Formation Reconfiguration of Cooperative UAVs via Learning Based Model Predictive Control in an Obstacle-Loaded Environment

摘要

Learning Based Model Predictive Control (LBMPC) is a new control policy that combines statistical learning along with control engineering while providing levels of guarantees on safety, robustness and convergence. The designed control policy respects the general rules of flocking such that when static obstacles appear, the UAVs are required to steer around them and also avoid collisions between each other. Also, each UAV in the team match the other team members velocity and stay close to its flockmates during flight. Our main contribution in this paper lays in solving the formation reconfiguration problem for a group of N cooperative UAVs forming a desired formation using LBMPC in the presence of uncertainties and obstacles in simulation.