Consensus-Based Distributed Collective Motion of Swarm of Quadcopters

Abstract

This article presents a scalable hierarchical distributed control framework to address the problem of collective navigation of a quadcopter swarm, which is a special form of Internet of Vehicles (IoV), in the presence of member drop-outs and interrobot communication link failures. The proposed control framework has three parts: 1) the swarm descriptors (the geometric interpretation of swarm statistics) estimator; 2) the cooperative swarm motion controller; and 3) the attitude controller. The proposed framework includes a novel distributed dynamic average consensus algorithm to estimate the swarm descriptors which represent the collective motion of the swarm in the abstract space (a lower dimensional space independent of the number and permutation of robots in the swarm). By employing a dynamic inversion approach, in the cooperative swarm motion controller, we design a control law that generates the desired thrust, yaw, pitch, and roll angle commands. In the attitude controller, we convert the desired commands from the cooperative swarm motion controller into the rolling, pitching, and yawing moments required to realize the collective motion of the swarm. The robustness of the proposed framework and the stability of the proposed swarm descriptors estimator, the cooperative swarm motion controller, and the attitude controller, and the overall cascaded system are mathematically proved. The significance of the proposed control framework is demonstrated via simulation in the presence of member robot drop-outs and interrobot communication link failures.