Bidirectional Potential Guided RRT* for Motion Planning

Abstract

Requirement for high accuracy and speed of grasping operation for motion planning is very important. Motion planning algorithms for avoiding obstacles in narrow channels play a vital role for robotic arm effectively operating grasp tasks. The potential function-based RRT*-connect (P-RRT*-connect) algorithm for motion planning is presented by combining the bidirectional artificial potential field into the rapidly exploring random tree star (RRT*) in order to enhance the performance of the RRT*. The motion path is found out by exploring two path trees from the start node and destination node, respectively, with the rapidly exploring random tree star. Two trees advance each other at the same time according to the attractive potential field and the repulsion potential field generated by the artificial potential field method of sampled nodes until they meet. The P-RRT*-connect algorithm is especially suitable for solving the problem of narrow channels. The simulation results prove that the P-RRT*-connect algorithm is more efficient than potential Function-based RRT* (P-RRT*) regardless of the number of iterations or the running time. The experimental data show that the time for the P-RRT*-connect to find the optimal path from the starting node to the target node is half than that of the P-RRT*, and the number of iterations of the P-RRT*-connect is also about one-third less than that of the P-RRT* which is useful for real time.INDEX TERMS Potential function based RRT*-connect, RRT*, artificial potential field, two path trees, narrow channels.

Comparison of Multiple Reinforcement Learning and Deep Reinforcement Learning Methods for the Task Aimed at Achieving the Goal

“…Motion planning is one of the most fundamental research topics in robotics. Some of the approaches have used traditional planning methods, such as RRT [18], RRT* [31], where structured tree methods are used to find the curve from point A to point B. Other approaches use modern techniques, such as RL/DRL, but both methods use Bézier curves to characterize complex trajectories and smooth motion planning [23].…”

“…Planning the trajectory of the robotic arm as one of the most basic and challenging research topics in robotics has found considerable interest from research institutes in recent decades. Traditional task and motion planning methods, such as RRT [18], RRT* [31] can solve complex tasks but require full state ob-servability, a lot of time for problem solving and are not adapted to dynamic scene changes. Advanced RL/DRL techniques can solve motion planning tasks for multiple-axis industrial robot [23].…”

Endüstriyel Robot Hareket Planlama Algoritmaları Performans Karşılaştırması

“…Bu algoritmalar ROS temelli robotik uygulamalarda kullanılabilirliği yüksek, yazılım geliştirmelerine açık algoritmalar olduğu için, günümüzde çok daha yaygın kullanılmaya başlamıştır [1]. Xinyu ve ekibi [25] OMPL'in algoritmalarından RRT* algoritmasının çift yönlü hareketteki etkisi üzerine çalışmıştır. Çalışmasında RRT* algoritmasının oldukça optimize olduğunu ancak işlem süresinin uzun ve yavaş olduğuna dikkat çekmiştir.…”

HPPRM: Hybrid Potential Based Probabilistic Roadmap Algorithm for Improved Dynamic Path Planning of Mobile Robots

“…For instance, the RRT* algorithm [43] is a variation of the original single-tree RRT that continually rewires the search tree for the shortest path search. Bidirectional RRT [44], [45] uses a bi-directional tree search for faster route planning. Similarly, [46], presented Lazy PRM, an improved PRM that minimizes the number of collision checks performed during the planning and therefore minimizes the running time of the planner.…”