Scheduled Composite Off-Line Output Feedback Model Predictive Control for a Constrained Hypersonic Vehicle Using Polyhedral Invariant Sets

Abstract

This paper presents a scheduled composite off-line output feedback model predictive control strategy for a constrained hypersonic vehicle (HV) in the presence of external persistent disturbances. First, multiple linear time invariant (LTI) models are constructed to represent the nominal longitudinal dynamics of HV without external persistent disturbances. Then, by combining the construction of asymptotically stable polyhedral invariant sets for explicitly handling asymmetric constraints of control inputs and angle of attack and the state estimator for estimating partial unmeasured states, a set of local off-line output feedback model predictive control schemes are first developed for multiple LTI models. Additionally, based on the strong nonlinear approximation ability of a recurrent cerebellar model articulation controller (RCMAC), a RCMAC disturbance observer (RCMACDO) is presented to estimate the actual disturbances, and then an auxiliary compensation controller is appended to attenuate the influences of external persistent disturbances. Furthermore, by designing a proper scheduling strategy, the proposed control strategy with the overlapped stable regions is proposed for the wide tracking task. Finally, the comparative simulation results for tracking reference commands of velocity and altitude verify the effectiveness of the proposed control strategy.