Adaptive control of hypersonic vehicles using intelligent allocation

Abstract

This paper proposes an intelligent allocation-based adaptive controller for the longitudinal motion of air-breathing hypersonic vehicles (AHVs). A control-allocation (CA) module is developed to deal with composite actuator servo constraints that have usually been neglected in existing AHV control works. This CA process is realized with the help of the recent deep deterministic policy gradient algorithm without involving any online optimization. In the followed adaptive controller design, several auxiliary signals are constructed to compensate for the possible CA error, while adaptive super-twist differentiators are employed to fast estimate the lumped effect of uncertain aerodynamic coefficients and unknown external disturbances. As a result, the adaptive control algorithm only needs three parameter updating laws, whose dimension is much lower than traditional adaptive control strategies for AHVs. Simulations are provided to verify the proposed intelligent allocation-based adaptive controller.