Neural Network Error Compensation Method for Analytical Redundancy of Angular Velocity

From the perspective of avoiding common mode problems, this article proposes the additional measurement component of Micro Electro Mechanical System(MEMS) gyroscope as backup input sources to achieve input resolution redundancy in inertial reference system(IRS). On this premise, a neural network-based multi gyroscope error compensation algorithm is proposed, which improves the single output accuracy of gyroscopes through neural network algorithms and uses confidence algorithm for data fusion processing, making the algorithm more practical in engineering.