With the improvement of frequency and bandwidth of fuse, the accuracy, target identification ability and anti-interference ability of detector are correspondingly improved. Therefore, the research on D-band detector for active terminal-sensitive projectile is carried out. Based on 120 GHz transceiver chip TRA_120_001, a D-band detector is designed. In order to improve the detection distance, a semi-spherical lens antenna is fabricated. The experimental scene of rotating scan falling with angular reflector as the target is built. In the experiment, the detector sets the target direction, gives the trigger signal, realizes the detection function and verifies the ability of range measurement.

The quasi-optical transmission characteristics of high-order mode of high-power gyrotron oscillator are analyzed based on geometrical optics and coupled-wave theory, and the simulation code of quasi-optical mode converter with perturbation waveguide is compiled. The design method and results of Denisov launcher in 140 GHz gyrotron oscillator are given. The power distribution curves of each mode and the current density distribution on the waveguide wall are analyzed. The launcher structure with disturbance length of 60 mm, total length of 136.5 mm and conversion efficiency of 99.1% are obtained. And the results fulfill the requirement of 140 GHz gyrotron. The software of Feko is adopted to analyze the reliability of the results from numerical calculation, which lays technical foundation for the gyrotron in the application of thermonuclear fusion.

Shale is one of the main type of oil and gas reservoirs. The accurate characterization of trace organic matter is a hotspot and bottleneck in oil and gas exploration. Terahertz Time Domain Spectroscopy(THz-TDS) is employed to scan and analyze shale with different crude oil contents(ppm level). The results show that there is a monotonic relationship between THz parameter and the content of organic matter. With the increase of crude oil concentration, the attenuation coefficient of THz amplitude per unit thickness increases linearly. Combined with the theory of effective media, the linear relationship between THz dielectric constant and crude oil content below 200 ppm is determined. Therefore, THz-TDS can be used as an effective method for characterization of trace crude oil in shale, which is of great significance for the improvement of the exploration efficiency of oil and gas resources.

The analysis on the thickness of polyamide plate using Terahertz Time Domain System (THz-TDS) in reflection mode is carried out. The refractive index, one of the optical parameters in terahertz band, is solved through the mathematics model, and its value is 1.88. A kind of polyamide plate sample with four kinds of thickness is designed and the ability of THz-based method to detect defects or foreign bodies in fiber glass is verified by attaching metal plates to the back of fiberglass. By the comparison of traditional method and THz method, the terahertz method has a measurement error between 2.5% and 10%. As the thickness increases, the error tends to increase. The reason about the deviations is analyzed, as well as the systematic factors affecting the thickness measurement accuracy, in order to improve the accuracy of THz thickness measurement system and provide theoretical basis for designing terahertz thickness measurement system in the future.

Aiming at the fifth-generation(5G) macro-cell Mobile-to-Mobile(M2M) communication scenario that requires an efficient signal transmission model, a three-dimensional(3D) cylindrical Multiple-Input Multiple-Output(MIMO) channel is proposed. It reflects the macro-cell propagation environment in the distribution area, and the three-dimensional cylinder is the object of interference in signal transmission. The receiver(Rx) is located at the center point of the cylindrical model, and the transmitter(Tx) is located anywhere outside the scattering area. The received signal consists of the sum of Line-of-Sight(LoS) propagation components and Non-Line-of-Sight(NLoS) propagation components reflected by different objects, making the model fully adaptable to various 5G wireless communication scenarios. In addition, this paper combines the Rice fading channel, uniform distribution, Gaussian distribution, Laplace distribution, and von Mises distribution to study the statistical channel propagation characteristics, that is, the transmitter and the receiver in different directions of motion and spatial Cross-Correlation Functions(CCFs), temporal Auto-Correlation Functions(ACFs), and Doppler Power Spectral Density(PSD) of motion time. The Matlab analysis results of the transmission characteristics of the transmitter and the receiver at different time fit well with the simulation results, proving that the proposed three-dimensional model is feasible for a real 5G macrocellular mobile communication channel.

Direction Of Arrival(DOA) study has always been focused on the accuracy and the computational complexity. While the existing algorithms, which are based on the compressive sensing theory, have advantages over the traditional, there are still problems with high computational complexity and low estimation accuracy because the signal model is on the equal spacing grid. To solve those problems, a partially refined grid method is proposed. The proposed method consists of fission process and learning process, the fission process is to refine angle space by inserting new grid points, the learning process is constantly approaching the direction of arrival. The proposed algorithm takes less time and has higher accuracy under sparse initial grid(the initial interval is 20°).

