Using Relative Localization Observations for Swarm Robots Search

摘要

Swarm robots searching for potential target in unknown environment with relative locations observed among robots is studied due to the absence of global localization mechanism.Comparing the similarities and differences of properties,we propose a concept of mapping between ideal particle in Particle Swarm Optimization algorithm and real robot in swarm robots search operation.With the definitions of simple behaviour rules and neighbourhood structure,we extend the Particle Swarm Optimization to model swarm system at the microscopic level and design control algorithm in terms of the swarm intelligence principles.As robot is very limited capable of sensing and communicating,the best history recognition position of each robot is decided to be either last or current moment by comparison of its detecting values at the two instants.Similarly,the local social optimization of robot is determined by comparing senses of all members belonging to the same neighbourhood.Based on the relative distance and relative bearing both between robot and its own history optimization and between robot and its local social optimization,the position of robot at next moment will be decided under individual coordinates rather than the world one.Finally,the simulation results indicate the validity of the control strategy presented.