Table of Content

15 July 2022 Volume 35 Issue 7

  

A Fast Analysis Method of Pathological Data Based on Wave Vector Grading Technology

KONG Fanshu,QI Jinpeng,GONG Hanxin,ZHU Junjun,CAO Yitong

Electronic Science and Technology. 2022, 35(7):  1-6.  doi:10.16180/j.cnki.issn1007-7820.2022.07.001

Abstract ( 132 )   HTML ( 0 )   PDF (1955KB) ( 19 )  

Figures and Tables | References | Related Articles | Metrics

In large-scale time series data analysis, traditional mathematical statistics and analysis techniques have problems such as time-consuming, low accuracy, and weak anti-interference ability. In view of the deficiency, based on the wave vector grading technology, this study presents a fast analysis method for the time series data of lesions. Based on the TSTKS mutation point detection algorithm and sliding window theory, this method uses multi-threshold segmentation technology to realize the multi-level classification strategy of fluctuation vectors, and then realizes the state analysis and rapid diagnosis of large-scale lesion time series data. The result of simulation experiments and brain epilepsy lesion signal analysis show that the proposed method has the advantages of faster speed and higher efficiency, and can provide a new method for the rapid analysis and research of large-scale time series data.

Partial Discharge Pattern Recognition of Cable Based on CNN-DCGAN under Small Data

SUN Kang,XUAN Xuyang,LIU Penghui,ZHAO Laijun,LONG Jie

Electronic Science and Technology. 2022, 35(7):  7-13.  doi:10.16180/j.cnki.issn1007-7820.2022.07.002

Abstract ( 260 )   HTML ( 1 )   PDF (2017KB) ( 43 )  

Figures and Tables | References | Related Articles | Metrics

In the process of cable partial discharge pattern recognition, the traditional manual feature extraction relies on the knowledge and experience of specific fields, and the workload of feature selection and optimization is heavy. In view of this problem and to avoid the overfitting problem of the classifier under the unbalanced small sample data of the model, this study presents a partial discharge pattern recognition method based on CNN-DCGAN in the case of small samples. Partial discharge time domain signals are transformed into two-dimensional image information by sliding time window. The DCGANs are constructed, and the data enhancement is carried out on the basis of the original data set. The original data and the enhanced data are taken as the system input. CNN is constructed, and its nonlinear encoder is used to automatically extract partial discharge features, and the feature classification model is trained by Softmax layer. Experimental results show that compared with artificial features, the recognition accuracy of CNN classifier based on automatic feature extraction is improved by 4.18%. Compared with the original data set, the system recognition accuracy based on the sample enhanced data set is improved by 3.175%.

Minimum Cost of Heterogeneous Directional Sensor Networks for Target Coverage

LIU Yongpan,WANG Ran

Electronic Science and Technology. 2022, 35(7):  14-21.  doi:10.16180/j.cnki.issn1007-7820.2022.07.003

Abstract ( 133 )   HTML ( 0 )   PDF (1448KB) ( 16 )  

Figures and Tables | References | Related Articles | Metrics

Deployment cost is an important indicator for evaluating the energy harvesting directional sensor network. Considering the network connectivity and the heterogeneity of sensors, the study proposes a heuristic two-stage selection algorithm (HTS algorithm) for the minimum cost target coverage problem of the energy-harvesting heterogeneous directional sensor network. The time complexity of selecting points is reduced by constructing a set of efficient candidate points, and the utility function is used to evaluate each deployment method. After sensing nodes are deployed in the best deployment method, eliminating redundant nodes and shrinking the sensing scope can further save costs. The simulation results show that the HTS algorithm can reduce the deployment cost of the sensor network by 10% to 18% when compared with the current directional target coverage algorithm.

