Authors: Songdong Xue, Jianchao Zeng
Addresses: College of Electrical and Information Engineering, Lanzhou University of Technology, Lanzhou 730050, P.R. China; Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, Taiyuan 030024, P.R. China. ' Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, Taiyuan 030024, P.R. China
Abstract: To control swarm robots with extended particle swarm optimisation approach for target search, target signals should be detected and fused as fitness evaluate due to the inherent parallel processing property caused by spatial interspersed of robots in search environment. Also, differences in sampling frequency of sensors and communication delays make it realistic to control such swarm systems asynchronously. Therefore, two asynchronous update principles, i.e., the communication cycle-based and evolution position-based control strategies are presented in case of target search. Besides, a concept of time-varying character swarm is proposed to facilitate decision-making on the best-found position. Each robot detects signals in a fine-grained parallel way and compares fusion of signals with the best in its character swarm. Then velocities and positions of individual robots are updated immediately. But the shared information within character swarm is updated asynchronously according to different control principles only. Simulation results indicate that the communication cycle-based strategy has advantage over the evolution position-based control strategy in search efficiency.
Keywords: swarm intelligence; swarm robots; target search; extended PSO; particle swarm optimisation; parallel process; asynchronous control; robot control; information sharing; simulation.
International Journal of Modelling, Identification and Control, 2009 Vol.8 No.4, pp.353 - 360
Available online: 09 Dec 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article