shift-and-add principle to restore temporal correlation between targets and aberration-modulated signals. Subsequently, a ??three-tier progressive velocity estimation architecture?? iteratively refines the velocity search range within constrained parameter spaces, optimizing motion parameters through hierarchical iterations. The implementation includes a 3σ statistical verification mechanism to validate target authenticity, optimizing computational efficiency and detection accuracy. Furthermore, periodic aberration perturbations characterized by Zernike polynomials are actively introduced to generate spatiotemporally coupled modulation signals, enhancing target discriminability. The method’s effectiveness is validated using a synthetic dataset incorporating motion trajectory simulation, Poisson noise modeling, and active aberration modulation, enabling comparative analysis with existing approaches. The validation process concludes with indoor and field experiments to verify practical efficacy.Results and DiscussionsThe VAMCM demonstrates superior performance in low-SNR moving target detection through comprehensive simulations, indoor experiments, and field tests. Simulation results indicate that VAMCM achieves a detection probabilities exceeding 90% and a sensitivity over 95% with a false alarm rate below 10% and trajectory errors stabilized at 0.39 pixel/frame when SNR is not smaller than 2 (Fig. 10 and Fig. 11). Traditional tensor decomposition methods (e.g., TT, TR, and ASTTV-NTLA) fail completely at SNR of smaller than 4, with false alarm rates approaching 50% (Figs. 13?17). Indoor experiments show VAMCM achieving 45% detection probability and 59.4% sensitivity at SNR of 3.39, substantially outperforming alternative methods (all 0%). At SNR of 9.47, it attains 100% detection probability and sensitivity with trajectory errors converging to 1.003 pixel/frame, while other methods demonstrate limited effectiveness even at high SNR levels (e.g., 4DST-BTMD achieves 100% detection probability but only 50% sensitivity) (Table 3). Field tests under complex astronomical conditions (SNR of about 2.22) confirm VAMCM’s successful target detection, while alternative methods fail entirely (Fig. 22). Analysis reveals that VAMCM addresses conventional methods’ insufficient target-background separation capability in low-SNR scenarios through spatiotemporal matching filters constructed via velocity iteration estimation and aberration modulation. However, computational efficiency remains constrained by high time consumption, with velocity iteration comprising 95.67% of processing time (Fig. 12).ConclusionsThis research addresses the detection of low-SNR moving targets under strong skyglow interference. Through the integration of velocity iteration estimation with active wavefront modulation and exploitation of aberration response differences between targets and background noise, the study presents a local enhancement-velocity estimation-target verification detection method for low-SNR target identification and trajectory resolution. Simulation analysis demonstrates superior performance with detection probability and sensitivity exceeding 90% at SNR of 2, while maintaining false alarm rates below 10% and achieving trajectory mean absolute error within 1 pixel/frame. Indoor experimental validation shows comparable performance (equivalent to simulation results at SNR of 2) at approximately SNR of 4.3 under practical conditions. Field experiments confirm the method’s system integration capability and effective detection of spatial targets with SNR of about 2.2 under strong daylight interference. VAMCM exhibits three key advantages: reliable target discrimination capability, sub-pixel level trajectory accuracy, and enhanced low-SNR adaptability. The approach demonstrates significant improvements over conventional tensor decomposition-based methods in detection thresholds, expanding active aberration modulation techniques’ application scope. VAMCM offers an innovative solution for real-time detection and tracking of faint moving targets under intense daylight conditions. Future research will focus on computational efficiency optimization and application extension to diverse dynamic target scenarios, enhancing operational adaptability across various environments.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1610002 (2025)
  • Yiming Zhang, Fangnan Hao, Zili Xu, and Guang Li

    ObjectiveThe advantages of low costs, non-contact and full-field measurements have contributed to the widespread adoption of digital image correlation (DIC) technique across various fields, including material science, bio-mechanics and structural engineering. In DIC technique applications, the displacement field solution for speckle images constitutes a fundamental component, with its accuracy directly influencing the quality of subsequent structural deformation behavior analysis. Although the traditional method’s theoretical framework has achieved considerable maturity, the inherent parameter selection challenge remains unresolved. The accuracy and stability of displacement field solutions are affected by factors such as subset sizes, shape functions, and optimization algorithms. These parameter selections typically depend on prior knowledge of deformation conditions. Recent advances in deep learning-based optical flow estimation have sparked interest in exploring deep learning methods for speckle image displacement field solutions. However, these approaches require extensive labeled datasets for network pre-training to establish the mapping between speckle images and displacement fields. The labeled datasets, primarily generated through specific displacement field simulations, exhibit limited generalization capability and struggle to deliver satisfactory results in practical applications. Additionally, the pre-training process demands substantial computational resources. This study introduces a self-supervised learning concept and proposes an adaptive displacement field solving method for speckle images to address parameter selection issues and eliminate dependence on pre-training with large-scale labeled datasets.MethodsThe traditional solution method’s concept of one-to-one correspondence between pixel coordinates in deformed and reference subsets (Fig. 1) provides valuable insights. This paper presents a self-supervised learning framework for speckle images (Fig. 2). The process begins with region of interest (ROI) delineation on the reference image and assembly of all pixel coordinates within the ROI. The framework utilizes an artificial neural network to characterize the mapping relationship between ROI pixel coordinates and those within the deformed image, taking ROI pixel coordinates as input and corresponding displacement as output. The reference image undergoes warping based on the neural network’s displacement field output. To address experimental illumination variations, the framework incorporates a gray-scale linear mapping module that assigns two learnable parameters to each warped reference image pixel for adaptive correction of illumination-induced gray-scale changes. A relative shape loss function, resistant to gray-scale linear changes, is implemented alongside the absolute gray-scale loss to establish illumination-robust self-supervised information. The framework’s learnable network parameters undergo continuous iterative updates through the back-propagation algorithm to obtain optimal ROI displacement field that minimizes the loss function.Results and DiscussionsThe proposed method undergoes experimental validation across multiple scenarios, and its performance is compared with traditional methods and other deep learning-based approaches. The method demonstrates superior performance in simple displacement field solutions, exhibiting exceptional smoothness and close alignment with ground truth (Figs. 3?6). The mean solution error remains below 0.008 pixel for rigid body motion displacement field and under 0.013 pixel for linear displacement field. In complex star-shaped displacement field solutions, the method maintains high accuracy across both high-frequency and low-frequency regions (Figs. 7?9), achieving a mean solution error of 0.0610 pixel. For rotation displacement field solutions, the method yields optimal results (Figs. 10 and 11), maintaining mean solution error below 1.9 pixel in fields with maximum displacement exceeding 40 pixel. Illumination robustness tests demonstrate minimal sensitivity to illumination changes (Fig. 13) and lowest solution error (Table 3), attributable to the integrated gray-scale linear mapping module and relative shape loss function. Computational efficiency analysis reveals enhanced performance compared to traditional methods, avoiding time-intensive pre-training processes and demonstrating strong application potential.ConclusionsThis paper presents a self-supervised learning-based method for speckle image displacement field solutions. Unlike conventional deep learning-based approaches requiring extensive pre-training datasets, this method enables direct displacement field solution implementation on analyzed speckle images through a self-supervised learning framework incorporating gray-scale linear mapping and relative shape loss functions. Experimental validation demonstrates the method’s superior performance in accuracy, illumination robustness, and computational efficiency. The approach effectively addresses traditional method limitations regarding manual parameter setting and rotation displacement field solutions while eliminating dependence on extensive pre-training datasets required by deep learning-based methods, offering novel perspectives for DIC technique implementation.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1610003 (2025)
  • Wenlong Li, Yanbo An, Duoyao Wang, Depeng Kong, Huangrong Xu, Zhen Liu, Yuan Yuan, Zhengquan He, Pengfei Wang, Yantao Xu, and Haitao Guo

    ObjectiveContemporary domestic radiation-resistant imaging systems predominantly utilize a reflective structure that combines ray shielding with reflective imaging to achieve radiation resistance. While light path reflection imaging prevents direct radiation damage to electronic components within the shielding shell, this approach necessitates substantial space for mirrors, resulting in systems with large size and complex structures. Furthermore, the utilization of high-density shielding materials, such as lead, substantially increases system weight, creating operational limitations and restricting application versatility. To address these challenges, this study presents a radiation-resistant imaging system based on optical fiber image transmission bundles, incorporating radiation-resistant optical glass technology and flexible fiber-optic transmission configuration. The design connects a front-end radiation-resistant objective optical system in the nuclear radiation zone with a rear-end coupling optical system and complementary metal oxide semiconductor (CMOS) image sensor in the non-radiation zone. This configuration facilitates physical isolation, effectively shielding high-energy rays and achieving superior radiation resistance. The approach resolves existing limitations of domestic radiation-resistant imaging systems regarding size, weight, radiation dose rate, and total dose, offering significant potential for nuclear energy utilization, nuclear power development, and equipment localization.MethodsWe designed a radiation-resistant objective lens using JGS1 quartz and ZF706 radiation-resistant optical glass, and prepared a quartz fiber image bundle with a pixel count of more than 500000 using the filament alignment and laminating technique. Combining those with a coupled lens, a CMOS camera, and the control and display system, a high radiation resistance optical imaging system was installed. The cobalt source was used to irradiate the radiation-resistant optical imaging system, with an irradiation dose rate of 1.0×104 Gy/h. During the irradiation process, the resolution performance of the optical imaging system was tested using the ISO 12233 standard resolution test card. The resolution test was performed when the imaging system irradiates to the corresponding dose points: 0, 2.50×105, 5.00×105, 7.50×105, 1.00×106 Gy and 1.21×106 Gy. To comprehensively analyze the primary factors contributing to the decreased transmittance of the radiation-resistant imaging system post-irradiation, the transmittance at various dose points of the matching objective lens was evaluated using a spectrometer. Due to the complexity of disassembling the image transmission bundle during irradiation, its transmittance was measured only before and after irradiation.Results and DiscussionsDuring the irradiation process, the radiation-resistant optical imaging system is tested using the ISO 12233 standard resolution test card to evaluate its imaging performance. The resolution test is conducted online at an irradiation dose rate of 1.0×104 Gy/h. The resolution test is carried out at the corresponding dose points: 0, 2.50×105, 5.00×105, 7.50×105, 1.00×106 Gy and 1.21×106 Gy. As shown in Fig. 11, the resolution test results indicate that the system maintained a resolution better than 25 lp/mm. No significant degradation in image resolution is observed, apart from slight browning during irradiation. The browning effect can be attributed to two main factors. 1) The target plate itself undergoes browningdue to irradiation exposure. 2) After irradiation, the transmittance at the shorter wavelengths decreases more significantly than that at the longer wavelengths, leading to severe absorption during short-wave transmission. To thoroughly analyze the primary factors affecting the decrease in transmittance of the radiation-resistant imaging system, the transmittance of its spectral lens at different dose points is investigated. Figure 12 shows the transmittance test curve of the accompanying objective lens of the high-radiation-resistant imaging system under different cumulative irradiation doses. The transmission first declines and then stabilizes during irradiation. After the irradiation dose reached 1.21×106 Gy, the average transmittance (between 400 nm to 760 nm with a sampling interval of 0.58 nm) decreased by 6.26% compared to the value before irradiation. The decrease of transmittance at the longer wavelenghts is relatively minor, registering 3.34%@600 nm, compared to the pronounced decrease at the shorter wavelengths, i.e. 18.23%@450 nm. This finding further supports the browning analysis discussed earlier. Due to the irradiation test conditions, where the maximum dose rate is 1.0×104 Gy/h and the test duration is relatively long, the experiment stopped when the total dose reached 1.21×106 Gy. However, this threshold in fact do not reach the limit of the self-developed radiation-resistant glass ZF706. The radiation resistance limit will be further studied according to actual needs in the future. During irradiation, because the image transmission bundle is difficult to disassemble, only the transmittance before and after irradiation is tested, and the results are shown in Fig. 13. In the wavelength range of 400 nm to 550 nm, the transmittance of the image transmission bundle before and after irradiation is less than 1%. Additionally, for visible light with wavelengths greater than 550 nm, no significant attenuation in transmittance is observed.ConclusionsIn response to the limitations of current high-dose radiation-resistant imaging systems, including their substantial size, weight, and limited continuous operation capability, this study presents an innovative high-radiation-resistant optical imaging system utilizing optical fiber bundles. The quartz fiber image transmission bundles demonstrate excellent bending flexibility and radiation resistance characteristics. The system exhibits enhanced spatial control flexibility, while its design enables miniaturization and weight reduction, making it adaptable for radiation-resistant video surveillance across diverse applications. This radiation-resistant imaging system addresses previous challenges related to equipment size, weight, and structural complexity that typically arise from increased shielding thickness requirements. Experimental results demonstrate the superior radiation resistance performance of the high-radiation-resistant optical imaging system based on quartz fiber image transmission technology. The system maintained continuous operation for 121 h under an irradiation dose rate of 1.00×104 Gy/h, achieving a cumulative total radiation resistance dose exceeding 1.21×106 Gy. Post-irradiation analysis revealed only minor browning of the objective lens, with its average transmittance (400 nm?750 nm) decreasing by 6.26% compared to pre-irradiation values. The system maintained consistent resolution before and after irradiation, achieving better than 25 lp/mm.

