BLNN-based predefined-time control for spacecraft close-range proximity operations with actuator faults and unknown nonlinearities

In this paper, a novel broad learning neural network-based predefined-time control (BLNNPTC) strategy is proposed for spacecraft close-range proximity operations under actuator faults, external disturbances and unknown nonlinearities. Firstly, in order to satisfy the constraint conditions, a novel prescribed performance function is designed. Secondly, for ensuring the predefined-time convergence, a nonsingular predefined-time sliding manifold is presented. Then, to mitigate the adverse influence of unknown nonlinearities and enhance the approximation ability of broad learning neural network (BLNN), a feature nodes selection strategy is designed. Ulteriorly, a super-twisting disturbance observer is constructed to estimate the lumped disturbances. It is shown that the predefined-time convergence can be achieved and all signals are bounded. Finally, several simulation examples are employed to demonstrate the effectiveness and superiority of the proposed control scheme.