Table of Content

15 January 2021 Volume 34 Issue 1

  

A Broadband High Directional Microstrip Directional Coupler

CHEN Xiaoli,XUE Huanjie,GUAN Boran

Electronic Science and Technology. 2021, 34(1):  1-4.  doi:10.16180/j.cnki.issn1007-7820.2021.01.001

Abstract ( 1006 )   HTML ( 52 )   PDF (911KB) ( 203 )  

Figures and Tables | References | Related Articles | Metrics

In this paper, a miniaturized directional multisegment microstrip directional coupler with high directivity is designed. By segmenting the coupled microstrip lines, a multisegment coupler composed of a quarter wavelength segment and two 1/8 wavelength segments is obtained. In addition, the difference of odd-even mode speed in the microstrip coupling line of non-uniform medium always leads to poor directivity of the microstrip directional coupler. Aiming at this problem, a quarter arc is adopted at the corner of the feed line of the coupler to improve directivity. A 10 GHz coupler is designed and implemented through optimized simulation. The coupling degree is 25 dB, and the directivity from 6 to 14 GHz is better than 30 dB. Compared with the traditional parallel coupled line coupler, the proposed coupler has smaller size has better directivity, and the measured and simulated results are basically consistent.

Pedestrian Detection Based on Improved YOLOv3 Algorithm

YE Fei, LIU Zilong

Electronic Science and Technology. 2021, 34(1):  5-9.  doi:10.16180/j.cnki.issn1007-7820.2021.01.002

Abstract ( 452 )   HTML ( 28 )   PDF (1519KB) ( 72 )  

Figures and Tables | References | Related Articles | Metrics

The YOLOv3 algorithm uses Darknet53 as the backbone in the target detection (pedestrian detection) of a single object, and the network appears redundant, which results in too many parameters and slow detection speed. Additionally, the traditional bounding box loss function makes the detection and positioning inaccurate. To solve these problems, the improved YOLOv3 backbone network is proposed in the current study. A new multi-scale fusion network based on Darknet19 is constructed to accelerate the training speed and detection speed, and a generalized intersection over union loss function is introduced to improve the detection accuracy. The experimental results show that the proposed algorithm improves the accuracy of the original algorithm by 5% in the pedestrian detection dataset such as the INRIA pedestrian dataset. Compared with Faster R-CNN , the detection speed of a single image reaches 0.015 s per image under the condition of good accuracy.

A Design and Optimization Analysis of the Centered Positioning Mechanism

SAN Hongjun, LI Pengyu, CHEN Jiupeng, HU Qiongqiong, LI Qi, ZHANG Kaixiang, WANG Chen

Electronic Science and Technology. 2021, 34(1):  10-16.  doi:10.16180/j.cnki.issn1007-7820.2021.01.003

Abstract ( 298 )   HTML ( 4 )   PDF (3019KB) ( 39 )  

Figures and Tables | References | Related Articles | Metrics

Display positioning fixture is the key equipment on the display assembly line, which plays an important role in realizing the automatic assembly of display. In this paper, a pair of centering positioning mechanisms for the crank slider is designed for the display positioning fixture. In order to improve the positioning accuracy and versatility of the fixture, the crank slider mechanism is optimized by the penalty function method, MATLAB is used to simulate it, and the key components are optimized by finite element method. The results show that the slider has a certain stroke in the process of clamping and positioning the components, and is affected by the length of the crank, the length of the connecting rod and the minimum transmission angle. When the length of the crank and the length of the connecting rod are 300 mm and 500 mm, respectively, the general performance of the mechanism is the best. Finally, the key components are optimized and analyzed by the finite element method, and the optimal results are obtained, so that the positioning accuracy of the centering positioning mechanism reaches 0.38 mm.

Study on the Identification of Various Growth and Monitoring of Pest and Disease of Rhododendron

PEI Xiaofang, HU Min

Electronic Science and Technology. 2021, 34(1):  17-22.  doi:10.16180/j.cnki.issn1007-7820.2021.01.004

Abstract ( 256 )   HTML ( 5 )   PDF (1818KB) ( 25 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem that the traditional BoF algorithm lacks spatial information, an improved BoF algorithm is proposed and applied to the identification of various growth stages and the monitoring of pest and diseases of Rhododendron in this paper. The ordered spatial information is integrated into the LAB-based color features by this algorithm to form a new spatial color aggregation feature to replace the traditional color histogram, which effectively solved the problem of the small scale of color feature change. Besides, SURF features are extracted to replace the original SIFT features. The image classification is realized by a multi-class features, and the leaf features are further analyzed to quickly identify the growth period and disease of Rhododendron plants. Simulation results show that the recognition rate of the improved BoF model of the LAB-based color aggregation vector is 90.6%. Compared with the image classification method of the traditional histogram, the image retrieval speed is increased by 3 times, making it easier to implement the feature fusion.

