RBF neural network based second-order sliding mode guidance for Mars entry under uncertainties

There are usually larger uncertainties in the atmosphere density of Mars and the aerodynamic parameters of entry vehicles, which inevitably degrades the performance of entry guidance and control algorithms. This paper studies the robust Mars atmospheric entry guidance design based on the radial basis function neural network (RBF-NN) and second-order sliding mode control (SOSMC). First, second-order sliding mode guidance (SOSMG) law is developed to robustly follow the pre-designed nominal trajectory under larger uncertainties and effectively reduce the longitudinal error. Then, the RBF neural network is included into the second-order sliding surface guidance law to online approximate the bounded uncertain terms and further improve the guidance accuracy. Finally, the effectiveness of the guidance algorithm proposed in this paper is confirmed by Monte Carlo simulations.