Autonomous estimation of Allan variance coefficients of onboard fiber optic gyro

Abstract

Satellite motion included in gyro output disturbs the estimation of Allan variance coefficients of fiber optic gyro on board. Moreover, as a standard method for noise analysis of fiber optic gyro, Allan variance has too large offline computational effort and data storages to be applied to online estimation. In addition, with the development of deep space exploration, it is urged that satellite requires more autonomy including autonomous fault diagnosis and reconfiguration. To overcome the barriers and meet satellite autonomy, we present a new autonomous method for estimation of Allan variance coefficients including rate ramp, rate random walk, bias instability, angular random walk and quantization noise coefficients. In the method, we calculate differences between angle increments of star sensor and gyro to remove satellite motion from gyro output, and propose a state-space model using nonlinear adaptive filter technique for quantities previously measured from offline data techniques such as the Allan variance method. Simulations show the method correctly estimates Allan variance coefficients, R = 2.7965exp − 4and tracks degradation of gyro performance from initail values, R = 0.651 • /h 2 , K = 0.801 • /h 1.5 , B = 0.385 • /h, N = 0.0874 • /h 0.5 and Q = 8.085exp − 5 • , to final estimations, R = 9.548 • /h 2 , K = 9.524 • /h 1.5 , B = 2.234 • /h, N = 0.5594 • /h 0.5 and Q = 5.113exp−4 • , due to gamma radiation in space. The technique proposed here effectively isolates satellite motion, and requires no data storage and any supports from the ground.

In-Orbit Performance Evaluation of a Spaceborne High Precision Fiber Optic Gyroscope

“…There are the intrinsic random noise and bias drift involved in FOG output data [ 11 , 12 , 13 , 14 , 15 ]. When the FOG fails or the performance degrades, the random noise will change significantly.…”

AMA- and RWE- Based Adaptive Kalman Filter for Denoising Fiber Optic Gyroscope Drift Signal

“…Deterministic errors are bias, scale factor, and misalignment, which are relatively easier to be compensated by suitable calibration methods in laboratory environment. Stochastic errors, which are induced from the environmental temperature changes, electronic noises, and other electronic equipment interfaced with it [ 4 , 5 ], are difficult to be directly eliminated. As an alternative, stochastic models and denoising methods are two main techniques to restrain the FOG random errors reported in the literature.…”

A novel IFOG Allan variance coefficients real-time detection method with temperature variation in moorage

“…Additionally, some other methods are presented [8,9] to estimate the coefficients of in-flight gyros, but they require support from the ground and after the data is collected. However, the references [10,11] proposed methods both in real-time and do not require large quantity of stored IFOG data. The accuracy and applicable conditions of auxiliary star sensor are skeptical especially for the marine IFOG gyrocompass, and it could not be used in all-weather conditions [12].…”