Controlling swarm robots with kinematic constraints for target search

An approach to control artificial swarm whose members are autonomous wheeled mobile robots is proposed, by applying Particle Swarm Optimization (PSO) to target search. First, swarm search is mapped to PSO based on similarities between the two cases. Then a distributed PSO-style algorithm is given, in which decision making on real inputs of linear and angular velocity of robot controller being explored. We obtain the required command sequences by constraining the computational expected velocities and positions with robot's non-holonomic properties in kinematics. In this way, swarm robots can work together cooperatively.