Saliency Detection by Progressive Structural Receptive Field and Global Attention

DONG Bo,ZHOU Yan,WANG Yongxiong

Electronic Science and Technology. 2021, 34(1):  23-30.  doi:10.16180/j.cnki.issn1007-7820.2021.01.005

Abstract ( 210 )   HTML ( 6 )   PDF (2136KB) ( 41 )  

Figures and Tables | References | Related Articles | Metrics

In view of the current deficiencies that the previous saliency detection algorithms are difficult to segment the complete salient region and sharp edge details in complex scenes, a novel feature fusion of saliency detection model is proposed in this paper. The proposed algorithm utilizes full convolution neural network to obtain the initial features of multi-level roughness and combines the feature pyramid structure to analyze its depth. In order to realize the gradual fusion and transmission of features, the progressive structural receptive field module is designed to transform features to different scales of space for optimization. The global attention mechanism is used to eliminate the background noise and establish the long-distance dependence between the saliency pixels, so as to improve the effectiveness of the saliency region, highlight the saliency region, and then obtains the saliency map by learning and fusing the hierarchical features. The comprehensive experiment show that the F-measure index is far beyond the other seven mainstream methods when the absolute error is reduced. The proposed saliency model combines the advantages of full convolution neural network and feature pyramid structure, and combines the gradual structure receptive field and global attention mechanism designed in this study to make the saliency map closer to the truth map.

Simulation Study of a Low Power 4H-SiC IGBT

SU Fangwen,MAO Hongkai,SUI Jinchi,LIN Mao,ZHANG Fei

Electronic Science and Technology. 2021, 34(1):  31-36.  doi:10.16180/j.cnki.issn1007-7820.2021.01.006

Abstract ( 454 )   HTML ( 14 )   PDF (975KB) ( 85 )  

Figures and Tables | References | Related Articles | Metrics

In this paper, a new trench gate 4H-SiC IGBT structure is designed and simulated by Silvaco TCAD for the problem that the traditional trench gate 4H-SiC IGBT has long turn-off time and high turn-off energy loss. By improving the structure of the conventional trench gate 4H-SiC IGBT, two sets of high doping P and N layers are introduced in the N ⁺ buffer layer to increase the donor concentration of the N ⁺ buffer layer, which compromises the forward voltage drop of the device and turn off the energy loss. The PN junction in the reverse bias state in the N ⁺ buffer layer optimizes the electric field distribution in the N ^-drift region when the device is in the turn-off process, which accelerates the electron extraction in the N ^-drift region and shortenes the turn-off time of the device. This ultimately reduces the turn-off energy loss of the device and increases the breakdown voltage of the device. Silvaco TCAD simulation results show that the new trench gate 4H-SiC IGBT has a breakdown voltage of 16 kV. Compared with the conventional structure with a breakdown voltage of 15 kV, the turn-off energy loss is as low as 4.63 mJ, which is 40.41% lower than the conventional structure.

Experimental Study on Milling Pure Copper with 50 μm D-Shaped Cemented Carbide Micro-Milling Cutter

YAN Guanghe, JIANG Chen, ZHANG Yongbin, HONG Xiaolan, QIAN Dabing

Electronic Science and Technology. 2021, 34(1):  36-42.  doi:10.16180/j.cnki.issn1007-7820.2021.01.007

Abstract ( 369 )   HTML ( 3 )   PDF (2597KB) ( 25 )  

Figures and Tables | References | Related Articles | Metrics

According to the requirement of miniaturization and high precision of folded waveguide slow wave structure, D-Shaped micro-milling cutter with diameter of 50 μm is fabricated by WEDG with the self-developed μEM-200CDS2 combined high precision machine tool. The surface quality of the microgroove is studied by the process test on the pure copper workpiece. White-light interferometer, ultra-depth-of-field microscope and scanning electron microscope are used to observe the variation of surface morphology and roughness of micro-grooves with milling distance, and the surface quality and burr formation during micro-milling are analyzed. These results show that the revolving track of the groove bottom tool is firstly concentrated in the cutting direction of the tool, and with the tool wear, it turnes to the feed direction of the tool. The burrs on the milling side of the micro groove are more than those on the reverse side, and increases with the increase of milling distance. Additionally, the roughness of the groove bottom increases with the linear growth trend.

Particle Swarm Optimization BP-PID of Rotor Variable Frequency Speed in Mine Hoisting System

ZHAO Shiyan, XIE Zidian, DING Kangkang, CUI Hanqing

Electronic Science and Technology. 2021, 34(1):  43-49.  doi:10.16180/j.cnki.issn1007-7820.2021.01.008

Abstract ( 176 )   HTML ( 6 )   PDF (965KB) ( 26 )  

Figures and Tables | References | Related Articles | Metrics

The control parameters of traditional PID controller applied in the mine hoist frequency conversion speed control system is fixed and difficult to be set, which aeaks leads to large speed overshoot and ripples of the electromagnetic torque and rotor flux linkage. In order to improve the performance of the system, an improved particle swarm optimization BP neural network PID controller algorithm of hoist rotor frequency conversion speed regulation system is proposed in the present study. The application of particle swarm algorithm which has fast convergence speed and global optimal characteristics in neural network can overcome the disadvantages of slow convergence speed and easy to fall into local optimum of BP neural network. In addition, the neural network convergence coefficient is designed to further accelerate the convergence speed. The simulation results show that the neural network control effect of particle swarm optimization is better than the neural network, and the effect is obviously better than the traditional PID controller. Compared with the neural network PID controller, the steady speed adjustment speed of the mine hoist speed regulation system of particle swarm optimization is obviously improved. Compared with the traditional PID controller, the ripple of the electromagnetic torque and rotor flux of the motor are significantly reduced, which has strong stability and robustness.

