A novel RRT*-Connect algorithm for path planning on robotic arm collision avoidance

Abstract

To address the limitations of the original algorithm, several optimization techniques are proposed. This article presents an original RRT*-Connect algorithm for the planning of obstacle avoidance paths on robotic arms. These strategies include implementing a target biasing algorithm, using elliptic space sampling to enhance the sampling process, the revision of the cost function to better guide path planning, and implementing an artificial potential field and gradient descent strategy to design adaptive step sizes. Furthermore, the use of segmented Bézier curves facilitates the generation of a more fluid trajectory when constructing the final path. The effectiveness of these augmentation strategies is corroborated by both simulations and experimental verification on a robotic arm. The simulations showed a 19.39% reduction in average run time and a 5% reduction in average path length compared to the existing RRT*-Connect algorithm. Therefore, The enhanced algorithm meets the requirement for optimal obstacle avoidance path planning by consistently finding the shortest path while avoiding obstacles.