Abstract:Radio frequency identification technology is one of the important technologies currently promoting the development of the Internet of Things. Because the wireless radio frequency signal is easy to be attacked during the transmission process, the RFID system needs to establish a complete and complete guarantee mechanism. Based on the security requirements and technical status of RFID systems, this paper proposes a lightweight RFID security authentication protocol. The protocol is based on LED cryptography and physical unclonable function, and uses the challenge-response signal pair of PUF for authentication. The LED algorithm encrypts and transmits the response signal of the PUF to ensure the security of the authentication information. After each authentication, the label information in the server will be updated. This paper uses the Verilog language to implement and simulate the circuit certification process, and conduct a comprehensive analysis of the circuit based on the standard cell library of the 40nm platform. Simulation and comprehensive results show that the lightweight RFID security authentication protocol can effectively resist common attacks, and the hardware overhead of tag storage and calculation is low, which is suitable for resource-constrained scenarios.

Abstract:Although the design of IPMC as the core of server management is becoming more and more mature, IPMC is still designed in the form of "integration" directly welded on the server motherboard, which has a high complexity in hardware update and maintenance. In this context, this paper puts forward the idea of "hardware plug-in", which makes IPMC into a pluggable board in the form of golden finger hardware interface. This board can not only solve the problems mentioned above, but also has Gigabit Ethernet switching function. The board hardware is mainly composed of microcontroller MSP432E401Y and L2 Ethernet switch chip KSZ9897R. In order to solve the hardware redundancy problem of Ethernet communication between boards, this design uses a transformerless Ethernet circuit interconnection. After the simulation and optimization of via parameters by HFSS software and the eye diagram and S-parameter test of high-speed signals between boards, the results show that the plug-in board hardware designed in this paper has a maximum voltage loss of 3.5dB when transmitting gigabit rate signals. The return loss is greater than 10dB, which fully meets the loss requirements of the IEEE802.3 standard for Ethernet signals.

Abstract:A MFA-SM control scheme based on RBFNN disturbance observer is designed to solve the problem that it is difficult to accurately control the output shaft speed of the valve-controlled electro-hydraulic rotation system during the drilling of surrounding rock due to the non-linear factors such as uncertain parameters, unknown load and external disturbance. First, an improved dynamic linearization method is used to linearize the electro-hydraulic system to an incremental model only related to the system I/O data. Unknown loads and external disturbances are combined into an unknown nonlinear time-varying term. Then, an RBFNN disturbance observer is designed to estimate the nonlinear term on-line in real time, and the time-varying pseudo-gradient parameters are estimated from the system I/O data. Finally, the corresponding controller design is given. The simulation results show that the MFA-SM controller can effectively compensate unknown load and external disturbance. Compared with other methods, this scheme shortens the system adjustment time by 10 to 15 seconds, reduces the maximum overshoot by about 7.4%, and the speed tracking error can converge to zero.

Abstract:Aiming at the problems of the insensitivity and inaccuracy of conventional controllers when the prpton exchange membrane fuel cell (PEMFC) power is high and the load current changes dynamically, a variable universe fuzzy PID control strategy was proposed in this paper. The proportion factor and scale factor in the fuzzy PID controller can be adjusted in real time by the factor to realize the contraction-expansion of the fuzzy universe, thereby improving the stability of the control. Build a 55KW fuel cell model and verify the feasibility of the model. Comparative experiments between the traditional control strategy and the variable universe fuzzy PID controll strategy. The research results indicate that the variable universe fuzzy PID controller is superior to the traditional controller in the overshoot, the steady-state error, the adjustment time, and the anti-interference performance.

