RRT-Based Path Planning for Continuum Arms

• Published in: IEEE robotics and automation letters Vol. 7; no. 3; pp. 6830 - 6837
• Main Authors: Meng, Brandon H., Godage, Isuru S., Kanj, Iyad
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.07.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Bending; Bending machines; Configuration space path planning; Cost function; Costs; Electron tubes; flexible robotics; Jacobian matrices; Kinematics; Motion and path planning; Obstacle avoidance; Path planning; Planning; Program processors; Robots; task and motion planning; Motion and Path Planning; Flexible Robotics; Task and Motion Planning
• ISSN: 2377-3766; 2377-3766
• DOI: 10.1109/LRA.2022.3174257

Continuum arms are bio-inspired devices that exhibit continuous, smooth bending and generate motion through structural deformation. Rapidly-exploring random trees (RRT) is a traditional approach for performing efficient path planning. RRT approaches are usually based on exploring the configuration space (C-Space) of the robot to find a desirable work space (W-Space) path. <styled-content style="color:black">Due to the complex kinematics and the highly non-linear mapping between the C-Space and W-Space of continuum arms, a high-quality path in the C-Space (e.g., a linear path) may not correspond to a desirable path/movement in the W-Space. Consequently, the C-Space RRT approaches that are based on C-Space cost functions do not lead to reliable and effective path planning when applied to continuum arms.</styled-content> In this letter, we propose a RRT* path planning approach for continuum arms that is based on exploring the W-Space of the robot as opposed to its C-Space. We show the successful applications of the proposed W-Space RRT* path planner in performing path planning with obstacle avoidance and in performing trajectory tracking. In all the aforementioned tasks, the quality of the paths generated by the proposed planner is superior to that of previous approaches and to its counterpart C-Space based RRT* approach, while the paths are generated in substantially less time.