Research on Obstacle Avoidance of Substation Robot Based on Spatiotemporal Networks

doi:10.16182/j.issn1004731x.joss.24-0081

Abstract

Abstract:

In order to improve the visual obstacle avoidance ability of substation robots in complex environments, a robot visual obstacle avoidance method based on spatiotemporal networks is proposed. The method utilizes traditional image processing techniques to enhance road information and designs a lightweight deep convolutional neural network structure to extract road features from a spatial domain perspective; based on the spatial characteristics of the road, a long short-term memory network is introduced to mine the changes in the road from a temporal perspective, and a classification regression prediction structure is used to predict the robot's obstacle avoidance direction and angle; considering the highly repetitive characteristics of substation roads, a feature filtering module is introduced to reduce redundant feature calculations and ensure real-time performance during network applications. The experimental results show that the proposed method can effectively extract spatiotemporal features of substation road scenes, this method can more accurately predict the next action of the robot and better complete navigation obstacle avoidance tasks.

Key words: substation robot, spatiotemporal network, attention structure, classification regression prediction, similar feature filtering

CLC Number:

TP391.41

Cheng Chong, Wang Lixia, Duan Songtao, Xiong Xiaoguang, Ge Xianjun. Research on Obstacle Avoidance of Substation Robot Based on Spatiotemporal Networks[J]. Journal of System Simulation, 2025, 37(6): 1522-1530.

