Attitude Tracking Control for a Quadrotor via backstepping and Adaptive Dynamic Programming

An online-learning control algorithm based on Backstepping control and action-dependent heuristic dynamic programming (ADHDP) is proposed for quadrotors to facilitate a satisfied control performance. In the paper, backstepping controller is used as the main controller and ADHDP controller is worked as the complementary controller. The ADHDP control strategy is designed based on the difference between the real attitude value response from the basic Backstepping control and the desired attitude reference value. An actor-critic neural network is served as the primary structure in ADHDP controller, which has the ability of online adaptive learning. The simulation results in Matlab/Simulink are presented to verify the effectiveness and robustness of the proposed control technology.