Abstract：The stability problem of memristor may affect the performance of memristive neural network. In order to explore it, a memristive back propagation(BP) neural network, in which the memristors are the synapses, is constructed based on the equivalent resistance topology memristor model. And it’s trained and tested on the MNIST dataset. The stability problem of memristor is simulated by setting fluctuations of the parameters in the model. Finally, it is found that the performance fluctuation of the memristor to a certain extent will promote the convergence of the neural network, but the excessive fluctuation will reduce the convergence speed of the network. To characterize this criticality, the maximum fluctuation range of each parameter in the model is obtained. Meanwhile, the fluctuations’ ranges of device parameters are obtained by tracing back to the memristor model. The result provides a reference for the fabrication and selection of memristor devices during hardware application.

Abstract：When high resolution range profiles(HRRP) are used to recognize radar target, few traditional recognition methods analyse the sparseness of HRRP samples. In order to overcome the tedious analysis problems and simplify the recognition procedure, sparse representation is an effective way to compress HRRP samples and extract the target features. Thus, a federated redundant dictionary and a fast sparse representation algorithm are introduced to implement radar target recognition here. Moreover, a sparse decomposition parameter is adjusted by SNR in order to suppresses noise. The simulation results show that compared with the same kind of RATR algorithms, the algorithm in the paper is practicable, simple and efficient. In contrast to the traditional dimension reduction recognition method, it has better noise robustness and higher recognition ratio.

Abstract：The ambient noise at any two different sensors, in practice, may correlate with each other, and the spatial distribution of ambient noise intensity may be directional because of wind and shipping noise sources. A linear harmonic noise model is considered in this paper to appropriately describe the ambient noise. By applying l2-norm penalization of fitting the source covariance model to the estimated spatial covariance and the linear harmonic noise model, super-resolution direction of arrival(DOA) estimation algorithm based on sparse spectrum fitting is proposed. Then, the influence of the regular parameters and the number of linear noise model on the performance of the algorithm is discussed by computer simulations. The performance of the proposed algorithm is verified by the data processing of the sea trial data.

Abstract：This paper introduces a way to realize the supervised neural network algorithms based on memristive characteristics on Field Programmable Gate Array(FPGA) for the problem that how to take the memristors into artificial neural networks and hardware implement. This design uses memristors module as weight store module in neural network to build supervised learning with error feedback mechanism. The memristive neural networks are used in pattern recognition and their hardware resource and processing speed are optimized. Experiment results show that the performance of pattern recognition is quite good. Further, the hardware resource occupancies and training time are 11 773 logic elements(LEs) and 0.33 ms on Cyclone II：EP2C70F896I8, respectively, and the test time of images is 10 μs, which gives a useful reference for combination of memristors and neural networks.

Abstract：In ultrasonic measuring distance technology, traditional signal processing algorithms are difficult to analyze signals with long distances and weak echoes, and it is also impossible to measure the distance of multiple obstacles at the same time. In order to solve this problem, an ultrasonic measurement distance method based on random forest is proposed in this paper. Firstly the time domain characteristics of the signal—the relative peak amplitude and frequency domain characteristics—the relative area of the spectrum are extracted to judge the number of obstacles in the detection area by random forest method, and then calculates distance of obstacles. After the experiment, the algorithm can effectively measure the distance of multiple obstacles within 10 meters, and the measurement error is within 3 cm, which satisfies the practical application and has high practical value and theoretical reference significance.

Abstract：Content Delivery Network(CDN) has recently become one of the most popular research fields. To reduce the user access latency and improve the quality of service the network, the content prefetching technology is introduced based on the CDN cache technology in related studies. The prefetching technology can overcome the lag time between the demands for user request and content caching, and improve resource utilization of the CDN. In this paper, the classification of CDN prefetching techniques is summarized from the perspective of the different objects in prefetching. The advantages and weaknesses of different classifications are summarized and compared between each other, and various evaluation criteria are clarified in detail. Finally, the research direction of content prefetching in CDN with new networks and technologies is pointed out.