    Aug. 08, 2025
  • Vol. 45 Issue 16 1611001 (2025)
  • Ziyue Guo, Ying Jin, Sunyong Zhu, Quanying Wu, and Guohai Situ

    ObjectiveTraditional computational tomography methods entail substantial computational costs for 3D reconstruction and time series prediction of flames in combustion fields, making real-time prediction of projection information challenging. The advancement of deep learning has facilitated 3D reconstruction, enhanced data processing efficiency, reduced experimental costs, and enabled real-time prediction capabilities. The computational efficiency and accuracy of combustion length prediction can be enhanced through the development of more efficient networks, optimization of model parameters, and improvements in computing capabilities.MethodsThe GAST-Net employs convolutional neural networks to capture complex spatial structural features, integrating gated recurrent units (GRUs) to address long-distance dependencies. The time processing module is integrated into the network’s skip connections to enhance temporal information modeling through spatiotemporal feature fusion. These skip connections mitigate the gradient vanishing problem during deep network training. Additionally, a hybrid attention mechanism in the time processing module separately processes critical temporal and spatial information at each feature layer, achieving comprehensive information fusion and utilization while significantly improving prediction and reconstruction accuracy. A projection acquisition device based on computed tomography of chemistry was constructed, and the three-dimensional structure distribution of CH* radicals in the combustion field was reconstructed using traditional methods as a ground truth reference for neural network training.Results and DiscussionsInitial validation of GAST-Net demonstrated its capability to predict reconstruction structure from historical multi-directional projection data within 7 ms (Fig. 4), maintaining high prediction reconstruction accuracy across multiple time intervals (Fig. 5). Further analysis comparing predicted 3D field structure data with 2D slices of real reconstruction data confirmed accurate prediction of internal combustion field structures. Quantitative evaluation through four quality assessment functions revealed that prediction accuracy decreases with increasing time steps while maintaining high performance levels. The model demonstrated effectiveness in inter-frame prediction of 3D fields (Fig. 8) and maintained robust performance when multiple instantaneous 3D data were inserted between consecutive time series (Fig. 9). Both future time prediction and inter-frame prediction scenarios achieved structural similarity exceeding 0.93, correlation coefficients above 0.9, root mean square error below 0.02, and peak signal-to-noise ratio not less than 34 (Figs. 7 and 10). Additionally, ablation experiments examined the impact of attention module placement and various network parameters on reconstruction accuracy.ConclusionsThis study presents a novel emission spectrotomography reconstruction prediction method incorporating gated recirculating units and attention mechanisms. The approach utilizes traditional chemiluminescence computational tomography algorithm reconstructions of CH* radical three-dimensional distributions as reference data for deep learning validation. GAST-Net leverages gated recurrent units for managing long-distance dependencies and attention mechanisms for importance weight adjustment, enabling three-dimensional reconstruction and future prediction of combustion fields from historical multi-directional two-dimensional projections. This methodology eliminates the need for complex acquisition devices or partial differential equation solutions. Experimental results demonstrate rapid and accurate prediction of future combustion evolution based on continuous historical multi-directional projection data. Furthermore, the model successfully addresses interpolation between continuous time series, accurately predicting multi-frame 3D structures using before-and-after multi-directional projections. These findings indicate significant potential for addressing time series challenges and suggest new approaches for solving long-distance dependency problems. The network’s prediction accuracy and applicability can be further enhanced through expanded training data.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1611002 (2025)
  • Xiaodong Zhang, Zhiguo Han, Lin Zhao, Jiahuan Zhang, Xiaoqing Xu, Suoyin Li, and Aihua Wu

    ObjectiveDriven continuously by Moore’s Law, semiconductor manufacturing technology is rapidly evolving along a dual-track technological path that combines “feature size scaling” and “three-dimensional heterogeneous integration”. As the core building blocks of integrated circuits, functional thin films’ interfacial properties and mechanical behaviors are directly related to the reliability and performance of devices. Currently, mainstream thin film deposition technologies represented by physical vapor deposition (PVD), chemical vapor deposition (CVD), and atomic layer deposition (ALD) have been able to achieve the preparation of thin films with nanoscale precision. However, in thin film systems formed under thermodynamic non-equilibrium states, there is a common problem of the dynamic evolution of intrinsic stress. This phenomenon stems from the intrinsic stress generated by lattice mismatch during the thin film deposition process, which is superimposed with the thermal stress caused by differences in thermal expansion coefficients of heterogeneous materials in the service environment, as well as combined mechanical loads such as tension, compression, and bending. The coupled effect of multiple stresses is highly likely to induce serious consequences such as interfacial delamination, lattice distortion, and even structural failure of devices. Therefore, conducting research on the generation mechanism of wafer thin film stress and establishing a unified curvature radius-stress characterization model for wafer thin films is of great engineering value for optimizing the process flow and improving device yield.MethodsIn order to monitor the evolution law of the stress during the growth process of the wafer thin film in real time, we conduct research on the stress testing theory of the wafer thin film. Firstly, through multi-physical field coupling modeling and simulation, the generation mechanism of the thin film stress and its spatial gradient distribution law are revealed. Secondly, a curvature radius-stress characterization model based on the stress testing theory and the least-squares method is proposed. This model optimizes the calculation process of the curvature radius by using the method of least-squares fitting of a circle, improving the repeatability and accuracy of the curvature radius. Finally, based on this model, a contact stress testing system with multi-parameter coordination is established. By precisely controlling key parameters such as elastic modulus, Poisson’s ratio, substrate curvature radius, thin film thickness, and substrate thickness, a quantitative conversion mechanism between curvature and stress is established.Results and DiscussionsThe static thermal load simulation reveals the phenomenon of stress concentration at the SiO?/Si interface. Results of the simulation show that the thermal displacement field exhibited significant axial gradient characteristics, and the displacement distribution presents a concentric circular ring topological structure. The maximum stress value is controlled at approximately 300 MPa. It is worth noting that material differences will cause nonlinear changes in temperature. In addition, extreme value regions of the stress are mainly distributed at the edge of the SiO? thin film, which is consistent with the gradient distribution law of the Stoney’s formula (Fig. 1). Then, compared with the traditional characterization model, a curvature radius-stress characterization model based on the stress testing theory and the least-squares method is proposed, and a contact stress testing system is established (Fig. 4). Two characterization models are used to conduct stress analysis on the data of stress calibrators with stresses of 40, 150, and 300 MPa respectively. These results show that when the stress is 40 MPa, the difference between these two characterization models is small (n-order polynomial method: 0.18?0.42 MPa; least-squares method: 0.19?0.40 MPa). However, when the stress increases to 150 MPa and above, the convergence of the least-squares method is significantly better than that of the n-order polynomial method (for example, when the stress is 300 MPa, the repeatability of the least-squares method is 0.16 MPa, which is 46.7% lower than 0.30 MPa of the n-order polynomial method) (Table 2). These research results indicate that the characterization model based on the least-squares method has stronger robustness in high-stress fields (≥150 MPa), and its optimized measurement repeatability can meet the accuracy requirements for thin film stress monitoring in semiconductor manufacturing processes.ConclusionsThrough finite element analysis method for modeling and simulation, the generation mechanism of the wafer thin film stress and its spatial gradient distribution law are revealed. A curvature radius-stress characterization model based on the stress testing theory and the least-squares method is proposed. This model optimizes the calculation process of the curvature radius by using the method of least-squares circle fitting, improving the repeatability and accuracy of the wafer thin film stress. The experimental data show that the test repeatability of this model is better than that of the traditional polynomial method. In addition, based on this model, a contact stress testing system with multi-parameter coordination is established. By precisely controlling key parameters such as elastic modulus, Poisson’s ratio, substrate curvature radius, thin film thickness, and substrate thickness, a quantitative conversion mechanism between curvature and stress is established to meet the accuracy requirements for thin film stress monitoring in semiconductor manufacturing processes.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1612001 (2025)
  • Haokun Zou, Jiying Chang, Kai Chen, Jining Li, Kai Zhong, Yuye Wang, Degang Xu, and Jianquan Yao

    ObjectiveSingle-photon lidar has demonstrated significant capabilities in long-range, high-sensitivity detection under conditions of low-pulse-energy laser and small-aperture optical systems. However, certain practical applications face challenges in detecting distant targets within high background noise environments, limiting operational effectiveness. This paper introduces an asynchronous correlation encoding detection method based on the gate-mode operational characteristics of single-photon detectors to enhance the ultimate detection range and improve detection reliability in strong background noise environments. The method executes target detection through alternating operation of the single-photon detector between two states (laser pulse-synchronized detection and asynchronous detection), followed by acquisition and subtraction of raw echo datasets from both states, culminating in cross-correlation operations between the resultant histogram data and correlation codes.MethodsThis study presents an asynchronous correlation encoding detection method. Monte Carlo simulations were utilized to compare detection success probabilities between this method and the pulse accumulation detection method under varying relative interference intensities, while analyzing the influence patterns of coding sequences on detection performance. The algorithm architecture was implemented on a field-programmable gate array (FPGA) platform for experimental validation, establishing an asynchronous correlation encoding single-photon lidar system. Outdoor ranging experiments demonstrated enhanced detection performance under strong background noise environments through comparative analysis of lidar measurements across different detection methods and noise levels, validating the method’s effectiveness in improving single-photon lidar ranging systems.Results and DiscussionsSimulation studies reveal that when the autocorrelation of the encoded sequence approaches similarity, the detection probability of the asynchronous correlation encoding method decreases with increasing code length. This decrease stems from reduced energy allocation to individual echo pulses corresponding to each code element in longer code lengths, adversely affecting system detection performance. However, at fixed code length, the system’s detection performance shows positive correlation with sequence autocorrelation. Under a relative interference intensity of 17.5 dB, the asynchronous correlation encoding method with a code length of 2 achieves a detection success probability of 63.1%, surpassing conventional pulse accumulation single-photon lidar by 22.9 percentage points [Fig. 5(a)], demonstrating superior detection performance under strong background noise interference. An asynchronous correlation single-photon lidar system was constructed for ranging experiments on an outdoor target. Multiple ranging trials were conducted on a building exterior wall at 10 m under varying background noise intensities with an integration time of 0.5 s. At a background noise photon count rate of 1600 s-1, both detection methods successfully identified the target [Fig. 9(a), Fig. 10(a), Fig. 10(b)]. At 198000 s-1, while the pulse accumulation method’s signal peak became indistinguishable from noise [Fig. 10(c)], the asynchronous correlation encoding method maintained a distinguishable signal peak [Fig. 10(d)]. In ten repeated trials, the pulse accumulation method achieved no successful detections [Fig. 9(b)], while the asynchronous correlation encoding method achieved five successful detections [Fig. 9(c)]. At 135000 s-1 background noise photon count rate, with the system repositioned 0.3 m back, the pulse accumulation method achieved six successful detections [Fig. 11(a)], while the asynchronous correlation encoding method achieved eight successful detections [Fig. 11(b)]. These results demonstrate the superior noise suppression capabilities and higher detection probability of the asynchronous correlation encoding detection method under strong background noise interference.ConclusionsThis study presents an asynchronous correlation encoding detection method based on single-photon detectors’ gated-mode operation characteristics and demonstrates the construction of a corresponding lidar system. The system exhibits enhanced detection probability under strong background noise interference. Monte Carlo simulations analyzed detection probability under various methods and conditions, particularly examining code length and autocorrelation effects on system performance. Research indicates optimal detection performance occurs with a code length of 2, where the single-photon detector operates in two states. The method demonstrates reliable target detection under strong background noise interference compared to conventional pulse accumulation of single-photon lidar. Experimental evaluations confirm excellent noise suppression capabilities and higher detection success probability in intense background noise environments compared to traditional methods. This research provides a practical approach for enhancing single-photon lidar system reliability under strong background noise interference, establishing groundwork for applications in challenging environments with intense background noise and long-distance requirements.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1612002 (2025)
  • Mingtao Jiang, Wenjie Xu, and Wei Chen