Abstract:Flyback converter has two working states: continuous inductance current mode (CCM) and discontinuous mode (DCM). Usually, the critical state of CCM and DCM is selected as the working point to design the compensation loop of flyback switching power supply. Although the compromise scheme can take into account two working modes, it has the problems of small bandwidth and poor dynamic response, especially under light load. To solve this problem, a flyback switching power supply with switchable working mode is designed. The loop compensation network is designed for CCM and DCM modes respectively. The load condition is judged by the primary side current of the transformer, and then the compensation network is switched to realize the switching mode of the power supply. A prototype flyback switching power supply with output power of 120W was designed, and the simulation and performance test of the prototype were carried out. The test results show that the power supply can not only switch the working mode according to the load, but also has a good dynamic response under light load. The effectiveness of the scheme is verified, which is consistent with the expected design results.

Abstract:In order to further reduce the power fluctuation of grid-connected photovoltaic arrays, this paper proposes a power smoothing method that used in photovoltaic array connecting to the grid based on variable-boundary tracking differentiator. Besides simulated annealing particle swarm optimization is used to optimize the boundary parameters. Firstly, a generalized form of variable-boundary tracking differentiator is constructed according to variable-boundary theory. Then, for the boundary parameter optimization problem, this paper proposes a fitness function based on short-time-scale volatility for real-time optimization, which improves the smooth ability of the algorithm. Taking a photovoltaic array in East China as an example, the experiment proves that this strategy has a better smoothing effect than the traditional strategy. The maximum fluctuation and average fluctuation of power are reduced by 64% and 16% respectively. And the charge and discharge energy of energy storage is reduced by 5%.

Abstract:In modern warfare, the status of infrared-guided missiles is increasing, and the most countries in the world attach great importance to the development of infrared-guided missiles. The performance of the seeker gyro, as a key component of missile guidance control, has a decisive influence on the operational capability of missiles. In this paper, we analyze and study the suspension, spin up and speed stabilization control problems in the gyroscopic tracking system of a certain type of seeker, and propose a new control method to realize the highly reliable and fast spin up control of the seeker gyro and complete the test verification. The experimental results show that the fast spin up control design of this paper can ensure the high reliability and shorten the start-up time of the seeker at the same time, which can effectively improve the operational performance of the missile.

Abstract:With the rapid development of underground utility tunnel, the safety of pipelines is becoming more and more important to ensure the safety of electricity and gas transportation in pipelines. Due to the harsh environment of utility tunnel, manual inspection is inefficient and has great security risks.In order to realize the omnidirectional detection of underground utility tunnel without blind spots, an omnidirectional robot for pipeline detection based on mechanical arm is designed. The mobile robot, mechanical arm, detecting probe are combined into a whole. The robot has the functions of path independent planning, tracking, obstacle avoidance, etc., and detects pipeline leakage through magnetic flux leakage technology, and realizes detection in complex environment through mechanical arm.Experiments show that the integrated robot can carry out omnidirectional detection and has good accuracy in defect identification.

Abstract:Leak detection of pipes in underground pipe gallery is one of the important guarantees for the safe operation of pipe gallery. In order to realize the automatic operation of pipe gallery, and for the shortcomings of insufficient flexibility and poor information interaction of traditional detection methods, this paper designs a new method based on STM32F103. Rail-mounted pipeline leak detection robot. The robot is equipped with a MIC sound sensor as a leak detection device, and the collected signal uses a neural network to identify the audio features, and the monitoring host makes an alarm or other measures on the judgment result. After the test of the pipeline leakage detection system experimental platform, the robot system runs stably, and the detection rate of pipeline leakage reaches 93%, which meets the practical application requirements.

Abstract:Ranging Radio Frequency Identification indoor positioning algorithm has the problem of large errors in positioning. An indoor positioning algorithm based on Particle Swarm Optimization based on Fireworks was proposed. The algorithm is divided into two stages: ranging and positioning. In the ranging stage, the Phase Difference of Arrival is used to measure the distance and construct the objective function to be optimized. In the positioning stage, the particle swarm optimization algorithm is improved. In order to improve the problem that particle swarm optimization is easy to fall into local extreme value during the iterative process, the explosion, mutation and selection operations of Fireworks Algorithm are introduced, the selection rules are improved; The algorithm also improves the speed update formula of particle swarm algorithm according to the firework explosion operator and mutation operator. The experimental results show that the algorithm can effectively locate the target, and the average error of positioning is 0.2773m, and compared with the indoor positioning algorithm based on the standard particle swarm optimization algorithm, it has a performance improvement of 39.61%.

