Adaptive Sliding Mode Control with Transition Process for Flapping Wing Aerial Vehicle

The flapping wing aerial vehicle (FWAV) is an unmanned aerial vehicle that is greatly affected by aerodynamic changes. The uncertainties of its aerodynamic model and the complicated external airflow disturbances make the existing controllers unable to meet the system performance requirements. Therefore, this paper presents a novel adaptive sliding mode controller framework based on cascade control and online learning. First, a virtual guided sliding surface was established to smooth the unstable output of the control quantity and eliminate chattering to avoid aggravating the airflow disturbances caused by the FWAV attitude changes. Then, a reinforcement learning method is designed to optimize the guided process established online, and actively compensates for unknown external disturbances and model uncertainties. Finally, this paper designs an attitude control system including an FWAV approximation model and unknown disturbances to verify the effectiveness of this adaptive framework. Compared with conventional controller optimization methods, simulation results show that the adaptive framework presented has better robustness and control accuracy, which can reduce the sensitivity of an FWAV to aerodynamic changes and attenuate chattering.