    ObjectiveTurbulence can be widely observed in natural environments and engineering systems, where accurate measurements of velocity and scalar fields are essential for understanding flow mechanisms and optimizing engineering designs. Traditional particle image velocimetry (PIV) has become a standard experimental technique in turbulence research. However, its reliance on tracer particles poses significant limitations in enclosed and high-temperature conditions or scenarios requiring multi-field synchronous observation. Specifically, tracer particles may interfere with scalar field measurements, thereby hindering simultaneous spatiotemporal acquisition. Scalar image velocimetry (SIV), based on passive scalar transport, provides a non-intrusive alternative that avoids these issues and offers greater adaptability. Nevertheless, current SIV methods are constrained in high-resolution velocity reconstruction and nonlinear structure perception due to algorithmic limitations. Recent advancements of optical flow neural networks in computer vision, with their powerful feature extraction capabilities, offer new possibilities for flow field estimation. However, most existing networks are built on rigid motion assumptions, making them unsuitable for the continuous deformation characteristics of fluid flows. To resolve this limitation, we proposed a physics-adapted SIV-RAFT (recurrent all-pairs field transforms) neural network architecture that integrated deep learning with fluid dynamic constraints. The model enabled accurate, non-invasive, and synchronized measurement of velocity and concentration fields in complex turbulent flows, demonstrating both theoretical significance and practical engineering applicability.MethodsThe proposed SIV-RAFT algorithm was built upon the strengths of the original RAFT framework while introducing systematic architectural enhancements tailored to the deformable fluid motion dynamics. In the feature extraction stage, deformable convolutional networks (DCNs) were employed to replace standard convolutional layers. By dynamically adjusting convolutional kernel sampling positions, the network could adaptively capture geometric deformation features induced by fluid motion. To enhance the optical flow iterative refinement process, an iterative attention feature fusion (iAFF) mechanism was integrated. This mechanism constructed a cross-channel and spatial dual-dimensional attention weight matrix, enabling the network to focus adaptively on multi-scale turbulent structures. Implemented within the gated recurrent unit (GRU) framework, the iAFF mechanism introduced a feature recalibration pathway that dynamically adjusted the contribution of different spatiotemporal scale features during each iteration, thereby improving complex turbulent flow representation. For model training and performance evaluation, a scalar image dataset tailored for the SIV task was constructed. First, the reliability of large eddy simulation (LES) data as a substitute for experimental measurements in jet-related problems was verified through physical comparison experiments. Subsequently, a series of inclined negatively buoyant jet scenarios under various conditions were simulated to generate scalar image sequences with realistic physical evolution, serving as training data. For model validation, both LES results and publicly available direct numerical simulation (DNS) turbulence databases were utilized to conduct rigorous quantitative assessments of the estimation accuracy and robustness of the SIV-RAFT model. These comprehensive evaluations confirmed the proposed model’s adaptability and effectiveness in complex flow velocity reconstruction tasks.Results and DiscussionsTo comprehensively assess the prediction capabilities of the proposed SIV-RAFT algorithm, extensive evaluations have been conducted on both the inclined jet dataset and the DNS turbulence dataset. Figure 8 shows velocity field prediction errors of the proposed SIV-RAFT algorithm, the PWC-Net algorithm, and the Horn?Schunck (HS) method. Quantitative evaluation reveals that SIV-RAFT achieves a maximum error magnitude of less than 0.29 pixel in 95% of the observed region, outperforming the PWC-Net algorithm (0.57 pixel) and the HS method (1.42 pixel). In particular, in high-velocity jet regions, the SIV-RAFT algorithm demonstrates significantly lower local errors than the other two methods, as shown in Fig. 8(a). These comparative results indicate that SIV-RAFT has a distinct advantage in maintaining the spatial consistency of the velocity vector field and capturing local features. Figure 11 presents the velocity profiles across sections perpendicular to the jet centerline at different streamwise locations, namely 20D and 30D downstream from the nozzle, corresponding to Figs. 11(a) and (b), respectively. The results reveal that SIV-RAFT achieves the best overall performance in jet velocity prediction. The root-mean-square errors (RMSEs) of velocity at the 20D and 30D sections are 2.91% and 1.41%, respectively. Furthermore, the predicted velocity profiles exhibit close alignment with the experimental data in both shape and magnitude, confirming that multi-scale feature extraction and spatiotemporal correlation constraints effectively enhance prediction robustness in complex shear flows. Figure 12 shows the instantaneous velocity magnitude at t=30, where Figs. 12(a)?(d) show the ground-truth DNS velocity magnitude and the predictions from the HS, PWC-Net, and SIV-RAFT algorithms, respectively. The HS method exhibits substantial reconstruction errors, with an RMSE approximately 37.14% higher than that of SIV-RAFT. Compared to the PWC-Net, SIV-RAFT reduces the RMSE of the velocity field by 14.10%. Figure 13 illustrates the vorticity fields from the ground-truth value of DNS and those of the other three algorithms. The results show that SIV-RAFT captures fine-scale vortex structures more accurately than PWC-Net and the HS methods, demonstrating enhanced capability in resolving intricate turbulent flow features.ConclusionsIn the present study, a deep learning-based SIV method, SIV-RAFT, is proposed. Based on the RAFT optical flow estimation framework from computer vision, the network is systematically optimized: DCNs are introduced in the feature extraction stage to enhance adaptability to large-scale fluid deformations, while an iAFF mechanism is integrated into the flow update module to improve the ability to capture fine-scale vortical structures. To support model training, a multiphysics benchmark dataset is constructed by performing multi-condition 3D inclined density jet simulations using the open-source CFD software OpenFOAM. The model’s performance is evaluated on both a jet flow dataset and a DNS turbulence dataset. Comparative results show that on the jet dataset, the proposed method reduces velocity magnitude prediction errors by 21.32% and 43.26% compared to the PWC-Net algorithm and the traditional HS method, respectively. On the DNS dataset, corresponding error reductions are 23.02% and 38.99%. Quantitative analysis demonstrates that the SIV-RAFT model effectively captures multi-scale turbulent structures from scalar concentration fields and achieves an accurate velocity field. This study provides a novel and data-driven solution for simultaneous concentration and velocity measurement in complex flow environments, demonstrating potential for deep learning applications in experimental fluid dynamics.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1612003 (2025)
  • Douming Hu, Libin Yao, Nan Chen, Jiqing Zhang, Yanzhen Liu, Changkun Cui, Kesong Chen, and Faming Li

    ObjectiveLarge-format mid-wave infrared (MWIR) focal plane arrays (FPAs) with a 5 μm pitch meet demanding size, weight, power, price, and performance (SWaP3) targets. However, at multi-megapixel scale, the readout integrated circuit (ROIC) suffers from reduced charge capacity, tight power budgets, high readout speed, and stringent linearity demands. We present a 5 μm digital ROIC (DROIC) of 1920×1536 that integrated direct-injection (DI) pixels, a column-level single-slope analog-to-digital converter (SS ADC), and a ramp-matching compensation network to overcome these limits.MethodsEach pixel adopts a DI structure operated in global-shutter integrate-then-read (ITR) mode. A centrally-symmetric 67 fF MOS capacitor inside each 5 μm cell yields a 0.75 Me- full-well; interlace mode lets odd/even rows time-share the same node, doubling capacity to 1.5 Me- and adding about 3 dB signal-to-noise ratio (SNR). Column signals are buffered by in-pixel source followers. Their body-effect non-linearity is cancelled by a ramp-matching network that let the SS ADC comparator see identical distortions on both inputs. One local-ramp branch now serves 16 ADCs, shrinking source-follower branches 16× and cutting static power. Additional power savings come from minimal bias currents, a 6-bit TSPC+8-bit DFF hybrid counter, and a gated clock tree. Noise is restrained by kick-back reduction via the compensated ramp and a pre-amplified comparator, as well as a large ramp capacitor plus tailored bias to suppress thermal and flicker components.Results and DiscussionsThe DROIC is fabricated in a standard 0.18 μm CMOS process. Measurements show a maximum nonlinearity below 0.1%, a readout noise of 178 e? (RMS), and a total power consumption of 249 mW at a 30 Hz frame rate. After hybridization with an HgCdTe MWIR detector, the FPA attains a noise-equivalent temperature difference (NETD) of 38.2 mK in normal mode; owing to the doubled charge swing and the square-root dependence of noise on charge, NETD further improves to 32.2 mK in interlace mode.ConclusionsThe proposed 5 μm-pitch MWIR DROIC simultaneously balances charge capacity, linearity, and power, and it constitutes the first Chinese large-format DROIC at this pixel size. By integrating DI pixels, interlace charge expansion, a low-power column-level SS ADC, and a ramp-matching compensation network, the design delivers high-performance digital readout for small-pixel FPAs and offers a viable solution for future high-resolution and low-SWaP3 infrared imaging systems.

    Aug. 11, 2025
  • Vol. 45 Issue 16 1636001 (2025)
  • Da Xu, Gaofei Sun, and Xianzhu Liu

    ObjectiveTraditional star sensors and emerging attitude navigation sensors face increasing limitations in calibration techniques. Current methodologies inadequately address complex space environment simulation requirements, high-precision calibration standards, and rapid testing needs for large-scale constellations (e.g., the Qianfan Constellation Project). This research presents a region-partitioned star-point independent modulation stellar simulation system for generating authentic star maps with precise stellar colors. The system enables independent spectral and energy modulation for individual star points, offering an alternative to conventional star observation tests and advanced calibration capabilities for next-generation navigation devices.MethodsThe research began with an analysis of the minimum modulation rate necessary for stellar spectral modulation, followed by the derivation of theoretical modulation rates meeting spectral simulation accuracy requirements. Based on predetermined digital micromirror device (DMD) partitioning criteria, the DMD was segmented accordingly. The system architecture incorporated a cylindrical collimating beam expansion system and a three-region independent modulation optical system. Image quality assessment confirmed the production of flat, curvature-free spot shapes and achieved spectral resolution exceeding 5 nm across the spectral plane, validating the system's capacity for accurate spectral modulation and simulation.Results and DiscussionsPerformance tests utilized an established experimental platform. Initial modulation capability assessment revealed a maximum full width at half maximum (FWHM) of 4.61 nm for monochromatic light output, satisfying the 5 nm spectral resolution requirement. Spectral simulation accuracy test across three regions employed target values of 3000 K, 5000 K, 7000 K, and 9000 K. Results indicated a maximum spectral simulation error of 7.2% across regions, with inter-region consistency variations below 1.8%. Independent spectral modulation verification assigned sub-regions 1, 2, and 3 to simulate 3000 K, 7000 K, and 9000 K, respectively, yielding spectral simulation errors of 4.7%, 4.8%, and 4.3%. These findings demonstrate the system’s effectiveness in achieving precise, independent spectral control for multiple star simulations.ConclusionsThis research establishes a region-partitioned modulation methodology for star map simulation, facilitating precise spectral control of individual star points. Experimental validation confirmed high accuracy (<7.5% error) and consistency (<1.8% deviation) across a 3000?9000 K range. The system demonstrates capability in simulating three stars with arbitrary color temperatures, indicating its potential as a replacement for traditional star observation tests. This technology advances calibration solutions for future navigation systems while reducing calibration complexity and testing duration, supporting efficient star sensor mass production.

    Aug. 13, 2025
  • Vol. 45 Issue 16 1622001 (2025)
  • Xiaobing Liu, Xingxiang Zhang, Tianjiao Fu, and Duo Wang

    ObjectiveThis research addresses optical system optimization challenges where local optimization algorithms with random starting points frequently converge to suboptimal local solutions, while heuristic global optimization algorithms demonstrate low efficiency in solution space exploration. To effectively utilize parallel computing capabilities and minimize computational resources spent on invalid solutions, this study implements a monitoring system that analyzes configuration differences between systems based on selective guidance of global optimal information. When global information guides systems from poor local minima toward improved regions, system aggregation inevitably occurs. This aggregation results in uneven distribution within the solution space, potentially overlooking valid configurations that might yield superior results.MethodsA diversity-enhanced selective guided gradient descent (DESG-GD) algorithm is proposed. Building upon the global information selective guided (GISG) framework, DESG incorporates a solution space density assessment indicator. This indicator quantifies distribution density through calculating the minimum Euclidean distance between system parameters and implements a dual-guidance point strategy: directing low-performing systems toward the loss-optimal point, while guiding systems with acceptable performance but high distribution density toward the diversity-optimal point. By separating systems with similar configurations, the algorithm explores more configurations within the solution space, enhancing system performance and generating effective, diverse results.Results and DiscussionsValidation through 4-lens and double Gaussian lenses demonstrates that initial structures optimized by DESG-GD achieve or exceed patent-level performance. In complex double Gaussian systems, the proposed method demonstrates more than 10% improvement in spot diagram size, efficiency, and diversity compared to existing advanced methods. DESG-GD substantially enhances the exploration of diverse optical system results while maintaining solution quality, offering novel approaches for reference-free optical design.ConclusionsThis study demonstrates how DESG-GD selectively guides high-performing but densely distributed solutions toward sparser regions of the solution space, facilitating exploration of previously unexplored areas. This approach enhances both diversity and performance in ab initio lens design. The root mean square (RMS) spot sizes for both the 4-lens system and the double Gaussian match or exceed patent performance, demonstrating its effectiveness in obtaining high-quality starting points for optical system optimization without references. The proposed DESG-GD, which enhances local optimization through gradient descent with system interaction strategies, presents significant opportunities for further development and research.

