An Improved Algorithm for Complete Coverage Path Planning Based on Biologically Inspired Neural Network
Linhui Han, Xiangquan Tan, Qingwen Wu, Xu Deng
Abstract
Complete coverage path planning (CCPP) requires the mobile robots to traverse every part of the workspace, which is one of the major challenges in cleaning robots and many other robotic systems. The biologically inspired neural network (BINN) algorithm has been extensively applied in path planning, recently. In this article, a new CCPP strategy with BINN is proposed. The planned path of cleaning robot is not only determined by the dynamic neural activities but also by the distribution of obstacles in the environmental map. By distinguishing the connectivity between different areas of the environmental map, and using the proposed path backtracking algorithm, the improved CCPP algorithm can autonomously plan a collision-free path and reduce the path repetition ratio. Besides, an improved dynamic deadlock escape algorithm is presented to select the optimal escape target point. The simulation results show that the proposed CCPP algorithm without any templates or learning procedures is able to generate an orderly path in both known and unknown environment.