Abstract：In the process of approaching the landing, the instrument landing system(ILS) is vulnerable to the external environment and airspace, resulting in the problem of reduced navigation accuracy. This paper proposes an inertial navigation system(INS) and GBAS landing system(GLS). The improved combined navigation algorithm uses the difference between the output position information of the integrated navigation system as the measured value of the improved unscented Kalman filter(UKF) of the BP neural network, and obtains the global optimality estimated value of the system through the optimal weighting method. Compared with the traditional federated filtering algorithm, the proposed algorithm can effectively reduce the measurement noise, reduce the error when the aircraft approaches the landing, and improve the navigation accuracy.

Abstract：With the development of computer technology, fire image processing technology combining computer vision, machine learning, deep learning and other technologies has been widely studied and applied. Aiming at the complex preprocessing process and high false positive rate of traditional image processing methods, this paper proposes a method based on deep convolutional neural network model for fire detection, which reduces complex preprocessing links and integrates the whole fire identification process into one single depth neural network for easy training and optimization. In view of the problem of fire detection caused by similar fire scenes in the identification process, this paper uses the motion characteristics of fire to innovatively propose the combination of fire frame position changes before and after the fire video to eliminate the interference of lights and other similar fire scenes. After comparing many open learning open source frameworks, this paper chooses Caffe framework for training and testing. The experimental results show that the method realizes the recognition and localization of fire images. This method is suitable for different fire scenarios and has good generalization ability and anti-interference ability.

Abstract：In order to improve the system of full digital phase-locked loop speed, reduce the power consumption of the system, and at the same time improve the dynamic performance and steady-state performance of phase-locked system,this paper proposes a full digital phase-locked loop based on assembly line. The electronic design automation technology is used to complete the design of the system, and the designed circuit is simulated and analyzed by computer. Simulation results show that the parameters of the digital filter in the phase-locked loop can be dynamically adjusted according to the magnitude of phase error, which can not only speed up the phase-locked speed, but also enhance the stability of the system. The integrated circuit optimized by pipeline technology can reduce the system delay and reduce the total power consumption. The phase-locked loop can be embedded as a functional module into SoC（Systerm on Chip）and has a wide range of applications.

Abstract：A 13～16 GHz bi-phase voltage variable attenuator(VVA) based on 0.25 μm gallium arsenide pseudomorphic high electron mobility transistors(GaAs pHEMT) process is presented. Balanced structure that adopt in this circuit provides less input output return loss. A cascade of T-attenuator and Pi-attenuator topologies are adopt in attenuator part. Parasitic reduce and linearity enhancement are achieved by using stacked double-gate switch transistor structure in shunt arms of T-attenuator and Pi-attenuator topologies. Simulated results show that in the range of 13-16 GHz, the input output return loss is less than -14 dB, the insertion loss is 12.5 dB, and the dynamic range is more than 20 dB, the input 1-dB power compression point(P1dB) is over 30 dBm while the chip area is 1.8 mm×1.2 mm.

Abstract：This paper presents an ultra-wideband(UWB) microstrip patch antenna with a simple pentagonal geometry and two notches. The UWB characteristics is achieved by using defective ground and impedance conversion methods, yielding a frequency band coverage from 3.1 GHz to 10.6 GHz where the voltage standing wave ratio(VSWR) is no more than 2. A rectangular slit and a narrow U-shape slot are etched in on the patch and feed-line respectively to achieve notch performance at WLAN band(5.15 GHz~5.85 GHz) and X-band communication downlink band(7.25 GHz~7.7 GHz). The antenna presents good radiation characteristics in its operational UWB, and measured results are in good agreements with the simulation, which accordingly validates the design.

Abstract：This paper proposed a hardware Trojan detection method based upon XGBoost(eXtreme Gradient Boosting) model by using the analysis results of ring oscillator network, and used the cross-validation method to optimize the model. It can utilize that train sample dataset to build the XGBoost classification model, and use the supervised learn mode to classify the data, thus realizing the separation of the original circuit and the Trojan circuit. Using RS232-T100 and RS232-T800 as Trojan circuits, the FPGA experiment was carried out. The experimental results showed that the detection rate of the Trojan data with RO at 0.1 ms integration time is 100% and 99.20%, which verified the validity of the method. In addition, when compared with traditional methods and other machine learning methods, the XGBoost-based detection method shows a higher detection rate, and can analyze the characteristic importance of the multi-dimensional vector correlation data instead of dimensionality reduction. It can maximize the key features required for Trojan detection.

