Reinforcement-learning-based fixed-time attitude consensus control for multiple spacecraft systems with model uncertainties

Abstract

In this study, the problem of fixed-time attitude consensus control was investigated for multiple spacecraft systems with model uncertainties. First, a distributed fixed-time adaptive observer is proposed for estimating the states of the leader. Subsequently, on the basis of the observation errors, transformed error dynamics are described and they are used to combine the unknown nonlinear terms of a spacecraft system. By using the non-singular fast terminal sliding mode technique and a reinforcement learning optimization algorithm, we implemented a neural-network-based fixed-time control strategy to achieve optimal attitude consensus control. The stability of the system and the fixed-time convergence of the tracking error are demonstrated by using the Lyapunov theory. Furthermore, the effectiveness and superiority of the control strategy are shown through numerical simulations.