Multi-parameter fusion-based bias compensation enabling 0.0194 (°)/h BI and 0.0001 (°)/s threshold in MEMS QMG

With the rapid advancement of error correction technologies such as closed-loop quadrature control and mode matching, the performance and accuracy of Micro-Electro-Mechanical Systems (MEMS) Quadruple Mass Gyroscopes (QMGs) have been significantly enhanced. However, internal signal cross-coupling and aliasing within the gyroscope system severely restrict further breakthroughs in room-temperature bias performance, presenting a critical technical bottleneck. Initiating from the closed-loop control architecture of the MEMS QMG, this study establishes a comprehensive bias cross-coupling model across the drive, quadrature, and force-rebalance loops. The distinct influences of internal coupled signals—such as the drive-mode excitation signal, drive-mode frequency signal, and quadrature suppression voltage—on bias drift are rigorously analyzed. By jointly evaluating the mechanical resonator, control circuitry, and environmental temperature, the critical significance of the rate of change of the quadrature suppression voltage on QMG bias drift is unveiled. Consequently, an innovative bias compensation methodology is proposed by incorporating this rate of change into a multi-parameter fusion framework. Experimental results demonstrate that within a full-scale range of ±200 (°)/s, the room-temperature bias instability (BI) of the QMG is drastically reduced from 0.163 (°)/h to 0.0194 (°)/h, representing an 8.4-fold performance improvement. The angle random walk (ARW) increased from 0.0075 (°)/√h to 0.0088 (°)/√h. Under thermal cycling tests spanning from −40 °C to 50 °C, the bias fluctuation range is suppressed from 0.3 (°)/s to 0.045 (°)/s, and the BI decreases from 1.89 (°)/h to 0.683 (°)/h, yielding a 2.7-fold enhancement in thermal robustness. Additionally, the method avoids degrading the gyroscope's threshold, ensuring the MEMS QMG consistently sustains an angular rate detection resolution of 0.0001 (°)/s. These findings conclusively validate the effectiveness and robustness of the proposed method for high-reliability applications.