Decentralized Control Strategy for Flocking of Multi-Agents in Polygonal Configuration in the Presence of an Obstacle

The avoidance of obstacles is a challenging problem during flocking and formation of multi-agents in a real-time environment. The technique of split and rejoin manoeuvres used in the earlier studies for obstacle avoidance cannot handle simultaneous flocking and formation of multi-agents. The concept of simultaneous polygonal formation and flocking configuration (in the presence of obstacle) will be of considerable importance for tracking of navigation in closed formations and coverage control of a multi-agent surveillance system. This paper proposes a decentralized control strategy for simultaneous flocking and formation of multi-agents in the presence of obstacle. The rationale for the proposed decentralized control strategy (Reinforcement Learning with adaptive strategy) is to achieve stability of flocking and octagonal formation configuration of multi-agents in a quicker response time. Simulation results substantiate that with the proposed decentralized control strategy, average position error lies within the position band and flocking error converges faster to zero in 0.2s.