A direct signals cancellation algorithm for passive radar based on Total Least Square(TLS) is proposed. First, the number and delay of direct signals are obtained by TLS, then the subspace of direct signals is constructed based on the available information of direct signals. Finally, the echo is projected into the sub-space to realize direct signals cancellation. Compared with the traditional extended cancellation algorithm, the direct signal sub-space constructed by this algorithm has low order and high precision, and it still has good clutter suppression performance in the case of short processing time or time-varying amplitude of direct signals. The simulation results verify the effectiveness of the proposed algorithm.

Aiming at the problem that traditional methods cannot estimate the bandwidth of under-sampled Linear Frequency Modulation(LFM) signals under low Signal Noise Ratio(SNR), a bandwidth estimation method based on Distributed Compressive Sensing(DCS) is proposed, which uses the joint sparse characteristics of multiple LFM signals with the same modulation type from the same source to estimate the bandwidth of LFM signal. Firstly, the under-sampled LFM signal model is constructed. Secondly, the LFM bandwidth is reconstructed by DCS algorithm. Then the parameter estimation ability of the proposed method under low SNR conditions is analyzed. Finally, the feasibility and validity of the proposed method are verified by simulation.

The triangulation constraint method is introduced into Synthetic Aperture Radar(SAR) image matching. This method uses the Scale-Invariant Feature Transform(SIFT) operator to generate feature points. After the SIFT feature points are generated, the RANdom Sample Consensus(RANSAC) algorithm is adopted to eliminate the false matching points, and the higher-precision points with the same name are obtained. Finally, the Delaunay triangulation is established by the same-named point obtained by the SIFT algorithm in the similar triangulation of the same name, the triangle center of gravity is used as the interpolated virtual name of the same name. A Normalized Cross-Correlation(NCC) constraint is performed on the virtual point of the same name, and the virtual point-name pairs that do not meet the threshold requirements are eliminated. Then, a new triangulation network is established according to the virtual same-named point obtained by the interpolation, and the triangulation network is dynamically updated to obtain more virtual names with the same name until the matching requirement is met. The experimental results show that the proposed method can effectively increase the number of matching feature points and improve the matching accuracy of radar images.

A precision analysis method of four-satellite Time Difference Of Arrival(TDOA) location based on Geometric Dilution Of Precision(GDOP) is proposed. Firstly, the 3D location precision model based on TDOA is investigated. Then, the influences of the time measurement error and location error on the above precision model are analyzed. Finally, the GDOP of representative four-Satellite of Y, T, square and anomaly are simulated. The influences of the length of baseline, the satellite and target altitude on location precision are analyzed. The simulation results show that four-satellite location of Y distribution based on TDOA achieves the best precision. The increase in the length of baseline and satellite altitude contributes to improving the precision of location.

In order to address the performance degradation when coherent sources, array perturbations and finite-sampling exist, a robust adaptive beamforming using Iterative Adaptive Approach(IAA)-based interference-plus-noise covariance matrix reconstruction is proposed. Firstly, the method uses the IAA to estimate the accurate power spectrum, and further reconstructs the interference-plus-noise covariance matrix with the power estimated. The method reduces the integral region to the three-dimensional domain to reduce the influence of noise, which improves the reconstruction precision of the interference-plus- noise covariance matrix. Finally, the proposed method suppresses the interference signal by beamforming. Since IAA does not rely on the non-coherent assumptions of the signal, it could estimate the high accuracy of DOA and power when the coherent interference exists. Computer simulations show that, compared with other beamforming methods, the output Signal to Interference plus Noise Ratio(SINR) can achieve the optimal output under coherent interference, array perturbations and finite-sampling situation.

To meet the requirements of 5G communication systems, the traditional scenarios are extended by assuming a general multi-hop cooperative wireless transmission scenario, and security analysis is performed in this scenario. The security capacity of the proposed scenario is derived and the key performance evaluation metric in integral form is obtained. The simulations based on Monte Carlo method are carried out to verify the correctness of the analytical results regarding security capacity. Through the observation and discussion on numerical simulations, a series of conclusions pertaining to the proposed multi-hop cooperative relay wireless communication network are given.

