An Improved Least Mean Square Demodulator Based on Adaptive Moment Estimation Algorithm for LFM MEMS Gyroscope

Lissajous frequency modulated (LFM) gyroscope based on Micro-Electro-Mechanical System (MEMS) craft have been mainly used for military and civil applications. The performance of LFM MEMS gyroscope is restricted by the phase noise of the demodulation algorithm and frequency readout circuits. However, the accuracy of conventional demodulators such as multiplication and original least mean square demodulation (LMSD) which are commonly used, is limited by the noise suppression performance. In order to break through the bottleneck, a novel improved LMSD based on adaptive moment estimation (Adam) algorithm for LFM MEMS gyroscope has been proposed in this paper. The proposed demodulator is appropriate for the demodulation of weak signal from noise, and can achieve the lower noise level and high precision output of sense mode of LFM MEMS gyroscope. The simulation verification of three different demodulation algorithms is carried out, and the results show that the Adam-LMSD algorithm has the lowest noise level, and the frequency demodulation accuracy can reach 1.7 × 10
⁻⁷
Hz.