A Finite-time Attitude Stabilization Control Approach for Spacecraft Based on Robust Adaptive Neural Network

Abstract

In order to deal with the attitude stabilization control problem of spacecraft, a finite-time control approach based on robust adaptive neural network is developed. In detail, a modified terminal sliding surface is proposed so that singularity problem is avoided. In spacecraft dynamics, all terms except control torque are regarded as total disturbance, which is estimated by three groups of adaptive variables and approximated by Chebyshev neural network. Robust adaptive control and neural network control are connected by a switching function for better control performance and the finite-time attitude stabilization control approach is proposed based on that. Then, the finite-time stability of the controller is proved and finally, some illustrative numerical simulations are carried out to show the performance of the controller. The proposed controller has quick convergence speed and excellent control precision.