The use of mobile Unmanned Aerial Vehicle(DAV) to collect data from Wireless Sensor Networks(WSNs) has attracted great attention recently. It is often necessary to relate the stream of sensed data to the deployed location of the data producing sensor nodes. Therefore, UAV-based Robust Node Localization(UAV-NL) algorithm is proposed. In UAV-NL algorithm, the location of UAV is unknown. Each sensor node receives beacon packets transmitted by the UAV at random positions and records an Received Signal Strength Indicator(RSSI) vector. The norm distance of two RSSI vectors, which is theoretically proved to be linearly related to the distance between two nodes, is used for ranging. Distance calculation method is theoretically derived by using RSSI measurements. Location of node is estimated by Semi-Definite Programming(SDP) algorithm. Extensive simulations in different environments validate high localization accuracy of the proposed algorithm even under noisy channel conditions.

Traffic Management System(TMS) is proposed to improve the traffic efficiency and minimize traffic congestion problems. Such applications are supported by Vehicular Ad Hoc Networks (VANETs). However, the data dissemination in VANETs is a challenging task due to the short-range communication and node mobility. Therefore, Metrics of Complex Networks based Data Dissemination (MCDD) scheme is proposed in this paper. In MCDD scheme, each vehicle has contextual knowledge of 2-hop neighbor to select the relay node based on two complex networks metrics, the betweeness centrality and degree centrality, in order to decrease the overhead and delay. Compared with Fully-distributed Traffic Management System(FTMS) based on districts, vehicle traffic congestion time of MCDD scheme is reduced by 18.11%, and the speed of vehicle is increased by 8%.

A high-efficiency broadband receiver based on broadband matching network is designed in order to solve the problems of intermodulation interference, blocking and saturation caused by strong electromagnetic interference at the front end of RF signal receiving system. The operating frequency of the receiver module covers the entire S-band, and it adopts a 4-wire SPI interface to complete link control according to the instruction of the host computer. The 7th-order Butterworth low-pass filter design is utilized to ensure the best flatness and port standing wave ratio in the band. The measured results show that the receiver module realizes signal amplification and down-conversion of 100?kHz-6?GHz, and outputs with intermediate frequency of 140?MHz and 100?kHz-19.999?MHz. It bears the internal and external 10?MHz reference time base automatic switching function. This module provides a single-ended 10?MHz clock for baseband board for baseband digitization. It shows that the optimized design of the RF receiving front end can achieve high selective tracking and pre-selection of signals in the target detection frequency band, so that the receiving system can effectively suppress the strong interference signal.

Low-frequency Electro-Magnetic(EM) waves have low transmission loss, but the traditional resonant antennas have large size at these frequencies which results in difficulties in engineering application. Convection current can generate time-varying electromagnetic fields, and the operational frequency is independent of the antenna size, which gives a new method to produce low-frequency EM radiation. In this paper, the electromagnetic fields and the equivalent model of the vibratory electret are studied through theoretical calculation and simulations. Some sound explorations are conducted to realize low frequency miniature antenna based on the vibratory electret. The research results show that the vibratory electret shares the same electromagnetic characteristics with the current element.

In order to improve the identification of transient electromagnetic system echo, a new waveform similarity algorithm is designed to analyze and process the received transient electromagnetic system echo data. Aiming at the fact that the sampled echo data is difficult to distinguish and identify due to various interference factors, and the traditional echo processing method is not ideal, this paper takes the sampling point as an evaluation index from the perspective of space vector, sets the index weight by using entropy weight method, synthesizes various evaluation indexes, and proposes a new waveform similarity algorithm. Through a large number of experimental data analysis, the algorithm has achieved good results in the identification of echoes, and the identification accuracy of partial thickness has reached 90%, which has good practical value.

Electro Magnetic Pulse(EMP) signals are generated in laser shooting experiment. The output signals of electric field probe utilized to measure these EMP signals are differential signals, which need to be integrated before the measurement results can be obtained. Because of the bandwidth limitation，analog integrator cannot meet the measurement requirements, the integration can be completed only through digital integrator. In this paper, the digital integrator is designed based on the wavelet transform and Fourier transform. The wavelet soft threshold method and the low frequency zero drift removal method are adopted to deal with the high frequency random noise and zero drift error of differential signals. After research and analysis, it can integrate the EMP signals whose rise time is less than 100?ns, and has achieved good results.

