Chebyshev Neural Noncertainty-Equivalent Adaptive Control of Wing Rock of Slender Delta Wings

Abstract

[Abstract] Aircraft with highly swept wings, flying at high angles of attack, exhibit selfexcited rolling motion (termed wing rock). This paper presents a noncertainty-equivalent adaptive control system for the wing rock motion control via a Chebyshev neural network. The unmodelled nonlinearity in the roll dynamics is approximated by a neural network with Chebyshev polynomials in roll angle and roll rate of the first kind. An adaptive control law is derived for the trajectory control of the roll angle. The control system includes a control module and a parameter identifier, and uses filtered signals for synthesis. Unlike the certainty-equivalent control laws, each estimated parameter is the sum of a judiciously chosen nonlinear function and a partial estimate generated by an update rule. The nonlinear function in the parameter estimate provides stronger stability property in the closed-loop system. Simulation results are presented which show that the adaptive Chebyshev neural controller is capable of suppressing the wing rock motion of the model with unknown nonlinearity and disturbance input at different angles of attack.