Abstract:In order to study the probability power flow calculation of multiple power plants with output correlation.A combined probabilistic power flow calculation method based on Box-Cox transform and nonparametric kernel density model is proposed for multiple photovoltaic power plants.Firstly, combining Gumbel Copula function and Gini coefficient, the density function is established to describe the correlation and intensity of photovoltaic output.Then, based on Box-Cox normal transformation method and improved LDU triangulation method, the PV output is normalized and independent respectively,and comprehensive modeling of PV output is carried out. Finally, the Cumulant Method combined with Cornish-Fisher series expansion is used to calculate the probability power flow, and the probability distribution of node voltage and branch power flow is obtained and the results are compared with traditional nonparametric kernel density estimation and Monte Carlo method.The simulation results of a practical example show that the proposed method has high accuracy and less time，it took just 3.17 seconds，it has certain practicability and solves the problem of low efficiency of the traditional Monte Carlo method.

Abstract:Load forecasting is crucial to the economic operation of the power grid. In order to improve the accuracy of short-term load forecasting and reduce the training time of the hybrid neural network, a short-term load forecasting method based on basic network with multilayer perceptron (MLP), simple recurrent units(SRU) and principal component analysis (PCA) is proposed. Firstly, the method considers various power load influencing factors to establish input feature sets of the load forecasting task, and uses PCA to transform and reduce the historical load and temperature characteristics which are the original inputs of the network. Then, the method uses new data information obtained after PCA as the inputs of the hybrid neural network model, and trains the network with Adam gradient descent algorithm. Finally, the outputs of the proposed model are load forecasting results. The results of the experiments show that the MAPE of the hybrid model including SRU on all test set samples is 2.126%, which is much lower than that of the single model with only basicnet and the hybrid model including DNN, and compared with the hybrid model including LSTM, the training time is reduced by 22.74%, and the application of PCA also accelerates the convergence of the model, which greatly reduces the number of training epochs.

Abstract:Achieving fast and accurate estimation of vehicle attitude is an important guarantee for successful mission execution. In order to improve the accuracy and speed of attitude estimation, an improved particle swarm algorithm is proposed to be applied in the spectral peak search, considering that the multiple signal classific-ation algorithm (MUSIC) algorithm is computationally intensive and slow in the spectral peak search. Firstly, the variation pattern between the attitude position of each electromagnetic vector sensor on the aircraft fuselage and the signal information transmitted from the ground base station is relied upon to form the steering vector required in the MUSIC algorithm, to establish the mathematical model expression of the electromagnetic wave signal of the signal receiving array composed of electromagnetic vector sensors, to find the covariance matrix, to decompose the eigenvalues of the matrix to obtain the noise subspace, and to construct the The attitude space spectral function is constructed to complete the establishment of the signal space spectrum and the search of the spectral peak, so as to obtain the unique spectral peak characterizing the attitude. Finally, the simulation shows that the improved particle swarm algorithm can effectively improve the estimation accuracy and search speed of the flight attitude.

Abstract:In conventional DOA methods, the subspace fitting algorithms have been widely used for its high resolution and high estimation accuracy. However, when subspace algorithm works, it needs to predict the exact number of sources, otherwise the performance of the algorithm will be seriously affected. In the view of the problem of high precision direction-finding under conditions where the number of signals is unknown, a DOA algorithm based on spatial secondary projection is proposed. Firstly, this algorithm constructs the primary-projection spatial spectrum by the relationship between vector and its spatial projection. Secondly, it takes a secondary-projection to removes the possible pseudo-peaks by taking the advantage of the uniqueness of the vector representation by complete spatial vector basis. The computer simulation results show that the spatial secondary projection algorithm can realize the accurate measurement of the coherent signals’ DOA without predicting the number of signal sources, and this algorithm has high spatial resolution and algorithm stability under the conditions of small aperture and limited number of snapshots.