The ability of outdoor power distribution equipment to protect against the Lightning Electromagnetic Pulse(LEMP) is studied by XFDTD full wave 3D electromagnetic field simulation software based on Finite Difference Time Domain(FDTD) method, including simulating the coupling of the near-field lightning electromagnetic pulse with the cabinet of outdoor box-type substation, which is characterized with slots. The coupling electric field strength and the electromagnetic power density of the three types of substations with different shapes of aperture array are obtained at 30 m from the return point, and at 5 cm, 50 cm and 100 cm from the opening position. It is concluded that when the LEMP whose energy is mainly concentrated on the low frequency band(below 2 MHz) couples on the anterior aperture array, the coupled electric field energy in the substation is mainly focused near the opening. The electromagnetic energy coupling into the substation with rectangular, square and circular aperture arrays decreases in turn, and the substation with circular aperture array has optimum shielding effectiveness. The electromagnetic hazard in the cavity far away from the aperture array is less than that near the opening position. The coupled electromagnetic energy will have impacts on the normal operation of the substation, but will not damage its internal electronic components.

Many image fusion methods, such as large image features to fuse image coefficients, ignore the correlation of edge features between images, resulting in block phenomenon and discontinuity in the fused image. This paper designs an image fusion algorithm with edge feature weighting based on the second generation of curved wave transform. Firstly, the input image is decomposed into low frequency and high frequency coefficients by Second Generation Curvelet(SGC) transform. Then, the region energy model is employed to calculate the energy features of the image, and the mean value model is adopted to calculate the brightness features of the image. By using the energy and brightness features of the image, the fusion function of low-frequency coefficients with two features is constructed to obtain the fusion low-frequency coefficients with excellent energy and brightness features. Sobel operator is introduced to detect the edge features of the image, and the fusion function of high frequency coefficients weighted by the edge features is constructed based on the detection results to obtain the fusion results. Finally, the fusion experiments are carried out with the algorithm. The experimental results show that the proposed algorithm has better fusion effect than current algorithms, and the edge continuity and clarity of the fusion image are better.

Multi-Chain Spheroidal Structure Universe Algorithm(MCSSUA) is proposed in order to improve the quality of tomato image surface defect segmentation. Firstly, multi chain spheroidal universe structure topology is established, and the universe can only communicate with the three universes around itself to maintain the balance of the universe. Secondly, the optimal threshold is adopted to segment the tomato defect image, and a separate element is taken as the stop condition for ending the image segmentation. Finally, the flow of the proposed algorithm is given. The simulation shows that the multi-chain spheroidal structure universe algorithm can segment the tomato image surface defect with clear results, and higher correct segmentation rate of various defects than other algorithms.

Due to the phenomena of huge illumination changes, partial occlusion and damage between images, robust and efficient image alignment is still a challenging task. An improved online image alignment algorithm is proposed. Firstly, the Principal Component Analysis(PCA) of image gradient orientations(IGO) is utilized to provide a lower dimensional subspace which is more reliable than the intensity of pixels. The alignment is sought in the IGO domain such that the aligned IGO of the newly arrived image can be decomposed as the sum of a sparse error and a linear composition of the IGO-PCA basis learned from previously well-aligned ones. The image alignment problem can be modeled as a norm minimization problem. The problem is relaxed to a convex optimization problem in this paper, and a convex optimization algorithm based on multiplier alternating direction method is proposed. IGO-PCA basis are adaptively updated based on incremental singular value decomposition considering the migration of IGO mean in this paper. The effectiveness of the proposed algorithm is validated on a large number of challenging data sets. The experimental results show that compared with the current typical SIFT algorithm, Robust Alignment by Sparse and Low-rank decomposition(RASL) algorithm and transformed Grassmannian Robust Adaptive Subspace Tracking Algorithm(t-GRASTA), the alignment effect of the proposed algorithm is the best, and it has the strongest robustness to illumination changes and occlusion phenomena of images.

Improved hill climbing algorithm is adopted in order to improve the subpixel positioning effect of micro-parts. Firstly, the focusing window is obtained by expanding the target area outward, which tracks the drift image, the boundary is determined by image distance and view angle. Secondly, the optimal convergence solution is obtained by using the optimized hill-climbing search algorithm, two-dimensional image information entropy is constituted by the focusing image evaluation function. Thirdly, modulus of higher moment of improved Zernike moment algorithm replace the edge parameters, which cut the computation amount. The convolution window matrix of difference of Zernike moment is constructed to improve the positioning accuracy. Finally, the flow of the algorithm is given. The experimental results show that the Zernike moment-even template has smaller subpixel positioning error than odd template edge method does, the improved hill-climbing algorithm is decreased by 43.24%,21.62%,32.43%,27.03% and 56.76% respectively compared with spatial moment algorithm, polynomial fitting algorithm, Zernike moment algorithm, region growing algorithm and template matching algorithm about the mean of positioning error of regular shape; the improved hill-climbing algorithm is decreased by 39.02%,21.95%, 26.83%,24.39% and 51.22% respectively compared with spatial moment algorithm, polynomial fitting algorithm, Zernike moment algorithm, region growing algorithm and template matching algorithm about the mean of positioning error of irregular shape. The positioning accuracy of the improved hill-climbing algorithm is higher.

