Research on trajectory tracking control method based on disturbance observer and MPC

Abstract

This paper investigates a disturbance observer combined with model predictive control (MPC) for trajectory tracking of quadrotor UAVs against model parameter perturbations, unmolded dynamics and external wind disturbances in complex low-altitude scenarios. Based on the small attitude angle assumption, the nonlinear dynamic model of the UAV is linearized and discretized to establish a practicable linear MPC prediction model. A Kalman filter-based disturbance observer is designed by augmenting internal model mismatch and external wind disturbances into system states, which realizes synchronous optimal estimation of motion states and equivalent disturbances. Furthermore, an active disturbance compensation strategy is embedded into the MPC rolling optimization, and a quadratic cost function is formulated to obtain optimal control variables for high-precision trajectory tracking and disturbance attenuation. Eight-shaped trajectory comparative simulations are implemented on the Gazebo platform. Experimental results demonstrate that the proposed scheme achieves lower overshoot and tracking deviation than the MPC method in windless environments. Under constant external disturbance, it can effectively compensate disturbance effects, eliminate steady-state errors and maintain unbiased trajectory tracking, which remarkably improves the tracking accuracy and robustness of quadrotor UAVs in complex operating conditions.