Abstract:Aiming at the shortcomings of the existing anti-pinhole camera with single function, low recognition accuracy, and low real-time performance, an experimental system of anti- pinhole camera device was designed. The system uses a four-quadrant infrared detector to expand the detection area. Based on the autonomous division of the area, the detection results are transmitted to the supporting APP through STM32 microcontroller, which is convenient for users to view the results in real time. Meanwhile, the WiFi module of ESP8266 is used to attack the network connected to the camera to block data transmission and ensure privacy security of users. Through this experiment, students have a deep understanding of detection technology and other related theories, and are familiar with the development process using single chip microcomputer, so as to improve their learning and practical ability of solving practical problems.

Abstract:The illegal action of exceeding the powdered food additives standard seriously threatens the health of people. In the study, a method for qualitative and quantitative prediction of azoformamide doping in flour through Raman hyperspectral images was developed by using a self-built Raman hyperspectral detection system. In this method, Raman hyperspectral images near 785 nm of azoformamide in samples were obtained through laser line light source. By preprocessing, selecting the region of interest, data dimensionality reduction and setting an appropriate threshold, the effective distinction between flour and azoformamide signals is realized. The method of image analysis was used to detect the azoformamide doped in samples with gradient concentration. Then the related quantitative analysis model was established. Finally the reliability of the quantitative analysis model is verified by the prediction sets and the correlation coefficient is more than 0.988. This study provides a new method for the detection of powdered food by Raman hyperspectral technology.

Abstract:Thevenin second-order equivalent circuit model is used to describe the state space of the power battery with high accuracy and moderate identification difficulty. The DEKF+GA algorithm is proposed to identify the parameters of the equivalent circuit model of Li-ion battery with noise interference, and the GA algorithm is added to search for the optimal solution at the initial value of the DEKF solution to improve the accuracy and robustness of the DEKF algorithm. Experiments show that applying the DEKF+GA algorithm reduces the URMSE and UMAX by 29mV and 27.73mV respectively on average compared to the DEKF algorithm.

Abstract:Deep learning algorithms are widely used in the field of network traffic classification and have achieved good results. However, the emergence of adversarial attacks has brought a serious threat to its security, and the accuracy of the current mainstream classification algorithms based on convolutional neural network models has been seriously reduced. In response to this, this paper proposes an encrypted traffic classification method that resists gray-scale adversarial attacks in traffic classification. The proposed method constructs a topology graph by extracting traffic interaction information such as packet load length, sending order, direction, and cluster, and transforms the encrypted traffic classification problem into a graph classification problem. Then, this paper uses the classification method based on graph convolutional neural network to learn and classify features. The graph convolutional neural network model can automatically extract features from the input topology and map features to different representations in the embedding space to distinguish different graph structures. The experimental results show that the proposed method can not only avoid adversarial attacks, but also improve the classification performance on public datasets by more than 5% compared with the existing typical methods.

Abstract:Aiming at the problem of low defect recognition rate in radiographic images, background estimation and differential operations are used to enhance defects and suppress complex background and noise. The method first used the mask image obtained by Otsu segmentation to extract the weld area; Secondly, the background estimation of the weld area was performed by the improved median filter, and the difference image containing the defect was obtained by inverting the background difference; Then, according to the difference of the gradient direction at the edge of the defect and the false detection, the multi-directional and multi-level gradient was used to effectively solve the background residual problem; Finally, t the differential image containing the defect was binarized by adaptive threshold segmentation. After experimental simulation, this method has a high defect recognition rate, with a recall rate and an accuracy rate of 91.90% and 90.95%, respectively, which has good application value in practice.