Surgical smoke generated during the minimally invasive surgery is normally regarded as an important aspect of surgical risk which could produce potential health problems for both clinicians and patients. Current researches in the area of surgical smoke simulation have limitations with regards to both accuracy and interactivity. In this paper, a novel approach based on the vortex particle method which combines the conventional fluid model and the meshfree method has been proposed. This method reconstructs the lost information during the simulation due to numerical dissipations by using a vorticity confinement force, and hence, successfully recovers the fluidic details including small-scale vorticities. GPU acceleration and image reconstruction based on medical images have also been employed to produce a patient-specific simulation. The results have shown to be high realistic and accurate, computational optimization and extensibility have also been achieved. Therefore, such simulation can be used in potential applications including surgical training and planning.

The inaccurate control of the valve positioner in industrial process control would cause the valve position to reciprocate and fluctuate. To solve this problem, a circuit board of pneumatic intelligent valve positioning control is designed. The hardware design is based on micro-processing STC12C5A6052 and outputs bipolar PWM(Pulse Width Modulation) pulse to drive the piezoelectric valve. The peripheral circuit includes power conversion circuit, I/V conversion circuit, A/D sampling circuit, button and liquid crystal display circuit, etc. Because the valve control and execution process is nonlinear time-varying, it is difficult to establish an accurate mathematical model. Therefore, a fuzzy PID(Proportion Integration Differentiation) control algorithm is utilized instead of the conventional PlD control in the control system. By improving the designs in software and hardware, the valve position fluctuation can be effectively avoided. The stable transition time is only 1.2?s and the system has no overshoot. In Matlab, the transfer function of the mathematical model and the measured data of the valve position are simulated by various mechanisms of the valve. Through waveform comparison, the control circuit board of designed valve positioner can make the valve control more stable, fast and accurate.

Aiming for the depletion of node energy consumption caused by the simplification of Ad hoc On-Demand Distance Vector(AODV) routing metrics, and the resulting network regionalization and high latency problem of path congestion, this paper proposes a routing protocol with multi-metric path selection mechanism—M-AODV. Based on the AODV routing protocol, in the route discovery phase, a node with higher energy and stronger signal strength is selected as a routing node according to the set energy threshold and the received signal strength threshold. In the path selection phase, firstly, according to the received signal strength, congestion degree and hop count of the physical layer, the stability value of each link is calculated; secondly, the stable value of each path is accumulated to obtain the stable value of the entire path; finally, the path with the largest stable value is selected as the optimal path. Simulation experiments are performed by using NS2.35 simulation software. The results show that the end-to-end delay of improved routing strategy is reduced by nearly 30% compared with that of AODV. It is superior to traditional AODV and I-AODV-SE routing algorithms in terms of packet transmission rate, throughput and node survivability.

Forwarding Area-Tune Energy Efficient Geographic(FAEG) routing is proposed to solve the problem of data transmission in Wireless Sensor Networks(WSNs). In FAEG, the forwarding region of packet is controlled, and forwarding node set is constructed. The node with big weight is selected to transfer packet. By calculating the weight based on both the remaining energy of the node and the distance from the destination node, the remaining energy of the node is large and the node close to the destination node is qualified to firstly deliver the packet, which improves the routing stability. When a node encounters a routing void, the forwarding area is adjusted to select the node transmitting the packet in a wider area. Simulation results show that the proposed FAEG routing reduces energy consumption and improves packet transmission rate.

Large-scale net flow training data sets are inevitable requirements for building high-quality, highly stable network traffic classifiers. However, with the increase of the feature dimension of the network stream and the expansion of the data set size, neither the analysis processing of the network stream nor the training of the classifier model based on Support Vector Machin(SVM) can obtain effective processing results in effective time. A distributed and parallel large-scale network flow based on Hadoop cloud computing platform is proposed. Distributed learning and training of SVM network traffic classifier is implemented by MapReduce technology on Hadoop cloud computing platform, and CloudSVM network traffic classifier is constructed. Through the distributed storage and processing of trace files of large-scale network traffic from the campus network export mirror, the sample data sets are classified, and the distributed storage and parallel processing of large-scale network data based on Hadoop platform is experimentally verified. The high efficiency of the set also verifies that the CloudSVM classifier can quickly converge to the best without reducing the accuracy of the classification, and with the increase of large-scale network flow samples, the training time of the SVM classifier is approaching constant.

