Data-Driven Learning and Receding Horizon Control for Quadrotors

This paper presents a control framework for the problem of unknown signal source localization based on quadrotors. There are two levels for the proposed control framework: a decision level and an execution level. In the decision level, the gaussian process regression approach is first used to provide the possible locations of all signal sources. Then, the reference trajectory of quadrotors is planned by the RHC approach. In the execution level, the reference trajectory is fed to the quadrotor controller, where position controllers and attitude controllers are designed with the PID controllers. Finally, through simulation experiments, the effectiveness of the proposed control framework is verified.