Abstract：The method of FIR filter coefficient quantization has influence not only on the filter performance, but also on the circuit system performance. Normal filter coefficient quantization method is analyzed, and a new method of filter coefficient quantization based on filter gain is proposed. A low pass filter based on the new quantization method is designed with MATLAB and implemented on FPGA. ModelSim simulation results show that the new method can improve the output signal′s amplitude and SNR effectively.

Abstract：In order to accurately receive and process large-capacity high-speed data in real time，based on the in-depth study of CPCI Bus，a high-speed data transmission system based on CPCI Bus and LVDS is proposed. PCI9054 is used as the bridge device and FPGA is the microcontroller. The transmission mode combining PCI Target single-cycle mode read/write instruction and DMA burst mode transmission data is designed to improve the data transmission speed while ensuring reliable data transmission.After practice testing，the design work is stable，the transmission is accurate，and the expected transmission task is successfully completed.

Abstract：Epilepsy is a chronic disorder of brain dysfunction caused by sudden abnormal discharge of brain neurons. The detection of epileptic seizure can be achieved by the detection and analysis of spike waves. In this paper, a method of detection based on wavelet transform and modulus maximum is proposed. Firstly, it uses the continuous wavelet transform of epileptic EEG signals in a certain scale to divide the frequency bands. Secondly, it applys the modulus algorithm and refining algorithm to detect singular points of EEG signals, which are taken as the suspect points of spike waves. Finally, through screening based on power spectral density analysis and space surface fitting, the final characteristic spike waves are detected to determine whether the epileptic seizure occurres. The verification experimental results indicate the efficiency and reliability of the proposed method with a diagnostic accuracy as high as 92.5%. It provides a valuable reference method for epileptic seizure detection.

Abstract：A monitoring system of salvaged pontoons based on FameView is designed, which consists of sensors, PLC and upper machine. Sensors and PLC are used for data acquisition, and PC is the monitoring system of FameView configuration software development for remote monitoring data. The system can real-time monitor the pose, pressure and buoyancy parameters of pontoons and display them dynamically. It has the functions of alarm display, historical data storage, web and mobile terminal. In addition, the buoyancy of pontoons is calculated by multi-physical field simulation and algorithm of piecewise curves turned into straight lines, which reduces the error. The system is easy to operate,and has reliable data and stable performance. It has good interactivity, reliability and stability in operation. The system can greatly reduce the labor intensity and difficulty of operators and improve work efficiency, and has good practical application value.

Abstract：A new balancing strategy was proposed to solve the problems of unbalanced batteries and the complexity of the existing battery balancing system. This balancing strategy combines active and passive balancing techniques, but it is different from the existing methods. This equalization strategy uses passive equalization to equalize the battery unit under battery charging and active equalization to equalize the battery pack under static or discharging conditions. It not only simplifies the equalization circuit, saves cost, but also has significant equalization effect. Based on the battery pack of electric bicycle, the experiment divides the battery pack into two groups. One group simulates the unbalanced battery pack with high current discharge, and the other group is in normal use.The experimental results show that the equalization system is feasible, and the battery voltage after active and passive equalization is basically restored and can be used normally.

Abstract：With more tasks physical layer needs to handle in 5G, the interaction between the physical layer and the high layer and the under layer is greatly increased. The confusing instruction causes the task conflicted becoming more serious based on the instruction triggering task mechanism. In order to improve the interaction efficiency and robustness of the system, the physical layer is divided into the null state, the cell selection state, the idle state, the random access state and the connected state according to the boot process of the equipment, and their physical layer tasks are designed in each state. Physical layer controls the physical layer status and tasks to realize interaction between the high layer and the under layer，so as to realize communication between the terminal and the base station. The physical layer of the terminal can demodulate or decode correctly under the parameter configuration of 5G TDD, bandwidth of 80 MHz, and subcarrier spacing of 30 kHz and refuse processing under the abnormal conditions.

