Conflict Resolution of Cluttered Multi-robot Systems Using Metaheuristic Optimization Algorithms

Abstract

Conflict resolution becomes crucial when one is dealing with a large number of robots working in cluttered environments. The majority of the developed conflict resolution approaches in the literature deal with a motion-liveness problem which fails to gain collision-free movements for a large number of robots. This paper develops a systematic approach for coordinating the motions of multi-robot systems by adjusting their speeds to avoid collisions and guarantee motion-liveness of the robots. We mathematically formulate the multi-robot motion as a constrained optimization problem to minimize the time it takes for each robot to reach its target while avoiding collisions. Using two metaheuristic optimization methods, multiobjective genetic algorithm and particle swarm optimization, we can solve the conflict resolution problem up to 30 robots in a highly cluttered environment. Results show that we can find collision-free movements for a large number of robots in cluttered environments, while also guaranteeing multi-robot motion-liveness.