Research on detection and data fusion algorithm on array gyroscope

Research indicates that outliers and noise significantly constrain the practical application accuracy of gyroscopes. To enhance gyroscope application accuracy, outlier identification, elimination, and correction are conducted by establishing a gyroscope output error model and computing interval estimates for current observations. Secondly, fusion weights are dynamically updated by calculating gyroscope noise coefficients to reduce the impact of noise on the fusion accuracy of MEMS gyroscope arrays. Experimental results demonstrate that the improved adaptive optimal weight fusion method effectively enhances the output accuracy of gyroscope arrays and mitigates gyroscope output noise to some extent. The results of Allan variance analysis show that the random error of the proposed method is reduced to 65.1% and the zero bias instability is reduced to 66.9%, which effectively improves the performance of MEMS array gyro in practice.