    Aug. 07, 2025
  • Vol. 45 Issue 16 1622002 (2025)
  • Chenxiao Li, Yuanming Zhao, Yu Zhang, Bo Zhang, Mingxu Piao, Yiqiao Yuan, and Junjie Gao

    ObjectiveVirtual reality head-mounted displays (VR-HMDs) have gained significant adoption across various sectors including education, healthcare, entertainment, and manufacturing. Nevertheless, lateral chromatic aberration (LCA) remains a substantial challenge, which is significantly exacerbated by pupil misalignment during head movements. This issue compromises image quality and diminishes user immersion. Traditional approaches, such as real-time image preprocessing or eye-tracking hardware, often introduce additional latency, increasing device complexity, or additional costs. To overcome these limitations and address the increasing demand for high-quality VR experiences, we propose an innovative refractive/reflective/diffractive hybrid optical system. The system features a compact configuration (total length <22 mm), extensive field of view (FOV of 96°), and robust pupil shift tolerance within a 6 mm×6 mm rectangular range. By integrating a single-layer diffractive optical element (SLDOE), the system effectively corrects chromatic aberrations while aligning diffraction efficiency with human visual sensitivity. This approach achieves high-resolution imaging (2304 pixel×2160 pixel) while maintaining a lightweight design, enhancing user comfort and immersion.MethodsA polarization-folded optical structure is adopted to reduce system volume while expanding FOV. This structure utilizes polarization elements including beam splitters, reflective polarizers, and quarter-wave plates to achieve optical path folding. The SLDOE, fabricated on a cycloolefin copolymer (COC) lens, utilizes its negative dispersion properties to counteract chromatic aberrations. Unlike conventional bandwidth-integrated average diffraction efficiency (BIADE) optimization methods, we develop a human eye sensitivity-diffraction efficiency (ESDE) model. This model incorporates the human eye’s spectral sensitivity curve, peaking at 0.555 μm, as a weighting factor for SLDOE optimization. By prioritizing wavelengths more perceptible to human vision, the model ensures higher diffraction efficiency in the visible spectrum range. The design process begins with constructing an initial pancake-type folded optical path. Then the optimization focuses on simulating six critical pupil positions (A?F) within the 6 mm×6 mm range to comprehensively evaluate axial and lateral chromatic aberrations, as well as the modulation transfer function (MTF). The SLDOE placement on the first lens surface minimizes stray light and simplifies fabrication processes. Finally, the SLDOE’s microstructural height is meticulously adjusted to maximize the bandwidth-integrated average ESDE (BIAESDE), ensuring compatibility with injection-molding manufacturing techniques.Results and DiscussionsThe hybrid system demonstrates remarkable improvements in chromatic aberration correction and overall imaging performance. Axial chromatic aberration decreases by 49.104%, with the maximum value reducing from 124.014 μm (pre-optimization) to 65.371 μm. Lateral chromatic aberration at the worst-case pupil position (E) decreases by 35.88% to 11.705 μm, falling below the 12 μm pixel size threshold, thus rendering color fringing artifacts imperceptible. At the Nyquist frequency (42 lp/mm), MTF values exceed 0.2 across all six pupil positions, with the aligned position (A) exceeding 0.5. This performance ensures compatibility with high-resolution displays and maintains image clarity even during pupil shifts. The ESDE-optimized SLDOE achieves 90.31% average diffraction efficiency under application conditions, representing a 3.5% improvement over traditional BIADE methods. Structural validation confirms the SLDOE’s manufacturability, with a minimum zone width of 76.33 μm achievable through injection molding. The compact design, incorporating an aperture less than 60 mm and lightweight plastic lenses, further enhance wearability and user comfort.ConclusionsWe present an innovative hybrid optical system for VR-HMDs that effectively addresses pupil shift-induced chromatic aberrations without requiring resource-intensive correction methods. The primary contributions include the integration of SLDOE to suppress both axial and lateral chromatic aberrations while maintaining a compact form factor, and the development of ESDE optimization model. This model aligns diffraction efficiency with human visual sensitivity, enhancing optical performance and user comfort. The system demonstrates robust imaging quality across a 6 mm×6 mm pupil shift range, providing an economical and lightweight solution for next-generation VR devices. By eliminating the need for additional hardware or computationally intensive algorithms, this design significantly reduces device complexity and manufacturing costs. Future work will focus on extending the FOV beyond 120° and exploring the integration of metasurfaces for enhanced aberration correction, further advancing the capabilities of VR optical systems.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1622003 (2025)
  • Hengwei Qian, Wenhao Miao, Jiazhen Zhu, Haixing Liang, Xiaobo Ma, Songfeng Kou, Zhongcheng Liang, and Rui Zhao

    ObjectiveDielectric elastomer (DE) is an emerging electroactive polymer material capable of significant deformation under an external electric field and is widely used in the fields of soft robotics and tactile sensors. Based on its electrically induced deformation properties, researchers have developed dielectric elastomer actuators (DEAs), which consist of a DE film coated with flexible electrodes on both surfaces. When a voltage is applied to the flexible electrodes, the electric field induces lateral expansion and thickness contraction of the film. DEAs offer advantages such as large strain, fast response, and low power consumption, making them well-suited for applications in liquid tunable lenses. Although the existing DE-based liquid tunable lenses can achieve a certain range of focus adjustment, the development of DE-driven liquid lenses with simple structure, good stability, and strong zooming ability still faces many challenges. In this paper, a dual-cavity dielectric elastomer zoom lens structure based on the principle of electrodeformation of dielectric elastomer is designed, featuring a compact structure, good stability, large aperture, strong zoom ability and fast response speed. This work provides new ideas and methods for the design and optimization of dielectric elastomer liquid lenses.MethodsIn this paper, a dual-cavity dielectric elastomer tunable lens structure is designed, which mainly consists of upper and lower cavities, transparent conductive liquid and DE films. The DE film between upper and lower cavities acts as an active film to realize the electrically induced deformation. The DE film on the top of the upper cavity acts as a passive film, with its curvature directly determining the focal length of the lens, and the DE film on the side of the lower cavity also acts as a passive film to equalize the pressure inside and outside the cavity. Upper and lower cavities are filled with the same kind of transparent conductive liquid, which serves as the driving electrode of the DE film. The zoom principle of the lens is theoretically analyzed, and a lens model is built in COMSOL to study the influence of the liquid volume parameters in upper and lower cavities on the lens’s focal length tuning performance. Based on the simulation results, the fabrication of the tunable lens is completed, and focal lengths of the liquid lens under different voltages are measured by a focal length meter and compared with simulation results. The focusing ability, imaging quality, focal spot size, and light intensity distribution of the liquid lens under different driving voltages are tested using a charge-coupled device (CCD). Additionally, an optical system is set up to evaluate the dynamic response performance of the lens.Results and DiscussionsTheoretical analysis reveals that in this dual-cavity lens structure, the liquid volumes in upper and lower cavities determine the initial focal length of the lens and the direction of focal length adjustment (Fig. 1). Based on COMSOL simulation results, the influence of liquid volume parameters on the zooming performance of the liquid lens is analyzed. It is found that within a 100 mm focal length range, when the total liquid volume is 3.15 mL and the volume ratio of the upper cavity is 0.21, the lens achieved the largest focal tuning range (Fig. 4). Based on the optimized parameters obtained from simulation, the liquid lens is fabricated (Fig. 5). The focal length of the lens is measured using a focal length meter, showing a decrease from 97.69 mm to 79.72 mm under a voltage of 5 kV, which closely matched the simulation results (Fig. 6). By adjusting the driving voltage, the liquid lens can clearly focus on objects at different distances (Fig. 7). With the increase of the driving voltage, the imaging resolution of the lens improves from 35.919 lp/mm at 0 kV to 45.255 lp/mm at 5 kV (Fig. 8). The imaging spot of the lens exhibits an approximately circular distribution, and the spot diameter gradually decreases with increasing voltage, reaching 178.68 μm at 5 kV (Fig. 9). Using ImageJ software, the light intensity distribution of the spot under different voltages is obtained. The full width at half maximum (FWHM) of the spot decreases with increasing voltage, reaching 93.12 μm at 5 kV, indicating that the focusing ability of the lens gradually enhanced. Under square wave excitation, the average response time of the lens reaches 56 ms (Fig. 10).ConclusionsIn this paper, a dual-cavity dielectric elastomer tunable lens structure is designed. The structural design and working principle of the liquid lens are introduced, and the deformation of the DE film under the driving voltage is simulated. The influence of the liquid volume parameter on the focusing performance of the lens is analyzed, and the focusing ability, imaging quality, focal spot size, light intensity distribution, and response time of the liquid lens under different driving voltages are tested. The results show that the initial focal length and the focusing ability of the liquid lens can be precisely controlled by adjusting the total volume and the volume ratio in the upper cavity. As the driving voltage increases, the focal length and focal spot diameter of the lens decrease, image details become clearer, and imaging resolution and energy concentration of the focal spot improve accordingly. When upper and lower cavities are filled with 0.66 mL and 2.49 mL of liquid, the focal length of the liquid lens is reduced from the initial 97.69 mm to 79.72 mm under the voltage drive of 5 kV, with a focal length variation of 17.97 mm, a resolution of 45.255 lp/mm, a focal spot diameter of 178.68 μm, and a response time of approximately 56 ms. The introduction of the passive film on the side wall of the lower cavity simplifies the design of the liquid lens structure and improves the zoom capability and control accuracy of the lens. Above results provide a theoretical reference and experimental basis for the optimal design and further practical application of liquid tunable lenses.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1623001 (2025)
  • Zhufang Chu, Jiahui Zhao, Hengjing Liu, Luyao Sun, Jidan Yang, and Bingxiang Li