A Survey of Intelligent Transportation Path Planning Algorithms

LU Dongxiang

Electronic Science and Technology. 2022, 35(7):  22-27.  doi:10.16180/j.cnki.issn1007-7820.2022.07.004

Abstract ( 1596 )   HTML ( 84 )   PDF (640KB) ( 286 )  

References | Related Articles | Metrics

In order to solve problems such as traffic congestion, traffic safety and environmental pollution of urban roads, a variety of technologies including path planning algorithm and vehicle self-organizing network have gradually become important research directions in the field of intelligent transportation and attracted a lot of attention. In recent years, the application and popularization of intelligent transportation provide rich and diverse research topics for the research of path planning algorithm, and the research of path planning algorithm also effectively promotes the development of cutting-edge technology of intelligent transportation. By reviewing the research of multiple path planning algorithms, this study summarizes the existing problems in the research of path planning algorithms. On this basis, this study deeply studies and analyzes the advantages and disadvantages of static and dynamic path planning algorithms, and puts forward the future research trend and methods of path planning algorithms in intelligent transportation.

Quality Inspection Algorithm of Chemical Packaging Bag Coding Based on Tesseract_OCR

ZHANG Maolin,YE Qingzhou,PAN Xin,LU Hua

Electronic Science and Technology. 2022, 35(7):  27-31.  doi:10.16180/j.cnki.issn1007-7820.2022.07.005

Abstract ( 270 )   HTML ( 4 )   PDF (965KB) ( 26 )  

Figures and Tables | References | Related Articles | Metrics

For the problems of low efficiency and high leakage rate in manual quality inspection of information printing on chemical packaging bags, a machine vision-based quality inspection method for printing codes on chemical packaging bags is designed in this study. Mean filtering and Gaussian bilateral filtering algorithms are used to pre-process the captured image, and then the character area is localized through a variable threshold algorithm based on local statistics. To solve the problem that the distance between the dots of the printout characters may be larger than the gap size between the characters, which leads to the formation of a connected domain with multiple characters sticking together after the binary image closure operation, the study proposes a dynamic character segmentation algorithm to improve the connected domain. The segmented character images are trained and recognized by Tesseract_OCR engine for classification. The experimental results show that the algorithm has the accuracy rate of 95.62% for coding quality detection, which can meet the requirements of chemical packaging bag coding quality inspection.

Crop Height Measurement Based on Ruler Image Recognition

SUN Xiang,PEI Xiaofang,ZHOU Wang,ZHU Ping

Electronic Science and Technology. 2022, 35(7):  32-39.  doi:10.16180/j.cnki.issn1007-7820.2022.07.006

Abstract ( 597 )   HTML ( 6 )   PDF (1393KB) ( 67 )  

Figures and Tables | References | Related Articles | Metrics

Crop height measurement is an important part of automatic crop observation, which can directly reflect the growth of crops. To solve the problem that the cost of artificial measurement of crop height is higher and the subjective influence of individual is greater, this study presents a method of measuring crop height by image processing. The ruler is placed behind the main stalk of the crop to take the photo of the ruler. The obtained images were converted to HSV color space on MATLAB to divide and filter the color blocks of the ruler. The pixel height of the color block is calculated by the connected domain analysis method. The plant height is calculated by the ratio relation between the ruler pixel height and the actual height. The shrubbery is taken as the experimental object to take photos and make measurement. Comparing the measured plant height data with manual measurement data, the results show that the measurement error is less than 0.0173 m, which meets the standard of automatic observation of plant height.

Improved YOLOv3 Model Based on New Feature Extraction and Fusion Module

ZHAO Xuan,ZHOU Fan,YU Hancheng

Electronic Science and Technology. 2022, 35(7):  40-45.  doi:10.16180/j.cnki.issn1007-7820.2022.07.007

Abstract ( 302 )   HTML ( 4 )   PDF (1663KB) ( 41 )  

Figures and Tables | References | Related Articles | Metrics

There is a certain optimization space for the feature extraction branch and multi-scale detection branch of YOLOv3 model. To solve this problem, this study proposes two structural improvement methods to improve the detection accuracy of the model on the target detection data set. For the three scales (13×13, 26×26, 52×52) of the YOLOv3 model, a priori anchor frames of different lengths and widths are used, and the label frames of the three scales are the same, and the feature fusion method between the design scales is used to improve the accuracy of the model. In view of the problem of convolutional layer spatial view sharing, the original convolutional layer can be replaced with deformable convolution to improve the accuracy of the model. The test result on the industrial tool library proves that the accuracy of the test set of the improved model is increased by 3.6 MAP when compared with the original YOLOv3.

