Multiagent reinforcement learning for a planetary exploration multirobot system

Authors:
Zhang Zheng;Ma Shu-gen;Cao Bing-gang;Zhang Li-ping;Li Bin
Affiliations:
School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, China;School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, China;School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, China;School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, China;Robotics Laboratory, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China
Venue:
PRIMA'06 Proceedings of the 9th Pacific Rim international conference on Agent Computing and Multi-Agent Systems
Year:
2006

Citing 2
Cited 2

Introduction to Reinforcement Learning

Introduction to Reinforcement Learning
Reinforcement learning: a survey

Journal of Artificial Intelligence Research

Applying Reinforcement Learning to Multi-robot Team Coordination

HAIS '08 Proceedings of the 3rd international workshop on Hybrid Artificial Intelligence Systems
The world of independent learners is not markovian

International Journal of Knowledge-based and Intelligent Engineering Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

In a planetary rover system called “SMC rover”, the motion coordination between robots is a key problem to be solved. Multiagent reinforcement learning methods for multirobot coordination strategy learning are investigated. A reinforcement learning based coordination mechanism is proposed for the exploration system. Four-robot climbing a slope is studied in detail as an instance. The actions of the robots are divided into two layers and realized respectively, which simplified the complexity of the climbing task. A Q-Learning based multirobot coordination strategy mechanism is proposed for the climbing mission. An OpenGL 3D simulation platform is used to verify the strategy and the learning results.