Efficient-Low Memory Path Planning Algorithm Based on Adaptive Thresholding

Abstract

Aiming at the problem of low efficiency and large memory consumption in mobile robot path planning, this paper proposes an improved version of the RRT*Fixed Nodes (RRT*FN)algorithm based on the adaptive threshold, named A-RRT*FN. The algorithm constructs a sampling tree in the two directions of the starting point, and the target point and adaptively sets the threshold according to the environment on both sides to improve the environmental adaptability and search speed of the algorithm. Secondly, a node elimination strategy is proposed to quickly locate low-performance nodes and reduce the number of iterations of the algorithm and the memory occupation of low-performance nodes. Finally, the path smoothing strategy is used to smooth the obstacle-free path to obtain a path more suitable for robot tracking control. The proposed algorithm is compared with RRT* FN, Fast-RRT, and Improved Bi-RRT* algorithms under three different maps of simple, narrow, and complex. The results show that the proposed algorithm has better performance in convergence efficiency and memory consumption.

Keywords

• Adaptive thresholding
• path planning
• RRT*FN algorithm
• memory allocation mechanism