2 September 2003 Framework for robot skill learning using reinforcement learning
Author Affiliations +
Proceedings Volume 5253, Fifth International Symposium on Instrumentation and Control Technology; (2003); doi: 10.1117/12.522188
Event: Fifth International Symposium on Instrumentation and Control Technology, 2003, Beijing, China
Abstract
Robot acquiring skill is a process similar to human skill learning. Reinforcement learning (RL) is an on-line actor critic method for a robot to develop its skill. The reinforcement function has become the critical component for its effect of evaluating the action and guiding the learning process. We present an augmented reward function that provides a new way for RL controller to incorporate prior knowledge and experience into the RL controller. Also, the difference form of augmented reward function is considered carefully. The additional reward beyond conventional reward will provide more heuristic information for RL. In this paper, we present a strategy for the task of complex skill learning. Automatic robot shaping policy is to dissolve the complex skill into a hierarchical learning process. The new form of value function is introduced to attain smooth motion switching swiftly. We present a formal, but practical, framework for robot skill learning and also illustrate with an example the utility of method for learning skilled robot control on line.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yingzi Wei, Mingyang Zhao, "Framework for robot skill learning using reinforcement learning", Proc. SPIE 5253, Fifth International Symposium on Instrumentation and Control Technology, (2 September 2003); doi: 10.1117/12.522188; https://doi.org/10.1117/12.522188
PROCEEDINGS
5 PAGES


SHARE
KEYWORDS
Switching

Computer science

Computer simulations

Machine learning

Neural networks

Psychology

Robot vision

RELATED CONTENT


Back to Top