For secure routing of Wireless Sensor Networks(WSNs), Energy-aware and Secure Multi-hop Routing(ESMR) is proposed. ESMR protocol increases the performance of energy efficiency with multi-hop data security against malicious actions. The network field is segmented into inner and outer zones based on the node location. Furthermore, in each zone, numerous clusters are generated on the basis of node neighborhood vicinity. Secondly, the data transmission from cluster heads in each zone towards the sink node is secured by using the proposed efficient secret key sharing scheme. The experimental results demonstrate the efficacy of proposed ESMR in terms of network lifetime, network throughput, energy consumption, average end-to-end delay in comparison with the existing work.

In order to improve the encryption effect and ensure the security of the data set, this study designs a privacy data set homomorphic encryption method based on the improved Rivest Shamir Adleman(RSA) algorithm. Based on the analysis of the operating principle of traditional RSA algorithm, the parameter selection and the prime number judgment condition, the improved RSA algorithm is proposed. In order to further improve the encryption speed, Data Encryption Standard(DES) algorithm is introduced. Firstly, DES algorithm is utilized to encrypt the plaintext data set, and RSA encryption is carried out for the key. On this basis, in the plaintext and ciphertext spaces, the addition homomorphism process is adopted to calculate the ciphertext, and the corresponding plaintext calculation result is obtained by decrypting the result. Experimental results show that, compared with the encryption method based on traditional RSA algorithm or DES algorithm, this method has higher encryption efficiency and higher success rate of resisting attacks. The encryption process takes time between 5 and 6?s, and the success rate of resisting attacks is kept around 95%, indicating that this method can effectively provide support for the security protection of private data sets.

Underwater Sensor Networks(USNs) technique is an effective way to develop water environment. However, when data is transmitted in the underwater environment, the signal attenuation is fast and the quality of communication link is poor, which reduces the packet transmission rate. Therefore, Link Quality and Location-based Cooperative(LQLC) routing is proposed. It refers to the idea of cooperative routing, which first uses the link quality and node energy information to build candidate forwarding node set, and then selects the node with the best link quality as the forwarding node from the candidate forwarding node set. The nearest node to the forwarding node is selected as the cooperative node in the candidate forwarding node set. Simulation results show that LQLC protocol has a good performance in terms of packet delivery ratio.

The new Package on Package(PoP) structure of memory is different from the conventional packages, resulting in current Destructive Physical Analysis(DPA) methods not fully applicable to the new PoP memory. In this paper, the structure and typical defects affecting the reliability of a new PoP package memory are analyzed by using 3D-X-ray, metallographic slices, stacked chip separation and internal inspection of non-top chip. A comprehensive DPA scheme with strong applicability and high efficiency is proposed, and the effectiveness of evaluating the reliability of the new PoP package memory is verified through an example. It can also provide the basis and help to the revision of subsequent standards and DPA of other advanced package devices.

A novel dual-channel monolithic circuit working at Q-band is reported which integrates the functions of Wilkinson power divider, digital attenuator and Single Pole Double Throw(SPDT) switch. Three-level parallel structure is adopted in the SPDT switch unit during the design process which can reduce insertion loss and improve isolation effectively. A kind of multi-section parallel-T structure is used in the digital attenuator unit for improving insertion loss and attenuation accuracy in high frequency band. By means of the above design，a monolithic circuit with 10?dB insertion loss and 26?dB isolation is produced successfully, which has the amplitude control function of 0.5?dB step to achieve 31.5?dB maximum attenuation.

Ultrasonic Test(UT) technology for dielectric defects of Multilayer Ceramic Structure(MCS) is studied. Firstly, UT is proved to be applicable and effective to voids and delamination through thin layer reflections theoretical calculating on Multi-Layer Ceramic Capacitors(MLCC) structure, as low UT frequency can get very strong reflection signal at extremely thin aperture. Simultaneously, suggestion of C-Scan testing parameters including equivalent focal length, surface wave time, data gate and gain is given. Then, the practical UT test of MCS indicates that 50 MHz is the optimum screening frequency. Finally, defect determining and locating is accomplished via A-Scan and transit time respectively, the conclusion is also verified by sectioned UT sample preparation and detection.