Due to uneven ground network coverage and unstable signal quality, it is impossible to achieve full, fast, and real-time monitoring of the railway transportation network, which hinders the rapid development of China's railway transportation network. At the same time, the integration of Low Earth Orbit(LEO) satellite communication and ground 5G has become a hot topic, and the fusion of the two can make satellite networks an effective supplement to ground networks. The current application of satellite ground fusion in 5G-Railway(5G-R) technology scenarios is relatively limited. This article first reviews the current research status and the development trends of railway satellite communication and satellite ground fusion networks, points out development trends, and then analyzes the communication needs of various businesses in railway communication systems. Based on this, a network architecture that integrates low orbit satellites with 5G-R networks is proposed, which not only fills the blind spots in railway communication network coverage, but also provides high-capacity information feedback. It can effectively ensure the safe operation of railway transportation and achieve information and intelligent management of railway transportation.

Traction spoofing is a highly covert type of spoofing interference that can induce the receiver to track spoofing signals without causing the tracking loop to lose lock. For scalar receivers, the tracking loops are independent of each other, so the traction spoofing for a single signal will not be affected by the signals of other channels. All tracking loops of vector receiver are coupled through receiver state and have mutual influence, which means that traction spoofing has different influence on vector and scalar tracking loops. Based on the assumption that there is no noise and the carrier frequency and carrier phase of the spoofing signal and the corresponding authentic signal are equal, this paper analyzes the influence of traction spoofing on the vector tracking loop, derives the success condition of spoofing, and verifies the analysis results by using the signal source simulator and software receiver. The results show that the conditions for successfully implementing traction spoofing on vector tracking loop are more stringent than those of scalar tracking loop, which reflects the inherent anti-spoofing capability of vector tracking loop.

In the Low Earth Orbit(LEO) scenario, beam hopping technology can flexibly allocate system resources and is suitable for the scenario of uneven service distribution. Time Division Duplex (TDD) mode has low requirements on the terminal and is conducive to the development of upstream and downstream asymmetric services. Therefore, a resource allocation algorithm is proposed for low-orbit satellite beam hopping based on TDD. Firstly, the resource allocation model of LEO reverse link supporting beam hopping and Multifrequency-Time Division Multiple Access (MF-TDMA) is established to minimize the time domain resource consumption in condition of satisfying the service requirements. Secondly, considering the dynamic time-delay compensation of satellite and earth, a multi-level beam hopping time slot architecture design is adopted to maximize the available time slot. The upstream and downstream time slot switching model is established, and a beam hopping time slot scheduling optimization method is proposed based on TDD. Simulation results show that compared with traditional MF-TDMA resource allocation method or fixed multi-beam splitting algorithm, the proposed algorithm can effectively improve the time slot utilization and throughput of the system.

Entity recognition technology, as an important step in constructing knowledge graphs, has been extensively applied in natural language processing applications such as semantic network, machine translation, and question answering systems. It plays a crucial role in promoting the practical application of natural language processing technology. According to the development process of entity recognition technology, the existing entity recognition methods are investigated in this paper. These methods can be classified as: early rule and dictionary based entity recognition methods, machine learning based entity recognition methods, and deep learning-based entity recognition methods. The core ideas, advantages and disadvantages, and representative models of each entity recognition method are summarized, especially the latest entity recognition methods based on Bi-directional Long Short-term Memory (BiLSTM) and Transformer. Additionally, the current mainstream datasets and evaluation criteria are introduced. Finally, facing the semantic requirements of future machine communication, we have summarized the challenges faced by entity recognition technology, and its future advancement in Internet of Things(IoT) business data is anticipated.

With the advent of the Internet of Things era, people's demand for higher bandwidth and faster speed is increasing. The introduction of the terahertz frequency band provides a new solution to this problem. Terahertz signals bear the characteristics of strong penetrating power, good penetrability to non-metallic materials, and high resolution, and show great potential in the fields of communication and imaging. Based on the principle of photon assistance, a terahertz transmission system that uses photon heterodyne beat frequency to generate terahertz signals at two frequencies of 386.2 GHz and 438.4 GHz is built. And the penetration of terahertz signals is measured under the obstruction of four common plastic materials: poly4-methylpentene(TPX), polyethylene(PE), polypropylene(PP) and polymethyl methacrylate(PMMA). The attenuation characteristics of these materials in indoor environments are obtained, and the experimental results are fitted with Beer's absorption law. From the experimental results, it is found that the attenuation of terahertz signals by 5 mm thick PP board and PE board is below 1 dB, and the blocking effect of thin PP board and PE board on terahertz signals at 386.2 GHz and 438.4 GHz is very weak; terahertz signals in the frequency band selected for the experiment can hardly penetrate the 15 mm thick PMMA plate. The exploration and research on the transmission attenuation characteristics of terahertz signals in plastic materials has reference value for the selection of suitable antenna materials in millimeter-wave communication systems, and plays an important guiding role in practical applications.

Modern medical devices are more inclined to high precision, high sensitivity, and miniaturization, resulting in more and more complex electromagnetic environments. External electromagnetic interference will also affect medical devices, and even affect doctors' diagnosis in severe cases. In this paper, combined with the heart sound electrocardiograph detector, the conduction emission test, radiation emission test and electrostatic discharge test are studied based on the medical device test standard. Some rectification methods are proposed, such as rectifying the shell, adding the filter circuit, adding the magnetic ring and adding the charge release circuit. It can be used as a reference for the analysis and rectification of other types of related medical devices.

