Function observer based on model compensation control for a fixed-wing UAV

Abstract

Fixed-wing unmanned aerial vehicle (UAV) is a nonlinear system characterized by strong coupling, high complexity, and sensitivity to external interference and model uncertainty. For the fixed-wing UAV having eight degrees of freedom, the uncertainty and wind disturbance parts of the model are clarified, and a model with lump disturbance is established through the mechanic analysis. To get the high performance of control, a highly accurate observer for the lump disturbance is necessary. A second-order compensation function observer (CFO) with feedforward is proposed with new structure. The CFO accuracy is much higher than the same order of the extended state observer (ESO) by convergence analysis. A hyperbolic tangent function is combined with the CFO for effectively reducing the large estimation error at the initial moment. A CFO-based model compensation control (MCC) algorithm is provided for controlling the longitudinal and lateral channels of the fixed-wing UAV. The MCC contains the known model, the compensation of unknown model via the CFO, feedback of error and the dynamic information of the given information. The MCC realizes the desired linear characteristics for the feedback system because it decouples the channels, restrains the lump disturbance. The stability of the closed-loop system is proved by Lyapunov theorem. The proposed CFO-based MCC is applied to the fixed-wing UAV, and compared with the traditional PID and ESO-based on sliding mode control (SMC) algorithm through simulation and the control test platform experiment based on Pixhawk, showing the superiority of the MCC and highly accuracy of the CFO.