A Novelty Temperature Compensation Model for Dual-Mass Vibration MEMS Gyroscope Based on Machine Learning and...

Area of Science:

• * MEMS (Micro-Electro-Mechanical Systems) technology
• * Sensor signal processing and compensation
• * Navigation systems and inertial measurement units

Background:

• * MEMS gyroscope outputs are susceptible to temperature variations, leading to drift errors and reduced navigation accuracy.
• * Accurate temperature compensation is crucial for enhancing the performance of MEMS gyroscopes in precision navigation applications.

Purpose of the Study:

• * To propose and validate a novel temperature compensation model for dual-mass vibration MEMS gyroscopes (DMVMG).
• * To significantly improve the navigation precision of DMVMGs by mitigating temperature-induced errors.

Main Methods:

• * Development of a hybrid compensation model integrating TTAO-VMD, 1D-CNN-Bi-GRU-Attention, and SHAKF algorithms.
• * Signal preprocessing using TTAO-VMD for noise decomposition and filtering.
• * Temperature error modeling using 1D-CNN-Bi-GRU-Attention with temperature, rate of change, time, and noise as inputs.
• * Reconstruction of optimized signal components for temperature-compensated output.

Main Results:

• * Significant reduction in angle random walk (N) from 18.56 °/h to 0.17 °/h.
• * Substantial decrease in bias instability (B) from 32.76 °/h to 0.82 °/h.
• * Experimental validation using the Allan variance method confirmed the model's effectiveness.

Conclusions:

• * The proposed integrated TTAO-VMD, 1D-CNN-Bi-GRU-Attention, and SHAKF model effectively compensates for temperature drift in DMVMGs.
• * The enhanced compensation significantly improves gyroscope performance, meeting critical requirements for high-precision navigation.
• * This research offers a robust solution for overcoming temperature-related challenges in MEMS gyroscope applications.