Silicon microgyroscope temperature prediction and control system based on BP neural network and Fuzzy-PID control method

We present a novel silicon microgyroscope (SMG) temperature prediction and control system in a narrow space. As the temperature of SMG is closely related to its drive mode frequency and driving voltage, a temperature prediction model can be established based on the BP neural network. The simulation results demonstrate that the established temperature prediction model can estimate the temperature in the range of -40 to 60 degrees C with an error of less than +/- 0.05 degrees C. Then, a temperature control system based on the combination of fuzzy logic controller and the increment PID control method is proposed. The simulation results prove that the Fuzzy-PID controller has a smaller steady state error, less rise time and better robustness than the PID controller. This is validated by experimental results that show the Fuzzy-PID control method can achieve high precision in keeping the SMG temperature stable at 55 degrees C with an error of less than 0.2 degrees C. The scale factor can be stabilized at 8.7 mV/degrees/s with a temperature coefficient of 33 ppm degrees C-1. ZRO (zero rate output) instability is decreased from 1.10 degrees/s (9.5 mV) to 0.08 degrees/s (0.7 mV) when the temperature control system is implemented over an ambient temperature range of -40 to 60 degrees C.