Dynamic surface control design of post-stall maneuver under unsteady aerodynamics

Abstract

This paper presents an efficient method that overcomes the problem of the control design of the post-stall maneuver under unsteady aerodynamics. On the basis of adequate data of the large amplitude oscillation experiment device in wind tunnel test, the unsteady aerodynamics model with nonlinearity, coupling and hysteresis is established by the improved Extreme Learning Machine (ELM) method. Considering the nonlinearity of the longitudinal model of the advanced fighter and the aerodynamics characteristics of the post-stall maneuver, the control law under large attack angle is designed combining the backstepping method and the daisy chain allocation method. The first order filter is adopted to prevent the “differential explosion” problem. The designed control allocation law guarantees that the conventional surfaces and the vector nozzle deflect coordinately within the position limits and the rate limits. The Radial Basis Function (RBF) network is applied to model the uncertainty, and the stability of the proposed control law which considering the uncertainty is also proved. Digital simulations of the typical “Cobra” maneuver under the unsteady aerodynamics are completed with comparisons under different conditions. Simulations results verify the validity of the proposed control law under unsteady aerodynamics and the aerodynamics uncertainty.