Improved RRT-APF algorithm for manipulator path planning

HU Haoxuan; ZHANG Chunyan

doi:10.16578/j.issn.1004.2539.2026.06.014

Abstract：[Objective]Aiming at the problems of local optimum,insufficient path smoothness and high computational complexity existing in the traditional rapidly-exploring random tree-artificial potential field(RRT-APF)path planning algorithm in industrial environments,an adaptive optimized RRT-APF(AORRT-APF)path planning algorithm was proposed to improve the path planning efficiency and path quality of manipulators in complex industrial environments.[Methods]Firstly,a goal-biased sampling strategy and an improved metric function were introduced to reduce the generation of redundant nodes and accelerate the convergence speed of the algorithm.Secondly,a dynamic parameter adjustment mechanism and an adaptive obstacle processing mechanism were constructed to enhance the adaptability of the algorithm to dynamic environments.Then,a path smoothing method combining cubic spline interpolation and collision detection was adopted to generate continuous and executable motion trajectories.Finally,two-dimensional and three-dimensional simulation environments were built based on Matlab software,and physical verification was carried out with the six-degree-of-freedom Fr3 manipulator as the platform.[Results]The results show that in the three-dimensional random environment,the running time of the AORRT-APF algorithm is only 0.059 s,which is 88.7%shorter than that of the traditional RRT algorithm,and the average number of iterations is reduced to 175,which is 88.8%less than that of the RRT algorithm.The number of inflection points of the generated path is significantly reduced,and the manipulator can smoothly complete the obstacle avoidance movement along the planned trajectory.This algorithm can provide a reference for improving the working efficiency and operational safety of industrial manipulators.