MEMS gyroscope robust self-adaptation control method based on neural network upper bound learning

The invention discloses an MEMS gyroscope robust self-adaptation control method based on neural network upper bound learning. The method includes the following steps that an ideal kinetic model and an MEMS gyroscope kinetic model are established, a sliding mode function is designed, a control law is obtained based on the sliding mode function, and an RBF neural network upper bound estimated value is used as a gain of a robust item on the basis of the control law together with a feedback item and the robust item; a parameter self-adaptation law and a network weight self-adaptation law are designed based on a Lyapunov method. According to the MEMS gyroscope robust self-adaptation control method based on neural network upper bound learning, the feedback item is added in the control law, the two-shaft vibration trajectory tracking speed and the parameter estimation speed of an MEMS gyroscope are greatly increased, and the vibration amplitude is decreased; the robust item based on RBF neural network upper bound learning is added in the control law, the buffeting problem caused by large external disturbance and fluctuation and the problem that the dynamic characteristics are changed worse are solved, the uncertainty of a structural formula and the uncertainty of a non-structured formula are eliminated, and therefore the robustness of the system is further improved.