A Resonant MEMS Electric Field Sensor Based on Feedback Capacitor Closed-Loop and Neural Network Method for Temperature Compensation

This paper proposes a resonant MEMS electric field sensor based on feedback capacitor closed-loop and neural network method to enhance temperature stability. Compared to previously reported sensors, the sensor employs closed-loop structural design and feedback capacitor detection to make its sensitivity independent of operating frequency and amplitude, greatly suppressing its susceptibility to temperature. Furthermore, temperature compensation is achieved by using a new particle swarm optimization extreme learning machine (PSO-ELM) algorithm to suppress the influence of temperature on the electric field response characteristics. The ELM is used to construct the neural network temperature model for the sensor, while the PSO algorithm is employed to optimize parameters of the model for optimal performance. Experimental results show that the maximum relative sensitivity drift and thermal sensitivity drift are reduced to 0.396% and 389 ppm/°C, respectively.