Abstract:Aiming at the problem that YOLOv4 has a huge backbone network and a large number of parameters, it cannot meet real-time requirements when applied to insulator defect detection. A lightweight YOLOv4 detection model is proposed. First, GhostNet with ECA integrated components is introduced as the feature extraction network, which greatly reduces the model parameters and speeds up the model inference while ensuring the feature extraction capability. Secondly, the K-means++ clustering algorithm is used to determine the initial anchor frame size to adapt to the size of the insulator defect and improve the accuracy of defect location. Finally, on the basis of the cross-entropy loss function, the Quality Focal Loss is introduced to improve the loss function to further improve the model detection performance. Experimental results show that compared with the original YOLOv4, the improved lightweight YOLOv4 has a reduced model size of 62.47%, Frames Per Second increased by 68.83%, and the accuracy of insulator defect detection has increased by 1.07%, significantly improving the detection speed. At the same time, the detection accuracy of the algorithm is guaranteed, and it performs outstandingly in small targets and complex backgrounds.

Abstract:Aiming at the problems of small datasets, unbalanced categories and poor diagnostic results in deep learning diagnosis of diabetic retinopathy (DR), this paper proposes a DR grading model based on parallel images and Swin Transformer. First, build a parallel image generation model based on StyleGAN2-ada to solve the problem of too few training images and class imbalance. After FID, KID and visual evaluation, the constructed parallel images meet the requirements of subsequent work. Then, a DR diagnosis model is constructed based on the attention and window shifting mechanism to improve the diagnosis effect. Finally, a diagnostic model is trained using the parallel images. After verification, the accuracy of the diagnostic model proposed in this paper is 93.5%, the highest specificity is 99%, and the highest F1-score is 0.96. Compared with the original images, the accuracy of the model is improved by 20% and the accuracy is improved by up to 70% after training the model with parallel images. Compared with the other three deep learning models, all the indicators of the method proposed in this paper are optimal. The above results show that the model constructed in this paper can achieve better diagnostic results under a small sample data set.

Abstract:The key of Coriolis mass flowmeter signal processing depends on the accurate estimation of frequency and phase difference. Frequency estimation methods exists the problem of low tracking accuracy for a long time, and phase difference estimation methods exists the problems of insufficient accuracy and poor real-time performance. Firstly, by introducing negative feedback control, the problem of long-time continuous tracking for adaptive notch filter can be effectively solved and the accuracy of frequency estimation can be improved. Then, the frequency estimation results are used to process the whole period data of the enhanced signal filtered by the adaptive notch filter. Then, Hilbert transform is performed on the signal after the period data processing. Finally, the signals before and after Hilbert transform are correlated, and the phase difference can be obtained by sinusoidal formula, so as to obtain the mass flow. The simulation results show that the proposed method has high frequency and phase difference estimation accuracy, and can be used for real-time signal processing for Coriolis mass flowmeter.

Abstract:Time delay estimation is often used in wireless positioning and ranging. In multipath environment, there will be problems of positioning accuracy degradation and time delay estimation distortion. To solve this problem, a time delay estimation algorithm based on orthogonal frequency division multiplexing (OFDM) frequency domain subspace smoothing on multiple signal classification (MUSIC) is proposed. First, the signal source is modulated by OFDM and the data stream is formed by using subcarriers. Then, the covariance matrix of data stream is bidirectional smoothed in frequency domain to make the best use of the data information of signal subspace. In the end, the signal or noise subspace orthogonality of MUSIC algorithm is used to detect the spectrum peaks one by one, and the pseudo spectrum is normalized to obtain more accurate time delay information. Computer simulation shows that the proposed algorithm has higher spectral peak and narrower sidelobe than the eigen space MUSIC algorithm, and has no error estimation interference. When the ranging signal interval is close, it can effectively solve the problem of estimation distortion. the minimum time delay interval can be close to 6ns, and the resolution is strong. At the estimation performance level, the estimation accuracy can be close to 3ns under the complex condition of signal-to-noise ratio of -15dB, which verifies the effectiveness and superiority of the improved algorithm.