    ObjectiveThe rapid advancement of optical communication technologies and optoelectronic display industry has highlighted the limitations of conventional nematic liquid crystal materials regarding response speed. The current millisecond-scale response time of liquid crystal materials fails to meet the requirements of high-performance devices for fast dynamic response, while limiting the performance enhancement of high-speed optical communication modulation devices. To address this technical challenge, this study examines a novel ferroelectric nematic liquid crystal material DIO, systematically investigating its electro-optical response characteristics under wide temperature range (40?68 ℃) and low driving electric field conditions, and exploring the relationship among response characteristics, voltage, and temperature. This research provides novel technical solutions to overcome the key limitations of slow response speed and high driving voltage in traditional liquid crystal materials, while establishing theoretical foundations and practical guidelines for developing next-generation high-performance optoelectronic display devices and high-speed optical communication components.MethodsThis study employed a combined approach of theoretical analysis and experimental verification. The theoretical framework was based on the physical mechanism of the Frederiks transition effect, describing liquid crystal molecular reorientation under applied electric fields, and the synergistic mechanism between dielectric anisotropy and elastic constants was analyzed. The experimental design incorporated an innovative liquid crystal cell with a cross-shaped electrode structure. This design addressed the limitations of traditional parallel-plate electrodes and enabled multidimensional precise control of liquid crystal molecular orientation. Electric field distribution of the cross-shaped electrode structure was simulated using MATLAB software. The simulation results confirmed that this design generated a uniform in-plane electric field, supporting the experimental findings. The experimental setup utilized polarized optical microscopy for texture change observation during molecular reorientation, photodetectors for transmittance intensity measurements, and high-bandwidth digital oscilloscopes for response waveform recording. This configuration enabled comprehensive monitoring of key parameters including liquid crystal director rotation process and transmittance intensity changes. Through controlled variable experiments, the study systematically examined the dynamic response characteristics of DIO material under various electric field conditions and temperature ranges, analyzing their correlations.Results and DiscussionsThe ferroelectric nematic liquid crystal DIO exhibits remarkable electro-optical response properties. Specifically, it demonstrates stable and fast switching under an ultralow driving electric field of 2×10? V/m across a broad temperature range of 40?68 ℃ (Fig. 6). This required electric field strength is four orders of magnitude lower than that of conventional nematic liquid crystals and two orders of magnitude lower than polymer-stabilized liquid crystals, substantially reducing the driving voltage requirements (Table 1). In terms of dynamic response, DIO shows microsecond-scale fast switching characteristics, with rise time and decay time of approximately 500 μs and 650 μs, respectively. Moreover, the transmitted light intensity increases significantly while the response time decreases markedly with increasing DC voltage, pulse voltage, and temperature (Fig. 6).ConclusionsThis study systematically investigated the microsecond-scale electro-optical response capability of ferroelectric nematic liquid crystal DIO based on the Frederiks transition principle. Experimental results reveal that increasing DC voltage, pulse voltage and temperature leads to enhanced transmittance intensity and reduced response time. DIO demonstrates significant advantages in both low driving electric field and wide temperature range operation compared to conventional nematic liquid crystals, polymer-stabilized liquid crystals and other ferroelectric nematic liquid crystals. The material’s excellent performance under wide temperature ranges and low driving voltages suggests promising applications in next-generation liquid crystal displays, liquid crystal lenses, and optical switching devices. However, DIO exhibits performance degradation below 45 ℃, primarily due to increased viscosity coefficient and decreased dielectric anisotropy at low temperatures. Future research directions may include improving the material’s room temperature performance through molecular structure modification by incorporating flexible groups to lower phase transition temperature, or developing DIO-based composite materials to enhance low-temperature response performance.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1623002 (2025)
  • Shuo Deng, Qiang He, Xuan Xiong, and Xiangshui Miao

    ObjectiveDynamic reconfigurable metasurfaces represent an advancing frontier in optical and materials science, enabling the modulation of optical properties for applications ranging from smart displays to optical sensing. Traditional methods, however, are predominantly restricted to single mode (reflection or transmission) displays and exhibit significant loss. This paper presents a novel all-dielectric metasurface incorporating low-loss phase change material (PCM) Sb2S3, designed to achieve concurrent high-efficiency reflective and transmissive structural color display in the visible spectrum. The research demonstrates dynamic optical control through reversible phase transitions of Sb2S3, and investigates the applications of such metasurfaces in multifunctional optical devices.MethodsThe metasurface architecture comprises TiO2 nanopillars integrated with a Sb2S3 PCM layer at the base, positioned on a transparent quartz substrate. The configuration utilizes Mie resonance effects in dielectric nanostructures to achieve robust light-matter interaction with minimal loss. Critical structural parameters include nanopillar diameter (D), gap between nanopillars (G), and thicknesses of the Sb2S3 layer (tPCM), upper TiO2 layer (tTiO2), and bottom TiO2 reflector (tsub). Through switching Sb2S3 between amorphous (low extinction coefficient) and crystalline (high extinction coefficient) states, the effective refractive index and loss of the metasurface are dynamically modulated, enabling regulation of electric dipole (ED) and magnetic dipole (MD) resonances. Time-domain finite-difference (FDTD) simulations were performed to examine the optical responses, including reflectance, transmittance, and CIE chromaticity coordinates, under varying structural parameters and PCM states.Results and DiscussionsThe metasurface structure, incorporating TiO2 nanopillars with a Sb2S3 PCM layer at the base (Fig. 1), employs Mie resonance to achieve independent tuning of ED modes through layer thickness adjustment. Increasing tPCM amplifies optical loss in the crystalline state, suppressing ED resonance (positioned at the nanopillar base) more substantially than MD resonance (distributed internally), resulting in stronger attenuation of the ED-related reflection peak (up to 70% intensity decrease) compared to the MD peak during phase transition [Figs. 3(a)?(b)]. Modification of D and G enables full visible spectrum coverage in reflective mode, with reflection peaks redshifting as nanopillar volume increases due to enhanced effective refractive index, and a broad CIE chromaticity gamut (exceeding 70% of CIE 1931 space) in the amorphous state that contracts upon ED suppression in the crystalline state (Fig. 5). In transmissive mode, the metasurface exhibits notch patterns corresponding to reflection peaks with high transmittance (>60%) due to low-loss materials, with minimal PCM impact as the transmission path primarily traverses the nanopillar and substrate, avoiding the PCM base [Fig. 2(d)?(f)]. The PCM’s volume directly affects tuning efficiency: larger tPCM enhances extinction coefficient in the crystalline state, promoting ED suppression and peak redshifting [Figs. 3(a)?(b)]. Furthermore, the metasurface demonstrates high sensitivity to environmental refractive index, with reflection peaks redshifting and colors changing from green to magenta as the refractive index increases from 1.0 to 1.44, enabling visual refractive index sensing (Fig. 6). The low extinction coefficient of amorphous Sb2S3 ensures high energy efficiency (reflectance >70%, transmittance >80%), with crystalline-state absorption primarily affecting ED modes while preserving MD functionality, demonstrating independent resonance control through PCM phase transitions.ConclusionsThis investigation presents a reconfigurable transflective structural color metasurface utilizing low-loss PCM Sb2S3, engineered for full-color reflective and transmissive display applications. Through optimization of structural parameters and film thicknesses, the metasurface achieves comprehensive visible spectrum coverage and dynamic optical tuning via independent control of ED resonances through Sb2S3 phase transitions. The architecture incorporates polarization-insensitive cylindrical TiO2 nanopillars with a Sb2S3 PCM layer embedded at the base, adjacent to a TiO2 reflector on a quartz substrate. This configuration optimizes energy efficiency by reducing polarization dependence and utilizing low-loss materials, facilitating simultaneous high reflectance (>70%) and transmittance (>80%) across the visible spectrum. PCM’s strategic positioning at the ED resonance region enables selective suppression of ED modes during crystallization, while MD resonance remains largely unaffected due to internal nanopillar distribution. Computational analyses demonstrate that the upper TiO2 layer primarily modulates MD resonance wavelengths, while the bottom TiO2 layer affects ED resonances. In the amorphous state, pronounced Mie resonances produce vibrant structural colors, whereas crystallization disrupts ED resonances through increased extinction, diminishing reflectance and constraining the color gamut. The tuning efficiency increases with PCM volume, as thicker Sb2S3 layers enhance loss-induced ED suppression. Furthermore, the metasurface exhibits significant sensitivity to environmental refractive index variations, with reflection peaks redshifting and colors changing noticeably across different media, indicating its potential for optical sensing applications. This research introduces an adaptable metasurface design that integrates dual-mode color display, dynamic tunability, and low-loss operation, providing novel perspectives for reconfigurable optical devices in smart displays, optical communications, and adaptive sensing systems.

    Aug. 07, 2025
  • Vol. 45 Issue 16 1624001 (2025)
  • Wenhui Xin, Yixin He, Jie Yao, Yu Zhang, Shichun Li, Yan Guo, Huige Di, Shan Gao, and Dengxin Hua

    ObjectiveLidar, a high-performance active sensing technology, has gained significant traction in recent years across fields such as meteorology, climate science, and environmental monitoring. It serves as a principal methodology for detecting atmospheric physical properties by analyzing echo signals generated from interactions between emitted narrow-pulse lasers and atmospheric constituents, offering high resolution and extended detection range. The quadratic attenuation of lidar echo signals with distance causes single-pulse returns to be overwhelmed by ambient noise, necessitating multiple accumulations (time integration) to enhance the signal-to-noise ratio (SNR). However, conventional lidar data acquisition and integration methods involve onboard storage of data chains collected from a single trigger event. After reaching the predetermined number of accumulations, the stored data are sequentially read out and averaged. This approach introduces a time overhead during data readout, as data acquisition cannot occur simultaneously with readout, creating an acquisition dead time. When the repetition frequency of the lidar system increases to the kHz level, the interplay between acquisition and readout becomes a limitation, resulting in extended dead time and potential pulse omissions.MethodsIn this paper, a “read-accumulate-store” intellectual property (IP) was proposed, which was developed within a field-programmable gate array (FPGA) with the dual-port RAM and an adder, enabling temporal integration of corresponding points in a data linked list. Its core innovation lies in concurrent acquisition and integration: during each acquisition cycle, prior results are retrieved, accumulated with current data, and stored iteratively until the preset accumulation count is achieved. The temporal integration IP architecture, implemented in FPGA, comprises the components of photoelectric conversion module (transforms optical signals into electrical signals), analog-to-digital (A/D) module, accumulation (ADD) module, dual-port RAM storage and control unit. To achieve the designed function, the control unit executes the following steps. First, prior to the first trigger, all storage units in the linked list are reset to zero. Second, the first data point is acquired when the first trigger arrives, and the value stored in the first position of the linked list is read. These two values are summed and stored back in the first position. This process repeats N times to complete the sampling for the first trigger. Since the storage units were cleared before the first trigger, each unit contains the result of a single acquisition after this step. For the second trigger, the same procedure as in the previous step is repeated. As the values in each storage unit of the linked list are read during this acquisition, each unit now contains the cumulative result of the corresponding points from the first and second acquisitions after completion. For the M-th trigger, this process continues, with each storage unit holding the cumulative result of the corresponding points from the previous M acquisitions. Finally, once the predetermined number of accumulations, P, is reached, the accumulated data are read out and transmitted to the host computer, yielding the final integrated results. This method ensures efficient data processing by integrating acquisition, accumulation, and storage, thereby facilitating high-fidelity temporal integration for lidar systems.Results and DiscussionsTo verify the effectiveness of the proposed design, a square wave signal, superimposed with a 0.2 V Gaussian white noise, with a period of 100 Hz and an amplitude of 1 V, was used as the input signal for testing. The results demonstrate that the signal becomes progressively smoother with the increase in accumulations, indicating significant noise signal attenuation (Fig. 5). The SNR increased by 42 dB after 1000 accumulations, confirming that the data acquisition and integration module can achieve multiple acquisition accumulation to reduce background noise and improve SNR. For further verification, this module was compared with the DPO5104 digital oscilloscope of Tektronix and the PXI-9826 acquisition card of ADLINK Technology in actual measurements of laser radar echo signals. After correcting the signals obtained by the three devices by the square of the distance, The RSCS, which eliminates the factor of attenuation factor of light transmission, revealed that despite some variations, the waveform trends were fundamentally consistent, and the position data of the thin cloud layer measurements were largely concordant. Finally, this method was applied to a high-repetition-frequency polarized Mie lidar system, with a 5 kHz repetition frequency, achieving data acquisition at a 50 MHz sampling rate and performing over 80000 cumulative averages, successfully determining the extinction and depolarization ratio coefficients (Fig. 8).ConclusionsA distinctive requirement in the digitization of lidar echo signals, setting it apart from other methods, is the need for temporal integration. While current acquisition techniques effectively handle data collection and temporal integration at low repetition frequencies, limited research addresses these processes at lidar repetition frequencies in the kHz range. This paper presents a novel “read-accumulate-store” method that enables temporal integration of corresponding data points within a linked list structure. This approach simultaneously reads previous acquisition results, accumulates them with current acquisition data, and stores the resulting sum, achieving seamless temporal integration. To implement this method, an intellectual property architecture for temporal integration was developed within a FPGA, utilizing dual-port RAM and an adder. SNR analyses and practical testing demonstrate that this method enables high-speed data acquisition and temporal integration in high-repetition-frequency lidar systems. Additionally, the method offers flexible configuration of parameters, as channels, sampling frequency, sampling length, and integration settings can be adjusted by the hardware description language. Its single-chip integration capability enhances both cost-effectiveness and compactness. Given its versatility and performance, this method shows potential for standardization as a modular component in lidar systems, promoting widespread adoption in advanced sensing applications.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1628001 (2025)
  • Mu Zhou, Linli Zhou, Wei He, and Liangbo Xie