Abstract：A new LTE and 5G BS microstrip antenna is designed. The antenna not only enhances the impedence bandwidth and radiation gain，but also creates low-profile height by loading a new EBG structure. The height of the antenna is reduced from 40 mm to 22 mm. The impedence bandwidth of the antenna is 1.35 GHz, from 1.65 GHz to 3 GHz, and radiation gain at 2.1 GHz reached 7.93 dBi. The result shows that it can achieve low-profile design by EBG loaded on patch antenna.

Abstract：Aiming at the problem that the communication environment is complex and the peak-to-average ratio of the baseband receiving signal of dPMR digital intercom is too large, a strategy for automatic gain control of signals processed by Gaussian channel and Rayleigh channel is proposed. The control strategy adopts feedback DAGC. The structure is optimized based on the moving average detection algorithm and the LSLDAGC gain control algorithm. The signal decision module and the finite gain adjustment mechanism are used to improve the system stability of the processing of the fading signal, and the signal misjudgment rate is reduced. The simulation results show that the AGC system can control the gain of stationary and fading signals with a large dynamic range of 96 dBm, a convergence time of less than 6.25 ms and a sensitivity of -105 dBm. At the same time, it has good stability and tracking performance when the signal power fluctuates.

Abstract：The new multiple access technology plays a very important role in 5G mobile communication system. In this article, a new non-orthogonal complex space division multiple(CSDM) technique is proposed and combined with traditional multiple input multiple output(MIMO) system to form CSDM-MIMO system. The CSDM-MIMO system can transmit multiple user data on one antenna at the same time and same frequency, which improves the data transmission rate and system capacity. The system special virtual channel is estimated by least squares(LS) algorithm and minimum mean square error(MMSE) algorithm respectively. At the same time, the optimal weight of the system is searched by the bit error rate(BER) fed back by the receiver and artificial fish swarm algorithm(AFSA).

Abstract：For 3×3 X channels, the theoretical degree of freedom of 9/5 can only be achieved by asymptotic interference alignment, which requires infinite resources to approximate, which is impossible to implement in practice. In a single antenna node scenario, the interference alignment scheme can be designed in the time domain. In order to minimize the transmission delay, it is considered to implement the transceiving processing of the message by increasing the cost of one extra time-slot. The 3×3 X channel message transmission is realized by reasonably scheduling the transmission time slot of the message and matching the propagation delay between the selected transmitter and the receiver. The specific transmission and reception schemes are given by the representation method of the cyclic polynomial, and the degree of freedom reaches 3/2. And comparing the total number of successfully transmitted messages, the result transmission efficiency is better than the time division multiple access scheme. Further, the feasibility conditions of the scheme in Euclidean space are derived, and the range requirements of each node arrangement are indicated.

Abstract：The system adopts the improved data encryption algorithm(DEA), and the hardware encryption system is constructed by complex programmable logic device(CPLD) chip EPM570. After frequency modulation, the acoustic signal is encrypted by key and sent to the receiving end. The receiving end is filtered and the original signal is restored after key decryption, so that the close-range encrypted transmission can be realized. The hardware circuit is divided into two parts: the sender and the receiver, which are composed of the main controller, CPLD, direct digital synthesizer(DDS),filter and reed-solomon(RS) codes module circuits. Experiments show that the system has high security and can be applied to the transmission of confidential or sensitive signals.

Abstract：The existing robot path planning(RPP) algorithms have the problems that the parameters are complexity. To solve this problem, this paper proposes a self-learning ACO(SlACO) algorithm for robot path planning. In SlACO, an improved grid map(IGM) method is used for modeling the working space and the 8-geometry is used as the moving rule of ant individuals. The strategy of multi-objective search is used for the whole ant colony. The SlACO has the feature that the whole algorithm only need set one computing parameter. Simulation results indicate that the SlACO algorithm can rapid plan a smooth even in the complicated working space and its efficiency is better than existing RPP algorithms.

