Formation control of multirotor aerial vehicles using decentralized MPC

Abstract

In this work, the authors propose a formation control strategy of a group of three multirotor aerial vehicles being able to avoid multiple obstacles and collisions. To deal with this problem, a decentralized architecture is proposed which has one model predictive controller per vehicle including a set of convex constraints on the vehicle’s position to prevent collisions with other agents and different shapes of obstacles. The resulting decentralized scheme controls the formation based on a virtual structure approach. For the purpose of avoiding collisions, each local controller considers the predicted position of every neighbor vehicles. The effectiveness of the developed scheme is demonstrated through numerical simulations considering a “figure-of-eight” as the reference trajectory, and the results show its capability to handle thrust force, obstacle and collision avoidance constraints.