Nonholonomic Vehicle Motion Planning by Generating a Mapping from Configuration to Control Output with Reinforcement Learning

Author

K. Naruse
Ming-Chuan Leu, Missouri University of Science and TechnologyFollow

Editor(s)

Stelson, K. and Oba, F.

Abstract

This paper presents the investigation of applying reinforcement learning to nonholonomic robot motion planning in an uncertain workspace. We propose a planning system for the nonholonomic robot, in which the continuous configuration and control output of the robot are discretized, and the reinforcement learning algorithm generates a mapping from the configuration to the output. The application results are presented with computational simulations.

Recommended Citation

K. Naruse and M. Leu, "Nonholonomic Vehicle Motion Planning by Generating a Mapping from Configuration to Control Output with Reinforcement Learning," Proceedings of the Japan/USA Symposium on Flexible Automation, American Society of Mechanical Engineers (ASME), Jan 1996.

Meeting Name

1996 Japan-USA Symposium on Flexible Automation Part 2

Department(s)

Mechanical and Aerospace Engineering

Keywords and Phrases

Computer Simulation; Learning Algorithms; Motion Planning; Obot Learning

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 1996 American Society of Mechanical Engineers (ASME), All rights reserved.

Publication Date

01 Jan 1996