Figures/Tables 15

Fig. 1

Fig.2

Fig. 3

Fig. 4

Fig. 5

Table 1

Table 2

Table 3

Table 4

Table 5

Fig. 6

Fig. 7

Table 6

Table 7

Table 8

References 18

1	张培. 智能变电站巡检机器人的应用[J]. 现代制造技术与装备, 2022, 58(6): 222-224.
	Zhang Pei. Briefly Analyze the Application of Intelligent Substation Inspection Robot[J]. Modern Manufacturing Technology and Equipment, 2022, 58(6): 222-224.
2	Dong Lijun, Chen Naichao, Liang Jiawen, et al. A Review of Indoor-orbital Electrical Inspection Robots in Substations[J]. Industrial Robot, 2023, 50(2): 337-352.
3	毛建旭, 贺振宇, 王耀南, 等. 电力巡检机器人路径规划技术及应用综述[J]. 控制与决策, 2023, 38(11): 3009-3024.
	Mao Jianxu, He Zhenyu, Wang Yaonan, et al. Review of Research and Applications on Path Planning Technology for Power Inspection Robots[J]. Control and Decision, 2023, 38(11): 3009-3024.
4	唐昀超, 祁少军, 朱立学, 等. 移动机器人避障运动研究[J]. 系统仿真学报, 2024, 36(1): 1-26.
	Tang Yunchao, Qi Shaojun, Zhu Lixue, et al. Obstacle Avoidance Motion in Mobile Robotics[J]. Journal of System Simulation, 2024, 36(1): 1-26.
5	Wang Lei, Yuan Luhai, Zhang Xin. Research on Navigation and Control Method for Inspection Robot in Smart Substation[J]. Journal of Physics: Conference Series, 2022, 2330(1): 012014.
6	黄绪勇, 王欣, 唐标, 等. 基于改进人工势场法的电力巡检机器人自动避障轨迹规划[J]. 机械与电子, 2023, 41(4): 22-25, 30.
	Huang Xuyong, Wang Xin, Tang Biao, et al. Automatic Obstacle Avoidance Trajectory Planning for Power Inspection Robot Based on Improved Artificial Potential Field Method[J]. Machinery & Electronics, 2023, 41(4): 22-25, 30.
7	何少华. 变电站巡检机器人复合导航与路径规划系统的设计[D]. 哈尔滨: 哈尔滨理工大学, 2023.
	He Shaohua. Design of Compound Navigation and Path Planning System for Substation Inspection Robot[D]. Harbin: Harbin University of Science and Technology, 2023.
8	Green J, Clounie J, Galarza R, et al. Optimization of an Intelligent Wheelchair: LiDAR and Camera Vision for Obstacle Avoidance[C]//2022 22nd International Conference on Control, Automation and Systems (ICCAS). Piscataway: IEEE, 2022: 313-318.
9	Hu Kewei, Chen Zheng, Kang Hanwen, et al. 3D Vision Technologies for a Self-developed Structural External Crack Damage Recognition Robot[J]. Automation in Construction, 2024, 159: 105262.
10	鲜开义, 彭志远, 谷湘煜, 等. 变电站巡检机器人避障方法研究与应用[J]. 科学技术与工程, 2021, 21(5): 1957-1962.
	Xian Kaiyi, Peng Zhiyuan, Gu Xiangyu, et al. Research and Application of Obstacle Avoidance Method for Substation Inspection Robot[J]. Science Technology and Engineering, 2021, 21(5): 1957-1962.
11	王志辉, 陈息坤. 融合场景理解与A^*算法的巡检机器人避障设计[J]. 无线电工程, 2022, 52(11): 2000-2008.
	Wang Zhihui, Chen Xikun. Obstacle Avoidance Design of Inspection Robot Based on Scene Understanding and A^* Algorithm[J]. Radio Engineering, 2022, 52(11): 2000-2008.
12	Venkat Raman Nagarajan, Singh Pavan. Obstacle Detection and Avoidance for Mobile Robots Using Monocular Vision[C]//2021 8th International Conference on Smart Computing and Communications (ICSCC). Piscataway: IEEE, 2021: 275-279.
13	Li Yong, Liu Yuzhe. Vision-based Obstacle Avoidance Algorithm for Mobile Robot[C]//2020 Chinese Automation Congress (CAC). Piscataway: IEEE, 2020: 1273-1278.
14	Shi Xingjian, Chen Zhourong, Wang Hao, et al. Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting[EB/OL]. (2015-09-19) [2023-11-21]. .
15	Wang Xiaolong, Girshick R, Gupta A, et al. Non-local Neural Networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7794-7803.
16	Howard A, Sandler M, Chen Bo, et al. Searching for MobileNetV3[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway: IEEE, 2019: 1314-1324.
17	Tan Mingxing, Le Q. EfficientNetV2: Smaller Models and Faster Training[C]//Proceedings of the 38th International Conference on Machine Learning. Chia Laguna Resort: PMLR, 2021: 10096-10106.
18	Tang Yehui, Han Kai, Guo Jianyuan, et al. GhostNetV2: Enhance Cheap Operation with Long-range Attention[EB/OL]. (2022-11-23) [2023-11-27]. .

数据集	数量	类别数	大小/像素
CamVid	701	2	360×480
Mnist	10 000	100	28×28
MovingMnist	1 000	—	448×448
RobotData	6 000	4	512×512

超参数	参数赋值
优化策略	SGD
学习策略	step
初始学习率	0.01
动量参数	0.9
权重衰减系数	0.005
分类损失函数	SoftMaxLoss
回归损失函数	MeanSquaredLoss

RGB	HSV	Sobel	Gabor	G/%	U/%	V/(帧/s)
√				91.1	83.3	25
√	√			91.6	83.5	25
√	√	√		91.8	83.8	24
√	√	√	√	92.3	84.2	24

网络	G/%	U/%	V/(帧/s)
MobileNet^[17]	92.3	84.2	24
EfficientNet^[17]	92.8	84.6	22
GhostNet^[18]	93.1	84.8	21
基础网络	90.6	82.3	33
基础+残差	91.2	82.8	32
基础+残差+注意力	91.7	83.7	31

网络	R	P/%	V/(帧/s)
LSTM	1.8	—	64
LSTM+自注意力	1.6	—	63
LSTM+自注意力+分类	1.6	95.6	63
LSTM+自注意力+分类+特征加权	1.3	96.2	61