Nonlinear Analysis And Adaptive Control Of MEMS Resonator Coupling Network

Micro-Electro-Mechanical System（MEMS）resonator is widely used in the fields of5G communication,unmanned driving and aerospace due to its high accuracy,low energy consumption,miniaturization,and many other advantages.However,the MEMS resonator will produce complex and rich nonlinear dynamic behaviors due to its microstructure characteristics and the impact of internal electromagnetic field,especially the high frequency chaotic oscillation will seriously destroy the stability of the system and even lead to the system paralysis.With the rapid development of wireless communication,artificial intelligence and other technologies,a single MEMS resonator cannot break through the limitations of band-pass,output energy and fault-tolerant ability.Meanwhile,the existing MEMS resonator system has problems such as inaccurate mathematical modeling,fuzzy dynamic cognition and lack of core control algorithm,making it gradually fail to meet the high-performance requirements of current intelligent sensing.Therefore,this paper takes MEMS resonator coupling network system as the research object,establishes its dynamic model,reveals its nonlinear characteristics,and explores the adaptive control theory of MEMS resonator coupling network by using multidisciplinary theories and methods such as dynamics,microelectronics,analog integrated circuit technology and neural network adaptive control.It eliminates the adverse effects of high frequency chaotic oscillation and unknown uncertainty,improves the system accuracy and response speed,and then provides theoretical and practical basis for the research and development of advanced and sophisticated equipment such as automatic driving,satellite navigation and intelligent terminal equipped with MEMS resonator coupling network.The specific research content can be divided into the following three aspects:（1）Dynamic analysis and adaptive control of a single MEMS resonator.Firstly,the nonlinear characteristics of a single MEMS resonator system are revealed by phase diagram,time history diagram and Lyapunov exponent diagram under different driving voltages.Then,based on the energy flow theory and electronic information technology,an analog integrated circuit of equivalent MEMS resonator system is constructed on Multisim,which further reveals and verifies the inherent chaotic motion of the system.Finally,based on radial basis function neural network（RBFNN）,an adaptive inversion control scheme is designed to suppress chaotic oscillation.Simulation results show that the proposed control scheme is feasible.The results of this study provide technical support for the modeling of MEMS resonator coupling network,the establishment of integrated circuit experiment platform and adaptive inversion control.（2）Modeling of MEMS resonator coupling network and establishment of integrated circuit experiment platform.Firstly,in order to improve the sensitivity,output energy and stopband suppression performance of the MEMS resonator system,a coupling network model consisting of four MEMS resonators is established considering the mechanical and electrostatic coupling effects.Then,the phase diagram and Lyapunov exponent diagram are constructed to reveal the dynamic evolution law of MEMS resonator coupling network.Secondly,the integrated circuit experiment platform of MEMS resonator coupling network is built by using proportional operation circuit,addition and subtraction circuit and integral operation circuit.Finally,the circuit experimental results indicate that the coupled network system exhibits chaotic oscillation at A=1.71V,f=287.94Hz.（3）Adaptive inverse control method of MEMS resonator coupling network based on IT2FNN.In order to effectively solve the chaotic oscillation of the MEMS resonator coupling network and realize its steady-state and high-performance output,an adaptive inversion control scheme is proposed on the basis of research content（2）.In the controller design,an interval type II fuzzy neural network（IT2FNN）is used to estimate the unknown nonlinear function of the system,and the cosine barrier function is constructed to ensure the boundary constraint of the output state.Lyapunov stability analysis proves that all signals of the closed-loop system are bounded.A large number of simulation results verify the correctness and effectiveness of the proposed control scheme,and the system tracking errors are converged within[-3×10^-3,1×10^-3].