Robust modification of nonlinear L1 adaptive flight control system via noise attenuation

Abstract

This paper describes the design of a nonlinear L1 adaptive control system based on angle and angular velocity feedback for the longitudinal attitude control of an unmanned aerial vehicle. An integral-form adaptive law based on Lyapunov's stability theorem is applied to estimate both the linear and nonlinear uncertainties. A modified algorithm is proposed based on the transformation between energy and autocorrelation function to solve the parameter drift problem caused by nonlinear terms in the adaptive law. The absolute value of the function is minimized by selecting an appropriate step length to eliminate the influence of noise on the adaptive law. A parameter design method for several common forms of noise is proposed. Compared with existing robust modification techniques, autocorrelation modification is demonstrated using MATLAB simulations to solve the parameter drift problem and ensure the robustness of the system. Flight tests examining the stability of the controller and the performance robustness of the nonlinear L1 adaptive control system to sensor noise and actuator failure are described. The variable gain control system is verified through flight tests. Comparative experiments show that the proposed modification suppresses the system oscillations caused by noise and solves the parameter drift problem within 5 s.