Adaptive extremum seeking controller via nonlinear variable gain for uncertainty model multirotor

In this paper, we propose an adaptive extremum search control (AESC) scheme for nonlinear underactuated multirotors in order to solve the cross-platform problem of adjusting the wheelbase and loads of multirotors. In AESC, an adaptive learning controller parameter reconstruction method is used, the adaptive extremum seeking cost function is employed to find optimal control parameters, and nonlinear variable gains (NLVG) are used to replace fixed gain in order to achieve real-time adaptive attitude and position tracking control. Experimental and numerical results with quadcopter were demonstrate the proposed AESC has superior tracking accuracy and robustness compared to the traditional cascaded PID control law. The simulation video are avaliable in Video-link: https://youtu.be/TRzlPReNUgY.