A deep learned fuzzy control for inertial sensing: Micro electro mechanical systems

This study presents a new fuzzy logic controller (FLC) for micro-electro-mechanical-system gyroscopes (MEMS-Gs). The nonlinearities and unknown dynamics are modeled by the designed non-singleton type-3 fuzzy system (NT3FS) and an adaptive control scheme is presented. To improve the control accuracy, the tracking error dynamics are identified by a deep learned Boltzmann machine (RBM) and the nonlinear model predictive controller (NMPC) is designed by the optimized RBM model. Finally, the approximation errors are tackled by an adaptive compensator. The parameters of RBM are learned by contrastive divergence (CD) method and rules of NT3FS are optimized by the tuning laws that are obtained through the stability investigation. In various simulations and comparisons with some other conventional FLCs the superiority of the designed controller is shown. A good tracking accuracy of chaotic references and well robustness performance are obtained by the suggested control scheme. (C) 2021 Published by Elsevier B.V.

Automated summary

This study proposes a new fuzzy logic controller for MEMS gyroscopes, utilizing deep learning and optimization to address nonlinearities and unknown dynamics, demonstrating superior performance through multiple simulations and comparisons.