Research on Improved Artificial Potential Field Path Planning Based on APF-CPRM Fusion Algorithm
Hao Zhai, Chen Zhen, Ying Lu · 2024
In response to issues such as falling into local minima, unreachable targets and excessive obstacle avoidance based on artificial potential field (APF), this paper presents a path planning algorithm based on improved circle probabilistic roadmaps method (CPRM) and artificial potential field method. Initially, the excessive nodes problem in the probabilistic graph approach is addressed by optimizing it with circular restricted domain constraints. Subsequently, the APF method is used to construct a potential field model that integrates target gravity and obstacle repulsion to produce path predictions on a two-dimensional map. When APF falls into the minimum value problem, the improved probability circuit diagram algorithm is introduced to solve the issues related to local minima, unreachable targets, and excessive obstacle avoidance. Simulation results show that the APF-CPRM path planning algorithm can well integrate the two algorithms and effectively improve the related defects. It reduces the number of redundant nodes in traditional PRM and the unsolvable local minimum issue in traditional APF algorithms is resolved. Moreover, with the same computer resources consumed, the time and the number of nodes are reduced by 61.6% and 75%, respectively. Compared with the unimproved fusion algorithm, the success rate and quality of path planning are higher.