Application research of RRT algorithm path planning based on reinforcement learning

Jianjian Wei; Wenyan Chen

doi:10.1117/12.2652462

10 November 2022 Application research of RRT algorithm path planning based on reinforcement learning

Jianjian Wei, Wenyan Chen

Proceedings Volume 12331, International Conference on Mechanisms and Robotics (ICMAR 2022); 123314I (2022) https://doi.org/10.1117/12.2652462
Event: International Conference on Mechanisms and Robotics (ICMAR 2022), 2022, Zhuhai, China

Abstract

In order to solve the problems of the traditional fast search random tree (RRT) algorithm in complex environment, such as low planning efficiency, zigzagging generation path and many inflection points, a Q-RRT algorithm combined with reinforcement learning method was proposed. By introducing q-value table and reward function in Q-learning algorithm to influence the expansion process of RRT random tree, the randomness problem of tree node expansion is overcome. By defining reward function, the number of invalid nodes is reduced and the search efficiency of the algorithm is improved. At the same time, a path optimization method is proposed, which uses the pruning method and cubic Bezier curve to shorten the planned path length, realize the smooth processing of the path, and satisfy the motion constraints of the mobile robot. Experimental results show that the planning efficiency of Q-RRT algorithm is greatly improved compared with traditional RRT algorithm in complex environment, and the planned path is smoother, which can better adapt to the operation requirements of mobile robot.

Citation Download Citation

Jianjian Wei and Wenyan Chen "Application research of RRT algorithm path planning based on reinforcement learning", Proc. SPIE 12331, International Conference on Mechanisms and Robotics (ICMAR 2022), 123314I (10 November 2022); https://doi.org/10.1117/12.2652462

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available