Distributed herding of multiple robots in cluttered environments

초록

This article introduces distributed herding controllers to make multiple robots (agents) follow a leader in environments with many unknown obstacles. The proposed herding controllers are developed considering agents which can only measure the bearing of a nearby agent using bearing sensors, such as camera. In addition, an agent uses Received Signal Strength Indicator (RSSI) sensors to detect the moment when it meets another agent. Every agent, except for the leader, is not equipped with sensors to localize itself. Every agent can only detect nearby agents and moves utilizing local sensing measurements. Since every agent moves in an unknown cluttered environment, communication or sensing between agents may be lost blocked by obstacles. Therefore, this article presents distributed herding controllers so that multiple agents follow a leader while maintaining network connectivity in unknown cluttered environments. The leader changes its velocity adaptively so that it can reach the goal while maintaining network connectivity with its followers. The proposed herding controllers are developed to overcome a network faulty situation. As far as we know, this paper is novel in presenting distributed herding controls of multiple agents, such that the network connectivity is maintained while agents move in cluttered environments. Under MATLAB simulations, we verify the effectiveness of the proposed herding approach in cluttered environments. © 2021 Elsevier B.V.

키워드