Drilling Tool Acceleration Denoising Based on GRNN Network Adaptive Filtering

TONG Xiaosen,YANG Jinxian

Electronic Science and Technology. 2022, 35(7):  46-51.  doi:10.16180/j.cnki.issn1007-7820.2022.07.008

Abstract ( 146 )   HTML ( 1 )   PDF (1849KB) ( 17 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem of serious distortion of gravitational acceleration information caused by bottom-hole vibration and noise during measurement while drilling, a method that combines generalized regression neural network and improved recursive least squares is proposed to remove noise. Using the principle of mutual cancellation of adaptive noise, the accelerometer measurement signal and drilling tool vibration signal are used as the main noise model, and the drilling tool vibration signal processed by the generalized regression neural network is used as the auxiliary noise model. Recursive least squares filter processing is used to improve the accuracy of drilling tool accelerometer measurement and calculate real-time drilling attitude information. The simulation results show that the proposed algorithm can effectively remove the vibration and noise of the drilling tool accelerometer and improve the attitude measurement accuracy. After denoising, the calculated deviation angle error is within 1.45°, and the tool face angle error is within 1.65°.

Improved Remote Incremental Update Scheme Based on BSDiff

CHEN Dirong,BAO Xiaoan,DU Peng,HU Yifei,SU Hongbin

Electronic Science and Technology. 2022, 35(7):  52-57.  doi:10.16180/j.cnki.issn1007-7820.2022.07.009

Abstract ( 407 )   HTML ( 7 )   PDF (955KB) ( 41 )  

Figures and Tables | References | Related Articles | Metrics

The traditional incremental update solution has weak cross-version update performance, and multiple incremental update packages need to be generated and issued, resulting in low terminal device update efficiency. In view of this problem, an improved remote incremental update scheme based on BSDiff is proposed in the present study. The scheme optimizes the firmware management method of the update server, and uses the BSDiff algorithm to instantly generate a unique incremental update package to reduce the amount of update data that needs to be transmitted. The BSDiff algorithm is improved to speed up the generation of incremental update packages. The simulation experiments show that the improved incremental update scheme has the same compression performance as the traditional incremental update scheme. It can reduce the time by 31.19% on average when generating incremental update packages and reduce the data transfer amount by 2.07% on average when updating across versions.

Improved Montgomery Modular Multiplication Algorithm and FPGA Implementation

CHENG Biqian,LIU Guangzhu,XIAO Hao

Electronic Science and Technology. 2022, 35(7):  58-63.  doi:10.16180/j.cnki.issn1007-7820.2022.07.010

Abstract ( 580 )   HTML ( 18 )   PDF (1639KB) ( 183 )  

Figures and Tables | References | Related Articles | Metrics

In order to ensure the online information security of users, the public key cryptosystem is used to encrypt the data information. As the core operation of public key cryptosystem, the computation efficiency of large integer modular multiplication is very important to the performance of public key cryptosystem. In this study, a polynomial expanded cross Montgomery modular multiplication algorithm is proposed, which is based on the classical Montgomery modular multiplication algorithm. By decomposing large bit-width logic operations, and performing modular multiplication and modular reduction operations with polynomial expansion, this algorithm can effectively improve the calculation efficiency of large integer modular multiplication operations, and reduce the resource consumption of hardware implementation. FPGA experiment verification shows that compared with other methods, the proposed algorithm can reduce AT1 and AT2 (the products of area and time) by 96.5% and 69% respectively, indicating that the proposed method achieves the balance between computing time and hardware overhead, has high flexibility and universality, and is suitable for cost-sensitive applications with a large number of encryption requirements.

