Neural longitudinal control of hypersonic vehicles with constrained aerodynamic surfaces

Abstract

This paper presents a neural adaptive flight control for longitudinal dynamics of air-breathing hypersonic vehicles (AHVs) with constrained aerodynamic surfaces. Multiple actuator constraints including magnitude, rate, and first-order dynamic model in both the elevator and canard are transformed into a specific control allocation problem, which can be readily solved using the standard model predictive control (MPC) technique. Furthermore, an adaptive control scheme is developed combining with the above control allocation and the recurrent cerebellar model articulation controller (RCMAC), which well handles actuator constraints and uncertain factors including aerodynamic coefficients, external disturbances, and flexible dynamics. Numerous simulation results verify performance and robustness of the proposed neural adaptive control.