Abstract：In order to achieve intelligent real-time target detection in high-speed scenes, a high-speed camera platform is designed based on FPGA of ZYNQ7000 and target detection algorithm is implemented on it. As a result, one prototype of integrated high frame rate real-time target detection system was created. This system connects the high-speed CMOS signal to the FPGA directly, doing data processing and implementing target detection algorithm through native FPGA, and outputs real-time target position sequence eventually. This system performed pipeline operation of background differencing, binarization, and centroid detection on high-speed video stream images by using a pipeline structure of FPGA design. Due to this operation, the system achieved the synchronization of image acquisition and target detection. The experimental results show that this system can achieve real-time target detection and output target position greater than 1 100 f/s at 560×480 resolution, meanwhile, the detection accuracy can be up to 3 pixels.

Abstract：In order to improve recognition accuracy of iris recognition，the iris image was disposed,as the result the iris region of image was accurately located and the normalized image was enhanced. The Haar wavelet transform was used in the feature extraction and K-means was used to cluster the feature,so that a small sample set of iris images was obtained.Combined with iris texture characteristics,the features were extracted by using Log-Gabor filter and the iris feature template was formed after quantization coding. The similarity of iris feature template was calculated by Hamming distance in small sample set, and the iris recognition was completed. The testing results illustrate that the proposed algorithm has a certain improvement in recognition accuracy and effectively avoids the problem of large amount of computation and long time in iris matching because of the variety and quantity of iris database.

Abstract：In the Cyber-Physical system(CPS), equipment generates a large amount of communication information and equipment data during the operation of the system, which are at risk of being modified and stolen. Block chain technology with distributed storage as the core is conducive to ensuring the data security of CPS system, but the existing consensus algorithms have the problem that data recording rights are concentrated in a few nodes and are easy to be hijacked by malicious nodes. By introducing node validity, the dynamic change of node behavior can be reflected. The experimental results show that the consensus mechanism based on the validity of nodes makes the distribution of nodes with access to data records more decentralized, better reflects the de-centralization characteristics of the system, and improves the security of data.

Abstract：The traditional digital signal processor(DSP) control system has been widely applied in the control and protection of low voltage power electronic devices and achieved good control effect, but it can hardly achieved good control effect for the high voltage power electronic devices is bloated. So, based on DSP, field-programmable gate array(FPGA) and complex programmable logic device(CPLD), an universal control system was proposed for the high voltage power electronic devices. In the system, DSP is responsible for the system core algorithm and logic control, FPGA is responsible for the system voltage and current high-speed sampling, fast logic operation and multi frequency pulse width modulation(PWM) signal output, CPLD realizes the output PWM signal detection and fault locking function at a high speed, and separation of the main control board and other signal input and output boards expand controller applicability. Feasibility of the design of the control system has been proven by experiment.

Abstract：This paper proposed a technique of maximum efficiency and constant voltage output for wireless power transfer system. A semi-bridge rectifier is used to replace the common diode rectifier on the receiver side, and a Boost circuit is adopted to regulate the system output voltage. The duty cycles of the semi-bridge rectifier and the Boost circuit are online adjusted so that the wireless power transfer system can keep both maximum efficiency and constant voltage output when load varies and mutual inductance of coupled coils changes. Experimental results verified the feasibility of the proposed technique and the correctness of the theoretical analysis.

Abstract：Both forgetting to unplug mobile phone chargers after using and generating large surge current when the charger is plugged into power, will accelerate the aging of high-voltage electronic components. It is also prone to failure and fire. In addition, the charger consumes 100~800 mW power when it is in standby mode for a long time. Therefore, a fireproof and energy saving mobile phone charger is proposed. This kind of charger can reduce the standby power consumption to 13 mW, so it is unnecessary to pull out the charger to meet the habit of the users. It can absorb the surge current, meanwhile, it also has the functions of timing charging, overcurrent and short circuit protection, so as to prevent the occurrence of fire.

Abstract：By analyzing the shock and spike voltage on the rectifier diode during converter process of the phase-shift full bridge ZVS DC/DC converter which is configured as full wave rectifier in second side, this paper aims at explaining the reason of shock and spike voltage after that equivalent circuits and mathematical models are established. Then by analyzed the mathematical model, we thus conclude that a relationship is between parasitic parameters and resonance oscillations. And last RCD snubber circuit experiments are carried out in Saber and experiment prototype of 500 W which proves resonance oscillations can be curbed effectively.