    ObjectiveNon-line-of-sight (NLOS) positioning technology has become an important research direction in the field of emergency response, as it can overcome line-of-sight limitations and achieve target positioning in obstructed scenarios. The core of the technology lies in the extraction accuracy of the photon flight time. However, existing methods exhibit deficiencies in extraction accuracy, limiting their application in complex environments. Moreover, current quantum-based positioning technologies fail to consider positioning performance under target occlusion. The coincidence counting method, while being fundamental for obtaining photon time-of-flight (ToF) in quantum techniques, suffers from high computational overhead that restricts practical efficiency. To address these challenges, this paper proposes a NLOS ToF positioning method based on coincidence count optimization.MethodsThis paper proposes a NLOS ToF positioning method based on coincidence count optimization. The system framework of the proposed method is illustrated in Fig. 1. First, entangled photon pairs are generated through the spontaneous parametric down-conversion (SPDC) process in a nonlinear crystal, providing a stable signal source for the positioning system. Next, based on the NLOS photon transmission model, a correlation-based photon selection algorithm is employed to identify entangled pairs. This algorithm records photon arrival timestamps using single-photon detectors (SPDs), obtains preliminary photon ToF through coincidence counting, and improves ToF accuracy by filtering out uncorrelated photon pairs according to the time-correlation characteristics. In the final stage, the optimized photon ToF information is combined with a back-projection algorithm to achieve high-precision positioning in NLOS environments.Results and DiscussionsWhen the ToF extraction error is 0.045 ns, the cumulative probabilities for raw unprocessed data, Gaussian filtering, median filtering, and wavelet filtering are 14.10%, 21.61%, 16.54%, and 12.11% respectively. In contrast, the quantum coincidence counting reaches 90.42% probability through photon-pair correlation filtering (Fig. 7). To assess ToF method robustness, light path length varies from 10 to 22 m across 7 surface points. As the light path length increases, photon loss during propagation reduces SPD counts, increasing ToF errors in both conventional and coincidence methods. Gaussian filtering shows an error increase from 0.05 to 0.32 ns, while median filtering shows an increase from 0.043 to 0.44 ns, and wavelet filtering from 0.07 to 0.23 ns. In contrast, the quantum coincidence counting method demonstrates superior robustness, with its time error only increasing from 0.01 to 0.09 ns. This enhanced performance is attributed to the method’s effective utilization of strong quantum correlations between entangled photon pairs (Fig. 8). Furthermore, with 5×10? photon arrival events, the quantum coincidence counting method requires 750 s of computational time at its optimal 250 ps window width, whereas the proposed method maintains computational costs below 9 s across all window widths [Fig. 10(b)]. Under a 600 ps coincidence window width, with 5×10? photon arrival events, quantum coincidence counting requires approximately 413 s for timestamp processing. The proposed method reduces this to just 5 s by filtering uncorrelated photons [Fig. 11(a)]. Finally, when the scanning point combination is A, B, C, and D, the ToF resolution performance of all methods deteriorates due to small relative distances between the points. The proposed method achieves horizontal, vertical and total errors of 0.012 m, 0.054 m, and 0.066 m, respectively, whereas the corresponding errors for the quantum coincidence counting method are 0.078 m, 0.141 m, and 0.219 m. When the scanning points are changed to C, D, F, and G, the extended light path reduces signal photons and increases noise. The total positioning errors for Gaussian filtering, median filtering, wavelet filtering, and quantum coincidence counting rise significantly to 1.27, 1.28, 1.80, and 0.04 m, respectively (Fig. 14). The proposed method attains a 90% confidence probability within a positioning error of 0.25 m, while the quantum coincidence counting method requires an error margin of 0.5 m to reach the same confidence level. In comparison, Gaussian filtering achieves 89.56% confidence at 3.25 m, while median filtering reaches 89.3% at 6.5 m, and wavelet filtering attains 95.4% at 8.25 m (Fig. 15).ConclusionsTo address the positioning accuracy degradation caused by low-precision photon ToF extraction, we propose a NLOS ToF positioning method based on coincidence count optimization. First, the second-order correlation properties of entangled photon pairs are utilized to obtain high-precision ToF through coincidence counting. Then, to mitigate the high computational overhead associated with quantum coincidence counting methods, entangled photon pair selection algorithm based on time correlation is proposed, which effectively reduces the time consumption of coincidence counting while maintaining timing accuracy. Finally, precise NLOS positioning is achieved using the optimized photon ToF. Experimental results demonstrate that the proposed method achieves a timing error of 12.75 ps at a coincidence window width of 350 ps, enabling NLOS positioning with an error of 1.2 cm.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1628002 (2025)
  • Changquan Zhuang, Heming Wei, Tao Jin, Xiao Wu, Mengshi Zhu, Fufei Pang, and Dengwei Zhang

    ObjectiveVibration sensors are of crucial importance in structural health monitoring and fault diagnosis within industries like aerospace, transportation, and precision machinery. Through the accurate capture and analysis of vibration signals, potential faults can be effectively predicted and prevented, thus enhancing the reliability and safety of the system. Compared with traditional electronic vibration sensors, fiber-optic vibration sensors have been widely studied due to advantages of high sensitivity, electromagnetic interference resistance, and remote detection capability. They are mainly classified into two categories: fiber Bragg grating (FBG) types and interferometric types. Among them, FBG vibration sensors are restricted by axial sensitivity characteristics of optical fibers and the temperature-strain cross-coupling effect, and their ability to decouple multiple physical fields and conduct broadband measurement is limited. Vibration sensors based on Fabry-Perot interferometry (FPI) provide a highly promising alternative solution to the above problems, but further breakthroughs are still needed in miniaturization, high dynamic response, broadband detection, and high-resolution signal demodulation. We endeavor to create an innovative micro-electro-mechanical system (MEMS) based FPI vibration sensor with the intention of tackling a diverse range of challenges. The sensor is equipped with a circular curved beam diaphragm, optimized spectral phase demodulation algorithm, and field-programmable gate array (FPGA) acceleration processing technology to achieve high dynamic response, ultra-wide measurement range, and strong environmental adaptability.MethodsWe integrate finite element simulation, MEMS micro-nano processing, ceramic packaging technology, and advanced signal processing techniques through a multidisciplinary collaborative approach to construct a high-performance fiber optic vibration sensing system. Firstly, by designing a circular composite elastic membrane structure with three pairs of bent beams arranged with 120° circumferential symmetry, a precision coupled spring mass system is constructed. Finite element analysis software is used to determine the resonance frequency of the membrane under different beam widths, beam thicknesses, mass block thicknesses, and mass block radius sizes. An optimized scheme with a resonance frequency of 6134 Hz is determined, and its wide working bandwidth within 1/3 of the resonance frequency meets mainstream industrial vibration detection requirements. Then, the 8 mm×8 mm micro-sensor is packaged using alumina ceramic encapsulation technology, combined with a demodulation algorithm based on spectral Fourier transform peak finding and phase estimation to effectively suppress phase jump errors and improve dynamic response characteristics. Finally, a demodulation system consisting of broadband light source, spectral module, and FPGA is built, and hardware acceleration of spectral preprocessing, fast Fourier transform, and spectral phase demodulation algorithm are achieved through a pipeline architecture. Finally, the cavity length data and interferometric spectra are synchronously transmitted to the upper computer via the universal serial bus (USB) interface. During the experimental verification phase, a reference level piezoelectric accelerometer with a sensitivity of 100.14 mV/g and a frequency response covering 1?10000 Hz is used for synchronous calibration. Accurate data calibration is achieved by placing the sensor on an electromagnetic vibration table, and the measured system’s dynamic response characteristics are significantly better than traditional demodulation schemes.Results and DiscussionsAfter the completion of the vibration testing system, the demodulation rate of 20 kHz is first verified using a testing device based on piezoelectric transducer (PZT), and the stability of the sensor is verified through a 60-s background noise test. Under soundproof conditions, the fluctuation of the Fabry-Perot (FP) cavity length is measured to be stable within ±0.1 nm, fully demonstrating the excellent stability and demodulation resolution of the designed sensor. Secondly, through frequency response testing, the operating bandwidth of the sensor is found to cover 0?1400 Hz when it is under 1g excitation. The resonance frequency measured by the impact test is 5356 Hz, which is basically consistent with the simulated resonance frequency. Subsequently, the vibration acceleration response of the fiber-optic vibration sensor is tested with a vibration table. These acceleration sensitivities measured at 200, 400, and 600 Hz are 18.135 nm/g, 18.347 nm/g, and 18.577 nm/g, respectively, with linear fitting coefficients of 0.999. This indicates that the vibration sensor exhibits extremely high acceleration linearity within its operating range. To verify the consistency of the vibration sensor, three repeated tests are conducted, and the linearity is measured to be 0.094%. The anti-interference test shows that the ratio of cross sensitivity to axial sensitivity under 600 Hz excitation is less than 7.649%, and the anti-interference ability in multi-dimensional vibration environment is verified. Finally, to evaluate the maximum measurement range of the sensor, limited by equipment conditions, a free fall impact test is used to successfully capture a 200g half sine waveform, and the dynamic range of ±182g is verified through sensitivity conversion.ConclusionsIn this paper, a circular curved beam structure elastic membrane is developed based on MEMS technology, and a fiber-optic FP vibration sensor is integrated with a ceramic bracket. Furthermore, a spectral-phase demodulation algorithm is constructed based on FPGA hardware, achieving high dynamic response detection capability at 20 kHz. Experimental tests have shown that the resonant frequency of the developed sensor is 5356 Hz, and its axial sensitivity is 18.577 nm/g @ 600 Hz. The repeatability is measured at 600 Hz to be less than 0.094%, the resolution can be reached to 0.01g, and the maximum measurement range has been extended to ±182g. Owing to its characteristics of wideband response, large measurement range, high resolution, and fast dynamic response, the proposed sensor has broad prospects in aerospace application.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1628003 (2025)
  • Xueqing He, and Mingming Zhang

    ObjectiveSurface plasmon resonance (SPR) represents a collective oscillation phenomenon generated by free electrons under electromagnetic wave excitation. This distinctive optical phenomenon transcends the diffraction limit of traditional optics and enhances light-matter interaction, garnering significant attention. These characteristics have catalyzed advancements in optoelectronics and nanotechnology. Conventional SPR sensors employ Kretschmann prism coupling or grating coupling for surface plasmon excitation. However, their substantial size and integration challenges fail to meet modern sensing technology requirements for miniaturization, high sensitivity, and multi-parameter detection. The photonic crystal fiber (PCF)-SPR sensor integrates photonic crystal light field regulation capabilities with SPR refractive index sensitivity characteristics, potentially overcoming traditional SPR sensor limitations. Despite the PCF-SPR sensor having notable advantages in refractive index sensing, it faces technical limitations regarding insufficient sensitivity in detecting low refractive index media (refractive index n<1.33), constraining its practical application in crucial areas such as biological solution analysis. We present a three-open-loop channel D-type PCF-SPR refractive index sensor designed for low refractive index sensing. The open-channel configuration enhances sensing capabilities by expanding the effective sensing area and increasing fiber core exposure to the target environment. This design innovation augments detection sensitivity while preserving mechanical stability.MethodsThe PCF incorporates three distinct pore sizes arranged in a double-layer rhombus configuration. The analyte microfluidic channels were constructed through the introduction of three open rings on the D-shaped polishing plane, with gold plating films applied to the open rings’ inner surfaces. Au was selected as the plasma material due to its established stability in chemical interactions. The open channel design reduces the distance between the fiber core and metal layer, facilitating faster detection response and generating enhanced SPR effects at the metal-dielectric interface. This study employs finite element analysis to evaluate how system structural parameters, including pore size, pore spacing, metal thickness, and fiber core-to-polishing plane distance, influence sensor performance.Results and DiscussionsThe proposed three-open-loop channel D-type PCF-SPR refractive index sensor demonstrates superior sensing performance through optimization of pore size, pore spacing, metal thickness, and fiber core-to-polishing plane distance. Initially, we examine the influence of pore diameters (d1, d2, and d3) on the fundamental mode loss spectrum for analyte refractive indices na of 1.26 and 1.27. In Fig. 3, distinct sharp SPR formant peaks emerge at pore diameters of d1=1.2 μm, d2=1.5 μm, and d3=1.8 μm. In Fig. 4, with fixed pore diameters (d1=1.2 μm, d2=1.5 μm, and d3=1.8 μm), we analyze the effects of pore spacing Λ, gold film thickness tAu, open-loop radius rs and distance from fiber core to polishing plane h on the fundamental mode loss spectrum. Pronounced SPR formant peaks appear at structure parameters of Λ=3 μm, tAu=50 nm, rs=1 μm, and h=6 μm. In Fig. 5, we examine the fundamental mode loss spectrum distribution across analyte RI na values from 1.15 to 1.29. The sensor achieves a maximum wavelength sensitivity of 14500 nm/RIU with a corresponding refractive index resolution of 6.90×10-6 RIU and a figure of merit (FOM) of 114.94 RIU-1. For enhanced measurement accuracy, the sensing range is subdivided into 1.15?1.22 and 1.23?1.29. Linear fitting achieves a correlation coefficient of 0.98299 for the lower range, while third-order polynomial fitting yields a correlation coefficient of 0.99663 for the higher range.ConclusionsWe present a high-sensitivity triple open-loop channel D-type PCF-SPR sensor. The finite element method was employed to systematically analyze the influence of structural parameters, including aperture size, pore spacing, distance from the fiber core to the polishing plane, and metal film thickness on sensor performance. The findings demonstrate that the sensor effectively detects RI within the range of 1.15 to 1.29. The sensor achieves a maximum wavelength sensitivity of 14500 nm/RIU, with a corresponding resolution of 6.90×10-6 RIU in the short-wave infrared band of 2?3 μm. Furthermore, the FOM reaches 114.94 RIU-1. The sensor’s superior sensitivity and capability for low refractive index detection make it particularly suitable for precise measurements of biochemical solutions, including liquid carbon dioxide, liquid medical oxygen, sevoflurane, and biological samples.

