Authors: Chun-Ying Wu; Wen-De Ke; Xiao-Guang Li
Addresses: Department of Sports, Guangdong University of Petrochemical Technology, Maoming 525000, China ' Department of Computer Science, Guangdong University of Petrochemical Technology, Maoming 525000, China ' Information and Telecommunication Branch, Qiqihar Power Supply Company, Qiqihar 161005, China
Abstract: In an effort to solve to the problems of dynamic relative mental, learning ability and moving time optimisation in the cooperation of multi-robots, a layer-dividing task cooperation model is proposed. Firstly, the changing features of the mental layer in a dynamic environment is shown through the belief, target, intention and knowledge. Secondly, the strategy learning and task planning under the condition of maximum united rewards are constructed by combining the reinforcement learning and particle swarm optimisation. Thirdly, the robot's fast moving method is designed based on the basic movement states and Bezier curve track. The validity of the model is proved through experiments.
Keywords: multi-robots; cooperation; layer division; learning.
International Journal of Wireless and Mobile Computing, 2017 Vol.13 No.2, pp.144 - 150
Received: 17 Nov 2016
Accepted: 12 May 2017
Published online: 14 Nov 2017 *