Dynamic Switching of Multi-agent Formation in Unknown Obstacle Environment

Abstract

Problem statement: In the smart electromechanical system (SEMS), formation control of multi-agent system is a typical group coordination problem, which has broad application prospects in the fields of geographic survey, environmental investigation, security rescue, mine clearance and target defense. In the case of performing a certain task, multi-agent system need to transform its formation to avoid obstacles according to environmental constraints in real time. This paper introduces a method of dynamic switching formation structure of multi-agent system under unknown obstacle conditions. For example, multi-agent system which is currently in the form of a triangle formation need to change its formation into a straight line as it pass through a narrow, long passage. Purpose of research: This paper proposes a strategy for dynamic switching formation of multi-agent system in unknown environment. In this strategy, the Polya enumeration theorem is used to build the set of multi-agent system formation, and controlling the switch of formation with different obstacles conditions is based on the finite state machine theory. In the case of formation switching, the collision avoidance strategy between different agents is considered when the motion trajectory intersects. In addition, the target point assignment problem of one agent switching from the current formation structure to the target one is solved by the minimum distance method in this paper. Results: A formation switching control system of multi-agent system was developed, which includes a network of interacting finite-state machines and a set of formation topologies, and an underlying control strategy. In order to verify the correctness and operability of the proposed method, simulation and real experiments were carried out. The results show that the multi-agent system can switch from one topology to another according to the shape of the obstacle during the movement, and avoid collision with each other during the formation switching. Practical significance: According to the proposed solution, the respond ability of group robots to changing environments in performing tasks is improved.