    Aug. 18, 2025
  • Vol. 45 Issue 16 1628004 (2025)
  • Jian Chen, Na Fang, Shenghui Shi, Binbin Luo, Chengyou Huang, Yao Huang, Xue Zou, and Shanghai Jiang

    ObjectiveRefractive index (RI) represents a fundamental optical parameter extensively utilized in chemical concentration detection, biomedical sensing, and environmental monitoring. Optical fiber RI sensors have garnered significant attention due to their compact structure, high sensitivity, and robust immunity to electromagnetic interference. Various optical fiber RI sensors have been developed, including straight-core, S-shaped, U-shaped, and D-shaped structures. U-shaped fiber sensors have emerged as a prominent research focus due to their enhanced evanescent field interaction, high RI sensitivity, compact structure, straightforward fabrication process, and suitability for localized sensing applications. This study proposes a reflective U-shaped micro-nano optical fiber RI sensor designed to enhance RI sensitivity and spectral performance. The sensor combines a tapered thin-core fiber structure with a U-bending geometry to amplify the evanescent field effect, while incorporating a reflective configuration to establish a dual-path interference mechanism. This design enhances the interaction between guided light and the surrounding medium while improving the spectral extinction ratio, thereby achieving superior RI sensitivity.MethodsThe beam propagation method was utilized to simulate optical field energy distribution under different bending diameters and taper waist diameters. A section of thin-core fiber (TCF) with a core diameter of 2.1 μm and a cladding diameter of 125 μm was first fusion-spliced to a single-mode fiber (SMF). The TCF underwent tapering using a hydrogen-oxygen flame tapering system. During the process, the hydrogen flow rate, stepper motor speed, and pulling length of the tapering machine were precisely controlled to achieve accurate regulation of the waist diameter. Subsequently, the micro-nano fiber was bent into a U-shape using a 3D-printed V-groove clamp, enabling precise control of the bending diameter. To examine the influence of structural parameters on sensing performance, two experimental configurations were established. 1) The waist diameter was fixed at 8.41 μm, while the bending diameter varied among 2 mm, 3 mm, and 4 mm. 2) The bending diameter was fixed at 2 mm while the waist diameter varied among 11.51 μm, 7.83 μm, and 5.58 μm. In RI sensing, variations in the surrounding RI modify the phase difference between the core and cladding modes, resulting in shifts in the resonance wavelength of the interference spectrum. A supercontinuum broadband light source (KG-ASE-D-1-1-FA, 1200?1600 nm) was coupled into the sensor via an optical circulator [(1550±50) nm]. The light reflected from the cleaved end face of the fiber re-entered the sensing region and was directed to an optical spectrum analyzer (OSA, Yokogawa AQ6370B, 600?1700 nm), where the reflected spectrum was monitored in real time. The RI sensitivities corresponding to different structural parameters were obtained by tracking the positions of the resonance dips.Results and DiscussionsSimulation results demonstrate that reducing both the taper waist diameter and the bending diameter substantially enhances the leakage intensity of the evanescent field, thereby strengthening the interaction between the guided mode and the surrounding medium. This theoretical prediction aligns closely with experimental observations. Under a fixed waist diameter of 8.41 μm, decreasing the bending diameter from 4 mm to 2 mm enhances the RI sensitivity from 1894.66 nm/RIU to 2007.79 nm/RIU (Fig. 6). Similarly, with a fixed bending diameter of 2 mm, reducing the waist diameter from 11.51 μm to 5.58 μm results in a significant increase in sensitivity from 1376.22 nm/RIU to 3511.33 nm/RIU (Fig. 7). Within the RI range of 1.3330 to 1.3519, the resonance wavelength exhibits a red shift with increasing RI, with a maximum shift of 67.86 nm. The linear fitting correlation coefficient R2 consistently exceeds 0.99, indicating excellent linear response characteristics. Compared to traditional transmission-type tapered micro-nano fiber sensors, the proposed reflective U-shaped structure improves the extinction ratio by approximately sevenfold, thereby significantly enhancing the interference spectral contrast. Fast Fourier transform (FFT) analysis confirms that the interference spectrum is predominantly governed by low-order modes, ensuring high spectral stability. Moreover, a one-hour stability test in deionized water reveals a maximum wavelength drift of only 0.22 nm, demonstrating excellent environmental stability (Fig. 8). Based on the spectral resolution of the optical spectrum analyzer (0.02 nm), the minimum detectable RI variation is estimated to be approximately 5.7×10?? RIU. These results collectively demonstrate that the sensor exhibits outstanding performance in terms of repeatability, linearity, and sensitivity. Overall, the integration of a U-shaped geometry with a reflection-based interference mechanism enables high-performance RI sensing while maintaining structural simplicity and fabrication feasibility.ConclusionsThis study presents and experimentally validates a reflective U-shaped micro-nano optical fiber RI sensor, integrating the advantages of U-shaped geometry and tapered micro-nano fiber. A systematic investigation examined the influence of bending diameter and waist diameter on sensing capabilities. The U-shaped configuration demonstrates enhanced evanescent field interaction with surrounding media compared to straight micro-nano fiber structures, making it particularly effective for contact sensing of biomolecules and chemical solutions. Additionally, the reflective configuration substantially improves the spectral extinction ratio. These combined features enhance sensitivity, spectral stability, and signal quality. Experimental findings demonstrate that RI sensitivity inversely correlates with both bending diameter and waist diameter. The resonance wavelength exhibits a red shift with increasing RI within the range of 1.330 to 1.3519, maintaining excellent linearity. The sensor achieves a maximum sensitivity of 3511.33 nm/RIU with robust spectral stability. Future research directions include sensitivity enhancement through surface plasmon resonance effects and integration of emerging two-dimensional materials characterized by strong light-matter interaction, large specific surface area, and excellent biocompatibility. These advancements aim to broaden the sensor’s applications in biomedical diagnostics, chemical detection, environmental monitoring, and food safety.

    Aug. 08, 2025
  • Vol. 45 Issue 16 1628005 (2025)
  • Xu Gao, Mian Chen, Zaiping Chen, Xiaopeng Yan, Zhe Wang, Zhihui Qian, Xiong Li, Jiajia Wang, Kaiwei Li, Lei Ren, and Luquan Ren

    ObjectiveFlexible tactile sensors demonstrate significant potential for applications in robotics, human-computer interaction, advanced prosthetic design, and medical monitoring. In robotic systems, the ability to detect both contact force and its position is essential for effective object manipulation. While electrically based flexible tactile sensors have been extensively researched over the past decade, their practical implementation faces challenges including high manufacturing costs, parasitic effects, complex circuits, and signal crosstalk. Optical sensing schemes have emerged as a promising alternative, with optical waveguide/fiber-based tactile sensors garnering particular attention. Initially, optical fiber Bragg gratings were employed for tactile sensing. However, the inherent brittleness and limited flexibility of silica-based optical fibers restrict their integration onto curved surfaces and spatial resolution capabilities. Polymer optical waveguides/fibers present enhanced flexibility, enabling seamless integration into curved surfaces and flexible structures, thus offering a more effective tactile sensing solution. This study proposes a flexible dual-core multimaterial optical waveguide tactile sensor featuring a multilayered structure with distinct cores for force detection and position discrimination. To address existing challenges, this research develops a distributed flexible dual-waveguide tactile sensor utilizing induced loss blocks.MethodsLeveraging the advantages of polymer optical waveguides, this research develops a distributed flexible dual-waveguide tactile sensor based on induced loss blocks through total internal reflection theory. The multilayer composite structure physically isolates force/position signals, effectively eliminating crosstalk. The integration of fiber thermal drawing technology in waveguide fabrication enables cost-effective mass production while maintaining excellent flexibility and structural reliability.Results and DiscussionsThe tactile sensor implements a multi-layer composite structure, as illustrated in Fig. 1. The design incorporates differentiated upper and lower waveguide structures to achieve decoupled measurement of contact force magnitude and position. The upper waveguide comprises a uniform rectangular cross-section with core and cladding layers, while the lower waveguide integrates rectangular loss-inducing block arrays into the conventional cladding-core configuration. The sensor's measurement principles were theoretically derived using total internal reflection theory and Lamé equations. Numerical simulations utilizing the beam propagation method (BPM) analyzed the optical field distribution characteristics of the lower waveguide core and output power attenuation profiles versus propagation distance (Fig. 2). The analysis examined the relationship between output light intensity and both loss-inducing block width and material refractive index (Fig. 4). Through the combination of fiber thermal drawing technology and conventional molding methods (Fig. 5), a flexible dual-waveguide tactile sensor of 70 mm×5 mm×3.5 mm was successfully fabricated (Figs. 6 and 7). Comprehensive performance testing (Figs. 9?11, Fig. 13, and Fig. 14) demonstrated 6 mm spatial resolution and 1 N stress resolution for contact force position and magnitude measurements. The sensor exhibits a wide dynamic measurement range (0?8 N) with excellent dynamic response characteristics. During extended cyclic loading, signal attenuation remained below 5%, with optical output baseline drift rate smaller than 0.008 h-1. After 3000 loading cycles, the sensor maintained structural integrity without surface cracks, plastic deformation, or delamination.ConclusionsThis research presents the development and fabrication of a distributed flexible dual-waveguide tactile sensor utilizing induced loss blocks through total internal reflection theory. The multilayer composite structure enables simultaneous dual-parameter measurement of contact force position and magnitude without crosstalk. The fabrication process combines fiber thermal drawing technology with conventional molding methods to produce a flexible dual-waveguide tactile sensor of 70 mm×5 mm×3.5 mm. The sensor achieves 6 mm spatial resolution and 1 N stress resolution for position and magnitude measurements, with a dynamic range of 0?8 N and superior dynamic response. Testing over 3000 cycles demonstrated performance degradation below 5% without structural deterioration. The sensor design offers cost-effective, batch-reproducible fabrication with array integration compatibility. Further optimization of materials, structural design, and manufacturing processes can enhance mechanical performance and resolution capabilities.

