Improved informed-RRT* combined with pruning strategy for manipulator obstacle avoidance planning

Abstract

Efficient and reliable obstacle avoidance path planning is crucial for robotic arms operating in complex environments. This paper addresses the path planning problem of the PUMA560 robotic arm in scenarios with cuboid and spherical obstacles. We investigate the performance improvement of an improved Informed-RRT* algorithm combined with a pruning strategy over the original RRT. The algorithm rapidly obtains an initial path via a guiding mechanism, followed by progressive optimization and path pruning, ultimately generating a smooth and collision-free trajectory. Three scenarios with different obstacle avoidance difficulties (cuboid slight blocking, cuboid full blocking, spherical full blocking) are designed. Experimental results show that, compared with RRT, the improved Informed-RRT* algorithm slightly reduces the path cost but significantly improves planning efficiency and stability (e.g., planning time reduced by approximately 26% in the slight blocking scenario). The pruning step further reduces the path cost substantially, by approximately 40% on average across all three scenarios, with negligible time overhead, showing strong reliability across scenarios. The method is simple to implement, significantly improves path compactness and smoothness, and is well suited for practical robotic arm obstacle avoidance tasks.