Obstacle Avoidance Path Planning for Power Inspection Robots based on Deep Learning Algorithms

Main Article Content

Yuxin Liu
Xiaoxi Ge
Haowei Jia
Lin Yuan
Min Zhou

Abstract

The current research on obstacle avoidance path planning methods for power inspection robots has problems such as poor obstacle avoidance ability and poor inspection effectiveness. Therefore, a planning method for obstacle avoidance path of power inspection robots is proposed. By utilizing motion relationships and the potential field theorem of robot motion, a three-dimensional model of the power inspection robot’s route is established to determine the direction of the robot’s route when obtaining action tasks. The fuzzy support vector algorithm is used to plan obstacle avoidance paths for the initialized walking path, making the inspection robot intelligent. The experimental results show that the average success rates for avoiding static and dynamic obstacles are 98.37% and 96.12%, respectively. The average time for obstacle avoidance path planning is 1.56 seconds, and it has fast, efficient, and accurate obstacle avoidance and path planning capabilities, which can improve the robot’s obstacle avoidance ability and path planning efficiency for dynamic and static obstacles.

Article Details

Section
Special Issue - Deep Learning-Based Advanced Research Trends in Scalable Computing