    Aug. 25, 2025
  • Vol. 45 Issue 16 1628006 (2025)
  • Liang Wang, Wei Liu, Lei Wang, Fan Yan, Yajing Wang, and Fuyuan Qin

    ObjectiveWhen measuring the length and diameter of rod-shaped nanoparticles by using depolarized dynamic light scattering (DDLS), it is often necessary to combine the inversion algorithms for calculation. Commonly used inversion algorithms include exponential fitting and Tikhonov regularization algorithms. The exponential fitting method will produce fitting errors due to improper setting of the initial value, while in the Tikhonov regularization algorithm, when inverting rod-shaped particles, the error of the horizontally depolarized signal will significantly amplify the diameter inversion bias. In order to improve the accuracy of particle inversion, scholars apply neural networks to the field of spherical particle inversion, which effectively improves the accuracy of inversion, and some scholars apply neural networks to non-spherical particles, but not to the length and size of rod-shaped particles. In order to improve the accuracy and repeatability of the inversion, we propose a generalized regression neural network (GRNN) combining Tikhonov and parameter optimization, which can effectively carry out the inversion of rod-shaped particle size.MethodsThe DDLS measurements are performed on the rod-shaped particle samples to obtain the scattering signals of vertical polarization and horizontal depolarization, from which the light intensity autocorrelation function is obtained, and the attenuation linewidth distribution is obtained by the Tikhonov regularized inversion of the autocorrelation function. The light intensity autocorrelation function and the attenuation linewidth distribution are sorted out into the dataset required for the Tikhonov-GRNN model, and the dataset is divided into the data set is divided into training samples and validation samples. After the training is completed, the data in the validation samples are inverted and analyzed: the predicted attenuation linewidth distributions in the vertical polarization and horizontal depolarization directions are obtained by the vertical vertical (VV)-GRNN and vertical horizontal (VH)-GRNN models, the horizontal diffusion coefficients and the rotational diffusion coefficients of the motion of the rod-like particles are calculated respectively, and the lengths and diameters of the rod-like particles are solved by the Tirado-Garcia de la Torre (TG) model.Results and DiscussionsBefore the experiment, 300 particles from experimental samples are observed by transmission electron microscopy first. Results show that actual sizes of the rod particles used in the experiment coincided are mostly consistent with the nominal size (Figs. 4 and 5, Table 1), so the use of nominal lengths and diameters has no effect on experimental results. In this paper, experiments are carried out on two different sizes of rod-shaped nanoparticles. The experiments use the DDLS method to measure the samples, and the data set required for the Tikhonov-GRNN model is obtained. After the training is completed, the autocorrelation function of light intensity is input for the prediction, and the attenuation linewidth distributions of VV and VH are obtained, which led to inverse results of lengths and diameters of rod-shaped particles (Table 4). Experimental results show that the inversion accuracy of the Tikhonov-GRNN model is better than inversion results of the traditional algorithm (Table 5), which is closer to the real size. After repeating experiments many times, repeatabilities of two particle lengths and diameters are 4.5%, 1.6% and 9.8%, 8.9%, respectively, which are better than 5.9%, 16.3% and 3.2%, 17.5% of the traditional Tikhonov regularization algorithm. All these can prove that the stability of the Tikhonov-GRNN model is stronger.ConclusionsWhen using DDLS method to measure the size of rod-shaped particles, the results of current inversion algorithms are not stable due to the complexity of the mathematical model of non-spherical particles. Therefore, based on the nonlinear mapping and generalization ability of GRNN, a GRNN combining Tikhonov’s regularization algorithm and parameter optimization is proposed, which can invert the length and diameter of rod-shaped particles. The method uses DDLS to measure the light intensity autocorrelation function in the VV polarization and VH depolarization directions of the rod particles, respectively, and obtains the attenuation linewidth by Tikhonov regularization inversion, and constructs a training dataset by combining the attenuation linewidth with the corresponding light intensity autocorrelation function. The trained VV-GRNN and VH-GRNN predict the attenuation linewidth to obtain the translational diffusion coefficients and rotational diffusion coefficients of the rod particles, and then calculate the size of the rod particles. Two kinds of gold nanorod particles are measured, and the experimental results show that the Tikhonov-GRNN algorithm can realize the inversion of the length and diameter of the rod particles, and compared with the Tikhonov regularization algorithm, the accuracy of the Tikhonov-GRNN inversion algorithm is improved, while repeatabilities of the length and diameter are lower than 4.5% and 9.8%, respectively.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1629001 (2025)
  • Tao Ma, Liping Liu, Xuehong Sun, Fei Yang, Tong Yu, and Haibo Yu

    ObjectiveThis paper aims to design a novel polarization-transforming metasurface for terahertz wavefront control. The design successfully achieves polarization conversion functionality and effectively reduces broadband radar cross section in the terahertz band, representing significant advancement in electromagnetic wave manipulation. The proposed design establishes an efficient approach for terahertz wavefront control while demonstrating the capabilities of encoded metasurface technology in electromagnetic wave regulation. In radar stealth applications, reducing radar scattering cross-sectional area remains a critical research focus with substantial practical value. This research presents both scientific innovation and practical applicability. Through comprehensive comparison with existing literature, this study demonstrates the superior characteristics of the designed polarization conversion metasurface. The design excels in several key performance metrics, including design simplicity, broadband performance, RCS reduction effectiveness, and wide-angle stability. These advantages provide valuable reference for academic research and introduce novel methodologies for related fields. This research advances terahertz wavefront control technology and radar stealth technology by enhancing and complementing existing techniques.MethodsThe research methodology encompasses the design and optimization of polarization-conversion metasurface units and the implementation of intelligent algorithms for optimal array configuration. The study initially proposes a polarization-conversion metasurface element structure for the terahertz band. The structure incorporates 2-bit supercells with digital encoding, based on the PB geometric phase principle. The genetic algorithm, an intelligent optimization approach that emulates natural biological evolution, identifies optimal solutions through natural selection mechanisms. The study achieves significant radar cross section reduction through scattering function design, fitness function implementation, and array layout optimization using genetic algorithms. This approach attains unit polarization conversion efficiency exceeding 0.9 across a broad frequency range. The array demonstrates radar cross section reduction greater than 10 dB across multiple frequency bands, maintaining stable performance over wide angles. For experimental validation, initial theoretical calculations utilize MATLAB’s mathematical computation capabilities to analyze the designed metasurface elements and arrays. Subsequently, CST simulation tools provide detailed electromagnetic simulation of wave-metasurface interactions. The dual verification through MATLAB and CST ensures consistency and accuracy between theoretical calculations and simulation results.Results and DiscussionsThe study presents a novel polarization-converted metasurface element characterized by simplicity, efficiency, and flexibility (Fig. 1). Experimental analysis of the progressively improved structure examines the Ryy values at each structural stage (Fig. 2). Resonant current analysis occurs at three resonant points (Fig. 3). Simulation yields reflection coefficients for same polarization and cross polarization, and polarization conversion efficiency at varying angles (Fig. 4). uv coordinate system simulations provide reflection amplitude and phase of copolarization reflection coefficient (Fig. 5 and Fig. 6). Component rotation angle modifications regulate electromagnetic wave phase (Fig. 7). The design includes four types of 2-bit encoded super units (Fig. 8). Optimization algorithms determine optimal arrangement and electric field characteristics (Fig. 9). Simulations examine scattering beams of different polarized waves under normal incidence and scattering characteristics of xoz surface array and metal plate (Fig. 10 and Fig. 11). RCS reduction exceeds 10 dB for different polarization waves under normal incidence within 0.695?0.846 THz, 0.872?1.195 THz and 1.476?1.674 THz, reaching maximum reduction of 32 dB at 1.05 THz, demonstrating effective RCS reduction (Fig.12). For X-polarized wave incidence angles from 0° to 30°, RCS decreases exceed 8.0 dB within 0.661?1.312 THz. For Y-polarized waves, RCS reduction exceeds 8.0 dB within 0.683?0.849 THz, 0.851?1.275 THz and 1.500?1.687 THz, indicating effective scattering control for oblique incidence from 0° to 30°. However, larger incidence angles diminish RCS reduction effectiveness (Fig.13). Analysis includes scattering beams of linearly polarized waves at 30°, 0.8 THz (Fig.14).ConclusionsThis study introduces a novel terahertz wideband polarization conversion metasurface structure and validates its practical applications through theoretical analysis and simulation experiments. The device demonstrates both linear polarization conversion capabilities and scattering control characteristics. The findings indicate that the polarization converter achieves conversion efficiency exceeding 0.9 for electromagnetic waves at normal incidence within the THz frequency bands of 0.706?0.862 THz and 0.911?1.744 THz. For electromagnetic waves at oblique incidence angles of 0°?40° within the THz frequency bands of 0.692?0.836 THz and 0.873?1.427 THz, the polarization conversion efficiency surpasses 0.7. The proposed structure exhibits enhanced performance in bandwidth, efficiency, and adaptability to small angles. Furthermore, utilizing the PB phase principle and incorporating phase gradient, effective wideband phase modulation has been achieved. The structure demonstrates RCS reduction exceeding 10 dB at normal incidence across multiple frequency bands, and greater than 8 dB reduction for oblique incidence angles from 0° to 30°. These results represent significant improvements over current polarization conversion metasurfaces in terms of RCS reduction and bandwidth limitations. The research advances existing technologies through enhanced wideband characteristics, conversion efficiency, and radar cross section reduction capabilities.

    Aug. 07, 2025
  • Vol. 45 Issue 16 1629002 (2025)
  • Zhilei Xiao, Songpo Xu, Duoduo Li, Zixi Li, and Zhengrong Wei

    ObjectiveThe interaction of strong laser fields with atoms and molecules provides critical opportunities for observing and controlling photoelectron dynamics on ultrafast timescales. The attoclock technique has significantly advanced the study of attosecond-scale electron tunneling dynamics, enabling the effective extraction of tunneling time delay through photoelectron momentum distributions (PMDs) in elliptically polarized laser fields. Over the past decade, studies on attoclock have gradually expanded from the infrared to the ultraviolet (UV) regime, and an above-threshold ionization (ATI) order-dependent photoelectron angular shift in UV attoclock measurements has been observed. However, the physical mechanisms behind this phenomenon remain unclear, including whether ATI order-dependent angular deflections correlate with tunneling time delays and whether changes in electron energy affect the extraction of tunneling time. Using the time-dependent Schr?dinger equation (TDSE), nonadiabatic model, and model potential (MP) theory, we investigated the PMDs of argon atoms in 400 nm elliptically polarized laser fields. From these results, we extracted the ATI order-dependent photoelectron angular distributions (PADs) and further obtained the tunneling time delay of argon atoms.MethodsTo simulate the photoelectron momentum distributions of argon ionized by an elliptically polarized (EP) laser field with ultraviolet laser wavelength of 400 nm, we numerically solved the three-dimensional TDSE in length gauge, as well as two other theoretical methods, i.e., nonadiabatic model and MP theory. For the time evolution of the electrons after the laser ends, we used Kepler’s laws to get the asymptotic momenta.Results and DiscussionsFig. 1 provides the PMDs calculated using the three theoretical methods under different laser intensities. The simulations are performed with a laser wavelength of 400 nm, intensities of 1.5×1014 W/cm2, 2.5×1014 W/cm2, and 3.0×1014 W/cm2, and an ellipticity of 0.71. All PMDs demonstrate that both the nonadiabatic model and MP theory can qualitatively reproduce the TDSE simulation outcomes, revealing the crucial role of nonadiabatic effects in electron dynamics. Fig. 2 displays the photoelectron energy spectra (PES) at each intensity. Fig. 3 presents the PADs of different-order ATI rings at 2.5×1014 W/cm2. As the ATI order increases, the angular distributions shift towards larger angles. The three theoretical models all qualitatively reproduce this ATI order-dependent angular shift phenomenon. Compared with the nonadiabatic model, the MP theory exhibits a modulation effect on the electron angular distribution, causing the final-state angular distribution to shift towards smaller angles. Fig. 4 shows the dependence of the peak angles of different ATI rings on photoelectron energy for two intensities (2.5×1014 W/cm2 and 3.0×1014 W/cm2). Fig. 4(a) reveals that the angular peak positions of ATI rings increase with higher ATI orders, while the peak angles decrease for the same ATI order when laser intensity increases. By comparing angular deviations between the nonadiabatic model, MP theory, and TDSE results, we can extract the tunneling time delay of photoelectrons for different ATI rings. Fig. 4(b) shows that the nonadiabatic theory produces predominantly negative angular deviations from TDSE results, leading to a negative tunneling time delay. In contrast, the MP theory yields positive angular deviations, giving a positive tunneling time delay with an upper limit of 15 as. Fig. 5 demonstrates that initial parallel and transverse momentum distributions are strongly dependent on ATI orders. With increasing ATI order, the peak of parallel momentum shifts towards negative values while the peak of transverse momentum moves positively. Concurrently, the tunneling exit moves closer to the nucleus at higher ATI orders. Fig. 6 reveals that electrons with tunneling exits closer to the nucleus experience larger deflection angles, while those exits farther from the nucleus show smaller deflections. Further analysis indicates that the atomic model potential slightly modulates the final-state polarization angles by influencing electron trajectories in the laser field.ConclusionsWe systematically investigate the ATI order-dependent PADs and tunneling time delay in UV attoclock experiments by TDSE, nonadiabatic model, and MP theory. Theoretical calculations generated PMDs of Ar atoms in 400 nm elliptically polarized laser fields, from which we extract the ATI order-dependent PADs. By comparing the results obtained from solving the TDSE and the other two models, we extract an upper limit of tunneling delay for Ar atoms in 400 nm elliptically polarized laser fields at the peak intensity studied, which is 15 as. We find that the tunneling time delay does not vary with increasing ATI orders at higher laser intensities, whereas it decreases with increasing ATI orders at lower laser intensities. The analysis reveals that the tunneling exits and initial momentum distributions induced by MP and nonadiabatic effects jointly influence the final-state PADs in attoclock measurements, thereby affecting the extraction of tunneling time delay.

    Aug. 15, 2025
  • Vol. 45 Issue 16 1632001 (2025)
  • Please enter the answer below before you can view the full text.
    Submit