Non-singular Terminal sliding mode control for landing on asteroids based on RBF neural network

Abstract

A method of non-singular Terminal sliding mode control was proposed for landing asteroids with uncertainty and strong nonlinearity based on RBF neural network. The dynamics of the detector in the landing environment was analyzed, and the nominal trajectory guidance method based on optimal polynomial was designed, by which the consumption of fuel was suboptimal. Controller was designed using non-singular Terminal sliding mode. The influences caused by unknown disturbance and uncertainty during landing phase was compensated by RBF neural network real-time compensation, which could effectively suppress the influence of external disturbance and weaken the system chattering. Simulation results show that the proposed method was effective.