Fast estimation method of fiber optic gyro stochastic error coefficients based on noise real-time decoupling

Abstract Allan variance (AVAR) is an important means to identify stochastic error parameters of fiber optic gyro, but this method has obvious defects. The traditional method based on AVAR to solve the power spectrum coefficient of stochastic error has the defects of large data demand, poor real-time operation and poor confidence of long-correlation noise variance estimation. Therefore, in view of the above deficiencies, the following studies have been conducted. Firstly, the differential equations corresponding to the error terms are established based on the power spectrum characteristics of the fiber optic gyro output data. Then, an improved angle random walk noise variance estimation method and an improved adaptive filter are proposed to decouple the bias instability and the rate random walk of fiber optic gyro with long correlation time. Finally, the real-time estimation of the three main stochastic error coefficients of fiber optic gyro is realized according to the estimation results of the noise information. The proposed algorithm can not only improve the identification efficiency of fiber optic gyro error parameters, but also improve the estimation accuracy of noise power spectrum coefficient. Simulation and experimental results show that the identification efficiency of fiber optic gyro error is improved by 90% and the estimation accuracy of error coefficients is improved by 68%.