An evolutionary artificial potential field algorithm for stable operation of a multi-robot system on domes

Abstract

In this paper, we propose a reasonable fast centralized to navigate a multi-robot system on domes. The robots, consisting of a leader and N-1 (N ≥ 2) followers, are connected to each other by N strings and make a ring around the dome. The leader uses the artificial potential field approach to plan a stable path toward its desired goal position while the other robots support it. To avoid getting into local minima we use genetic algorithm(GA) and multi-rapidly exploring random trees (multi-RRTs) before the robots get stuck into local minima. Our method guarantees the leader can reach its goal on the dome and all robots remain stable throughout the path. Finally, we have simulated the behavior of a 4-robot system on a spherical dome to show the effectiveness of the method.