The phase noise level of the frequency source directly restricts the upper limit of radar performance, thus low phase noise frequency synthesis is a key technology for high performance radar systems. Existing low phase noise frequency synthesis methods are often realized by high order frequency doubling, and the overall performance is heavily dependent on the low phase noise Oven Controlled Crystal Oscillator(OCXO), therefore the cost is high. In this paper, a low additional phase noise frequency synthesis method is proposed, the technique of minimizing the additional phase noise on the link is adopted, and an ordinary OCXO is employed to cascade the low phase-noise amplifier, comb generator, and phase-locked loop to ultimately realize the frequency synthesis with low phase noise. The measured data show that the proposed method uses a 100 MHz OCXO as a reference, the time jitter in the integration interval [ 1 kHz, 30 MHz] is 11 fs, the phase noise of the frequency synthesized at the 5.8 GHz carrier is -119 dBc/Hz@1 kHz, the time jitter in the integration interval [ 1 kHz,30 MHz] is 13.7 fs, the total additional time jitter is 8. 17 fs, and the additional phase noise is only 1.9 dB. It reaches the industry-leading level, which can effectively improve the imaging performance of millimeter-wave radar systems and outperforms the traditional frequency synthesis methods.

As an important component in smart grid, the Feedback Terminal Unit(FTU) occasionally faces the unexpected shutdown due to the extreme operation environment. The failure rate prediction of massive Feedback Terminal Units(FTUs) is investigated by using the Stacked Denoising Autoencoder (SDAE) failure rate estimation method improved by the Dropout Regularization operation to prevent overfitting. Adadelta algorithm is employed to optimize the learning rate. An accurate failure prediction is realized with satisfied learning rate. A series of experiments are conducted to verify the advantages of the proposed method in solving the FTU failure estimation problem

In order to effectively reduce the frequency estimation error of single complex sinusoid, aiming at the estimation error of the traditional Rife algorithm caused by the judgment error of the maximum amplitude and sub maximum amplitude spectral line under the condition of low signal-to-noise ratio, replacing the Fast Fourier Transform(FFT) result with frequency estimation result of Root Multiple Signal Classification(RootMUSIC) algorithm is adopted to determine the maximum amplitude and sub maximum amplitude spectral line. In view of the problem that the estimation error of Rife algorithm in some frequency intervals is larger than that in other intervals, frequency shift andre-estimation method is employed to further reduce the estimation error in the whole frequency estimation range. Simulation results show that the proposed RootMUSIC-Modified-Rife(RM-Rife) algorithm has better estimation performance than Rife algorithm, RootMUSIC algorithm and Modified-Rife(M-Rife) algorithm under low SNR conditions, and can be applied to related engineering practice.

In recent years, radar-based gesture recognition technology has been widely used in industry and life, and more complex application scenarios also put forward higher requirements on the accuracy and robustness of gesture recognition algorithms. A high-precision gesture recognition algorithm based on millimeter-wave radar is desgined. By comparing the existing classification algorithms, a Convolutional Neural Network-Long Short Term Memory(CNN-LSTM) deep learning algorithm model is constructed for gesture recognition. At the same time, the Blackman window is employed to suppress the problem of spectrum leakage in gesture signal processing, and efficient clutter suppression and data enhancement is achieved through the combining of wavelet threshold and dynamic zero-padding algorithm. The actual measurement results show that the proposed gesture recognition algorithm achieves a correct classification rate of 97.29%, and can maintain a good recognition accuracy rate under different distances and angles with very good robustness.

In order to accurately predict the potential safety hazards of the Yellow River diversion tunnel, a fuzzy Bayesian safety evaluation model based on fuzzy set theory and subjective Bayesian method is proposed. Combined with the actual inspection data, this model uses Bayesian Network(BN) to process fuzzy information and uncertain information, fuzzies polymorphic fault events from the inspection data, and then finds out logical relationship among different fault events through conditional probability table of Bayesian network. Finally, the overall evaluation of the headrace tunnel safety model is obtained, and the feasibility of this method is verified by example analysis of the headrace tunnel safety evaluation model.

In order to achieve the integrated function of absorption/transmission/reflection of airborne antenna covers, a multifunctional metasurface based on graphene and Positive Intrinsic- Negative(PIN) diode is presented, which bears the function of wideband absorption/refection/ transmission. When the PIN diode operates in forward biased condition, the transmittance of the metasurface is close to 0, and the reflection can be adjusted by changing the Fermi energy of graphene. When the Fermi level of graphene is 0.3 eV, the absorption of the metasurface is more than 90% from 1.55 GHz to 5.51 GHz, and the reflection coefficient is 45% at this band with the graphene energy level of 0eV. On the contrary, with the PIN diode acting in reverse biased condition, a transmission peak appears at 4.76 GHz with the transmission coefficients of 61% and 71% respectively when the Fermi level of graphene is 0.3 eV and 0 eV. Due to its advantages of having a simple structure, wideband absorption, and switchable refection/transmission, the proposed metasurface shows the potential in the fields of stealth of communication equipment and radar.

Based on silicon Micro-Electro Mechanical System(MEMS) technology, an X-band four- channel 3D integrated T/R module suitable for array antenna system is designed. The module is stacked by three layers of chips connected to each other by Through Silicon Via (TSV) and Ball Grid Array(BGA). The module integrates functions such as 6-bit digital control phase shifting, 6-bit digital control attenuation and serial-to-parallel conversion, negative voltage bias, power modulation. The dimension of the module is 12 mm× 12 mm× 3.8 mm. The test results show that in the X-band, the saturated transmitting power of the single channel reaches 30 dBm, the gain of transmitting channel reaches 27 dB, the gain of receiving channel reaches 23 dB, the noise figure is less than 1.65 dB. The module bears excellent performance and high integration, which is suitable for mass production.