Finite-time ESO-based model predictive control for hierarchical robust tracking of UAVs

Abstract

This paper investigates the problem of accurate and fast trajectory and attitude tracking for a quadrotor unmanned aerial vehicle (UAV) against unknown external disturbances. To address this issue, a finite-time extended state observer-based model predictive control framework is proposed to enable hierarchical robust tracking performance. Specifically, a finite-time extended state observer (FTESO) with a simplified finite-time convergence term is developed to estimate the lumped disturbances, reducing the structural complexity of the observer while retaining fast finite-time convergence. Next, the real-time disturbance estimates are incorporated into the prediction model to compensate for the mismatch between the nominal model and the actual quadrotor dynamics. On this basis, a Lyapunov-based model predictive control (MPC) scheme with a contraction constraint is constructed to guarantee the closed-loop stability of the position subsystem. Furthermore, a nonsingular fast terminal sliding mode control (NFTSMC) strategy is designed for the attitude subsystem to achieve finite-time convergence and improved tracking accuracy. Theoretical analysis establishes the stability of the overall closed-loop system. Finally, numerical simulations are conducted under representative disturbance scenarios. By incorporating FTESO, the proposed framework achieves a 14.43% reduction in the position-tracking root mean square error (RMSE), demonstrating the effectiveness of online disturbance compensation in improving tracking performance.