Adaptive Neural Network Extended State Observer-Based Finite-Time Convergent Sliding Mode Control for a Quad Tiltrotor UAV

Abstract

The internal and external disturbance of the quad tiltrotor UAV stem from its rotor tilting motion and wind disturbance, which affect the control stability of the quad tiltrotor UAV in full flight. This paper proposes a control law design method based on the Adaptive Neural Network Extended State Observer-based Finite-Time Convergent Sliding Mode Control (ANNESO-FTCSMC). The adaptive neural network obtains the unknown total disturbance term that needs to be compensated in the finite-time convergent sliding mode control. An extended state observer is used to estimate the state variables of the controlled plant. The stability of the ANNESO-FTCSMC controller is proved with the Lyapunov stability theory. The unknown total disturbance as the expanded state variable of the state observer has the advantage of high adaptability, and the ANNESO-FTCSMC has a strong ability to compensate for the internal and external disturbances of the system. Compared with traditional sliding mode control, ANNESO-FTCSMC does not contain switching terms, so ANNESO-FTCSMC is chattering-free and converges faster. Based on this method, the flight control system of the quad tiltrotor UAV is designed and verified by hardware-in-loop simulation. The hardware-in-loop simulation results of attitude control in different flight modes and trajectory tracking control in full flight show that the ANNESO-FTCSMC controller is suitable for the flight control of quad tiltrotor UAV. Compared with the PID controller and ADRC controller, ANNESO-FTCSMC controller has more stable flight state of quad tiltrotor UAV under different flight modes and the same total disturbance. The transition process of ANNESO-FTCSMC controller from helicopter mode to airplane mode is also smoother and will not have large fluctuations. Therefore, ANNESO-FTCSMC controller has strong anti-disturbance and robustness.