Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:A3L-D
Session:
Number:A3L-D-02
Mental Simulation on Reservoir Computing as an Efficient Planning Method for Mobile Robot Navigation
Yoshihiro Yonemura , Yuichi Katori,
pp.83-86
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.A3L-D-02
PDF download (971.3KB)
Summary:
Machine learning methods have been applied for autonomous mobile robot navigation. Despite the achievement of the methods, their learning cost is the most significant remaining problem. We propose a mental simulation framework on reservoir computing to perform efficient learning and action planning. Mental simulation is a process that simulates the interaction between the model and the environment. Reservoir computing is appropriate for mental simulation because it can process complex time series efficiently. In this research, we implemented action planning with mental simulation on reservoir computing, and we confirmed that the robot could reach the target point by the planning.