Abstract:In order to meet the requirement of reliable multi-channel high-speed data transmission in space telemetry system, a four-channel data transmission design based on FPGA controller and Serial RapidIO(SRIO) protocol is proposed. Xilinx A7 series FPGA is used in the design, and four SRIO IP cores are used to design the internal logic and realize the high-speed data transmission of four-way SRIO. Use its internal integrated Gigabit transceiver (GTP) to meet the SRIO transport protocol physical layer requirements. The hardware circuit uses four high-speed receiving and luminous modules to complete the photoelectric conversion; A high quality clock chip is used to generate 125MHz differential clock signal as the reference clock of the SRIO IP core. The data transmission rate of four channels can reach 440MB/s without frame loss and error. The design has been successfully applied to a ground test platform project of telemetry system, which can realize stable transmission of four channels high-speed data.

Abstract:In order to improve the stability and veracity of offshore wind turbine installation, it is necessary to perform the heave compensation of the floating crane. But the ship heave motion measured by the inertial system has the characteristics of the random drift and the advanced phase, which seriously affects the real time ability and accuracy of the heave compensation system. So this paper proposes a ship heave motion measurement method based on UKF multi-step observer and BP residual correction. Based on the Stewart wave motion experiment platform, the ship heave motion is simulated, and the inertial measurement system is used to collect heave displacement and acceleration information, then the state space model of the motion is established. Thus, a UKF observer is established by the state space model, and then the state transition matrix is used to perform multi-step observation according to the advanced phase of the heave motion of dynamic detection, to eliminate the random drift and the advanced phase of heave motion. And then, UKF multi-step observer and observation residual are applied to train the BP residual prediction model, and then the prediction residual is used to correct the observation of the multi-step observer online to improve measurement accuracy. The experiments based on Stewart platform show that the proposed method solves the problems of the random drift and the advanced phase of ship heave motion, and the measurement accuracy can be improved to 90%.

Abstract:Aiming at the problems that the fault definition standard of analog circuit board chip is not clear and it is difficult to realize fast and accurate classification, this paper proposes a fault diagnosis model based on binary convolution logistic atom search algorithm to optimize BP neural network. Firstly, the temperature of circuit board chip in different states is collected and feature extracted, and the features are fused by Euclidean distance to establish a fault feature model containing chip fault definition criteria. Then, the binary convolution logistic map is used to initialize the population size and location of the atomic search algorithm to improve the convergence speed and accuracy. Then, the optimization process of BP neural network is optimized by BCL-ASA to obtain the optimal weight and threshold. Finally, input the chip fault characteristic model into the BCL-ASA-BP neural network for training and testing to complete the circuit board chip fault diagnosis. The experiment uses the power supply circuit board for reliability analysis, and the results show that the accuracy of BCL-ASA-BP's comprehensive diagnosis of chip faults reaches 98.35%, which is 13.9% higher than the traditional BP algorithm.

Abstract:The thermal aging effect of glass fiber reinforced composites (GFRP) under the influence of different time needs to be studied urgently. In this paper, the complementary nondestructive testing technology which combines Acoustic Emission(AE) and Digital Image Correlation(DIC) technologies was used to monitor the damage process of three-point bending specimens after thermal aging at 0, 4, 8 and 16 days respectively. Finally, the effect of prefabrication defects on the fiber-matrix interface properties of composites after thermal aging was analyzed. K-means cluster analysis was performed on the AE signals and global strain field measurements were performed by DIC to characterize the effects of prefabrication defects and thermal aging time on the mechanical behavior and deformation damage process of GFRP. Acoustic emission signals can classify various types of damage, and through the analysis of debonding signals, it was found that when crosslinking inside the laminate, the debonding signal was reduced and distributed after the failure load; as the thermal aging progresses to the stage of moisture volatilization, the debonding signal increased and was evenly distributed in the pre-fracture and post-fracture stages. For prefabricated fiber fractured laminates, the crosslinking phenomenon will reduce the number and amplitude of fiber fracture signals during laminate loading, and the maximum strain value will be reduced, which will effectively improve the carrying capacity of the laminate. The combination of acoustic emission clustering analysis and digital image correlation technology can better describe the damage evolution of fiber-reinforced composites under different thermal aging conditions, which is helpful to further reveal the influence of defects and thermal aging time on internal damage mechanism.