Control of Quadrotor Autonomous Aerial Vehicle via Finite-Time Observer

Abstract

This paper investigates the attitude tracking control problem for quadrotor autonomous aerial vehicles (AAVs) subject to unknown time-varying attitude measurement errors. The measurement errors, acting as a mismatched disturbance, cannot be directly attenuated by conventional control methods. To address this issue, a finite-time observer-based control method is proposed. Firstly, an attitude dynamics model of the quadrotor incorporating attitude measurement errors is established. Secondly, a finite-time observer is designed using the super-twisting sliding mode technique to accurately estimate the unknown time-varying measurement errors online. Subsequently, the bias estimates are incorporated into the design of an integral terminal sliding mode surface, based on which a robust finite-time attitude tracking controller is developed. Rigorous Lyapunov-based analysis proves that the proposed observer and controller guarantee finite-time convergence (rather than asymptotic or fixed-time convergence) of the estimation errors and the attitude tracking errors. The convergence time bounds are explicitly derived, which depend on the initial estimation/tracking errors and the observer/controller gains. By appropriately increasing the observer and controller gains, the convergence time can be effectively shortened. Finally, HIL-in-the-loop semi-physical simulation platform experimental results are presented to demonstrate the superiority of the proposed method over traditional SMC control and integral terminal sliding mode control (ITSMC) in terms of convergence speed and tracking accuracy.