Neural-networks and event-based fault-tolerant control for spacecraft attitude stabilization

This paper proposes a neural network and event-based fault-tolerant control scheme for spacecraft attitude stabilization in the presence of lumped disturbances, which consists of space disturbances, inertia uncertainties, and actuator faults. A neuro-adaptive estimator is employed to approximate the lumped disturbances, with the help of its powerful adaptive estimation capability of approximating any unknown smooth nonlinear function with arbitrary accuracy. The estimation is then utilized to formulate an integrated event-based dual-channel control scheme that can both guarantee the system's convergence and ensure the event triggering sequence possessing no-Zeno behavior simultaneously. The proposed control scheme provides a new and straightforward way for spacecraft attitude control to deal with lumped disturbances while requiring a low actuator updating frequency, thus saves on-board communication resources. Numerical simulations show the effectiveness of the algorithms.