An Event-based Distributed Economic Scheduling Strategy for Micro-Grid

ZHU Qiaohong,LONG Yingwen,YU Su

Electronic Science and Technology. 2021, 34(1):  50-54.  doi:10.16180/j.cnki.issn1007-7820.2021.01.009

Abstract ( 236 )   HTML ( 7 )   PDF (735KB) ( 46 )  

Figures and Tables | References | Related Articles | Metrics

To solve the problems of heavy communication load and economic distribution of active power in distributed control in the traditional micro-grid system, an event-based active power control strategy is proposed. Based on the traditional consistent algorithm, each distributed power source adoptes the event triggering method, which effectively reduces the communication burden. Each distributed power source communicates with its neighbors through the event trigger controller and exchanges active power information with each other. The optimal active power value is obtained through the event trigger control algorithm, and the economic distribution of power is realized under the minimum cost. Finally, a micro-grid system model is established by MATLAB/Simulink platform, and the simulation experiment verifies the effectiveness of the control strategy.

Research on Complex Control Algorithms for Non-Minimum Phase Systems

WANG Jiaren,WANG Yagang

Electronic Science and Technology. 2021, 34(1):  55-59.  doi:10.16180/j.cnki.issn1007-7820.2021.01.010

Abstract ( 353 )   HTML ( 9 )   PDF (724KB) ( 53 )  

Figures and Tables | References | Related Articles | Metrics

For the undershoot caused by zeroes in the right-half plane of non-minimum phase systems, in order to suppress the undershoot and time lag of non-minimum phase systems, a two-step PID tuning method combined with the Smith predictive controller is proposed as a complex control algorithm. The algorithm achieves the effective of suppressing the undershoot of non-minimum phase systems by adjusting the pre-coefficient and PID parameters and applies it to the Smith predictive control structure to achieve the purpose of improving the system response time. The simulation model is established by using Simulink in MATLAB software to simulate a typical second-order control system. The simulation results show that the complex control algorithm eliminates the negative adjustment of the system, and verifies the effectiveness and robustness of the complex control algorithm.

Simulation Study of Anode Short Circuit 4H-SiC IGBT with Low Lifetime Region

MAO Hongkai, SU Fangwen, LIN Mao, ZHANG Fei, SUI Jinchi

Electronic Science and Technology. 2021, 34(1):  60-64.  doi:10.16180/j.cnki.issn1007-7820.2021.01.011

Abstract ( 251 )   HTML ( 4 )   PDF (955KB) ( 29 )  

Figures and Tables | References | Related Articles | Metrics

The traditional trench type 4H-SiC IGBT has a large turn-off loss, and it is difficult to compromise the turn-on voltage drop and turn-off loss. Aiming at this problem, a LS-IGBT structure was proposed, in which the emitter region contained a low-life region while a stepped collector was introduced by collector region to reduce the turn-off loss of the device. This method reduced the turn-off loss of a device by simultaneously controlling the number of hole carriers injected in the collector region and the lifetime of the P base region carriers, while maintaining the device breakdown voltage. The Silvaco Atlas device simulation tool was used to simulate the improved structure and compare it with the traditional structure. The simulation results showed that under the premise of consistent breakdown voltage, the turn-off loss of the new structure was increased by 84.5 % compared with the conventional structure, and the turn-on voltage drop of the device was reduced by 8.3%, which verified the correctness of the design.

3D Map Construction of Micro UAV Based on Depth Camera

ZHANG Zhenwei, ZHANG Wei, LONG Lin, YAN Chenhang

Electronic Science and Technology. 2021, 34(1):  65-70.  doi:10.16180/j.cnki.issn1007-7820.2021.01.012

Abstract ( 641 )   HTML ( 18 )   PDF (2892KB) ( 77 )  

Figures and Tables | References | Related Articles | Metrics

In order to solve the problem of obstacle avoidance and path planning when the small drone is insufficient in indoor light. In this paper, an indoor map construction system of UAV based on depth camera is designed and implemented. The pixhawk control board and low-cost embedded structure optical depth camera hardware platform are used to provide indoor environmental information for obstacle avoidance and path planning. Anti-sensor model algorithm is adopted, the information provided by the depth camera, pose sensor to filter are utilized to select and process effective obstacle information. A 3D map of the interior is established, and the depth camera acquires the description information of the obstacle point cloud by means of laser scanning. The position sensor is used to acquire the height information of the drone. The experimental results show the system could quickly acquire indoor maps, and the accuracy of the obstacles is relatively high, and it is not affected by light. The proposed system can be widely used for indoor navigation of drones, and realizes an indoor drone map construction system that does not rely on external light sources.