Fault Voting Method Based on Transient High Frequency Components

WU Jun,XIE Yifei,ZHANG Changsen

Electronic Science and Technology. 2022, 35(7):  64-70.  doi:10.16180/j.cnki.issn1007-7820.2022.07.011

Abstract ( 130 )   HTML ( 1 )   PDF (867KB) ( 18 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems of extraction and processing of transient line selection characteristics of single-phase grounding faults in existing low-current systems, as well as line selection accuracy, this study analyzes the transient characteristics of the faults and finds that the high-frequency components of the transient zero-sequence currents of various lines have obvious differences. Accordingly, a voting line selection method based on waveform similarity of high-frequency components of transient zero sequence current is proposed. Combined with VMD algorithm and Bhattacharyya distance algorithm, the zero sequence current of each line is decomposed to obtain the transient high frequency components, and then the Bhattacharyya coefficients of the transient high frequency components among the lines are obtained. The fault line is voted by the method of preliminary voting and k-value test on the coefficient matrix. The simulation results under different fault conditions show that the method is simple and reliable, with high accuracy and credibility of line selection, wide applicability and certain anti-interference.

Adaptive Finite-Time Estimation and Simulation of Friction Torque and Friction Model Parameters of Robot Systems

FANG Yueming,WANG Xian

Electronic Science and Technology. 2022, 35(7):  71-78.  doi:10.16180/j.cnki.issn1007-7820.2022.07.012

Abstract ( 194 )   HTML ( 3 )   PDF (1559KB) ( 31 )  

Figures and Tables | References | Related Articles | Metrics

The nonlinear friction torque in the robot system affects the performance of the control system. To address this problem, an adaptive parameter estimation algorithm is proposed to estimate the key parameters of a continuous friction model and realize the friction torque modeling. In order to avoid the use of acceleration signals, the robot system model is reconstructed and an unknown system dynamic estimator is designed to realize the overall estimation of the friction torque. The estimator has a simple structure, only needs to adjust one parameter to realize the estimation of the friction torque, and the estimated convergence speed is fast, which is beneficial to the realization in the actual system. On the basis of obtaining friction torque estimation, a finite-time adaptive parameter estimation algorithm based on parameter estimation error information is constructed according to the Lyapunov method. Under the premise of ensuring the convergence and robustness of the algorithm, the key parameters in the continuous friction model can be identified accurately in finite time.

Comparison of Coordinated Control between Front and Rear Stages of Isolated AC-DC Solid-State Transformers

ZHENG Zheng,JIANG Pengfei,ZHANG Guopeng,LI Zihan

Electronic Science and Technology. 2022, 35(7):  79-86.  doi:10.16180/j.cnki.issn1007-7820.2022.07.013

Abstract ( 155 )   HTML ( 1 )   PDF (2014KB) ( 22 )  

Figures and Tables | References | Related Articles | Metrics

Isolated AC-DC solid-state transformers widely use the front and rear multi-stage cascade structure of the bidirectional converter. Due to the difference in the control bandwidth of the front and rear stages, the transmission power of the two stages is different. When the power suddenly changes, the difference will seriously threaten the operational safety of this type of SST. In view of this problem, this study analyzes the mechanism of the power difference between the stages, summarizes various control strategies to reduce the power fluctuations between the stages, and derives the mathematical model of the load disturbance on the voltage between the two stages under different strategies. Using MATLAB, the closed-loop Bode plots of various typical control strategies and the corresponding step responses in the time domain are obtained, and the advantages and disadvantages of various strategies are compared horizontally. Finally, simulation and experiment results verify the correctness and validity of the analysis and comparison results.