Intelligent chattering free controller design for MEMS gyroscopes usingadaptive fuzzy-neural networks based SMC

This paper presents an intelligent chatter-free controlmethod for MEMS gyroscopes. The proposed controlmethod is a combination of intelligent methods based onneural networks, fuzzy logic and nonlinear control methodwhich is a combination of sliding mode and backsteppingmethod. Considering that in the dynamic model used todescribe the dynamics of the Mems gyroscopes, parametricuncertainties and external disturbances are considered, sothe dynamics of the system is unknown in general, andtherefore first, a fuzzy-neural network is used to estimatethe dynamics of the system. Fuzzy neural networkssimultaneously use the advantages of neural networks andfuzzy logic in the form of an estimator, and the accuracyand speed of its performance is better than when we useonly neural networks or fuzzy logic to estimate systemdynamics. The central part of the control system consists ofthe sliding mode method, which is also combined with thebackstepping control method in order to take theadvantages of both methods. To mitigate the chattering andcompensate for the neural network's approximation error,an adaptive fuzzy system is used to fine-tune the slidingmode control gain. To ensure the planned control system'sasymptotic stability, the weights of the fuzzy neuralnetwork and fuzzy system parameters are changed in realtime. The simulation results show the good performance ofthe proposed control method in terms of tracking accuracy,as well as eliminating chattering and compensating thepresence of uncertainties in the dynamic model.