Answers are grounded in this paper and cite their sources.
The initial alignment accuracy of microelectromechanical system (MEMS) inertial measurement unit (IMU) arrays is affected by sensor noise uncertainty and time-varying global navigation satellite system (GNSS) measurement noise. To address these issues, this paper presents a noise-adaptive initial alignment method for MEMS IMU array/GNSS integrated navigation systems. First, a second-order mutual difference method is employed to estimate the noise variance of individual MEMS IMUs online. The estimated variances are then incorporated into a weighted least squares estimator to obtain the fused angular velocity and specific force. Subsequently, a variational Bayesian adaptive left-invariant Kalman filter is developed on the SE2(3) Lie group to accommodate variations in GNSS measurement noise covariance while maintaining robustness to large initial attitude errors. Simulations and experiments using a 16-IMU array are conducted to evaluate the proposed method. Experimental results from five motion datasets show that the proposed method achieves RMSEs of 0.040°, 0.045°, and 0.303° for the final roll, pitch, and yaw errors, respectively. Compared with the conventional LIEKF using SOMD-based array fusion, the yaw error is reduced by 38.66%. These results demonstrate the potential of adaptive noise estimation to improve the initial alignment performance of MEMS IMU array/GNSS systems under uncertain noise conditions.
No takes yet. Share an insight, caveat, or question.
Sun et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: