Research On Modeling,Analysis And Compensation Of Fiber Optic Gyroscope Random Drift

Fiber optic gyroscope(FOG) is a kind of angular rate sensor based on Sagnac effect with solid state, small size, light weight, wide dynamic range, fast starting, low cost, high reliability and so on. Many countries pay much attention to FOG for the military and civilian purposes. FOG is widely used in the world because of its outstanding features. However, FOG is easily affected by the performance of its components, parasitic effects and the surroundings. This leads to the FOG output drift and deteriorates the precision of the inertial navigation system. Therefore, it is significantly important to suppress the FOG drift error to improve the FOG measurement accuracy.This thesis designs and implements a FOG testing platform which provides support for the FOG experiments. The corresponding FOG testing platform software is developed based on Labview, which is helpful to improve the level of automation and the efficiency of experiments. This thesis is based on the digital closed-loop interferometric FOG, illustrating its system structure, working principle and performance indexes.Allan variance method is utilized to evaluate the FOG random drift characteristic. Time series analysis, Kalman filter, and FLP filter are utilized to model and filter the FOG random drift sequence. This thesis firtly preprocesses the FOG random drift sequence and secondly builds an ARMA model. A state space model is induced through the ARMA model, and then a Kalman filter is built to filter the sequence. This thesis uses Allan variance method to identify and quantify five main error terms in the FOG random drift. Bias stability and five error term coefficients are used to assess the filtering effect. A better step size selection rule is proposed for the iterative procedure of the FLP filter. The experiments show that the FLP filter based on the proposed method gets a good filtering effect in the static and dynamic tests of the FOG.An improved AR model and its modeling method are proposed in order to model, predict and compensate the FOG temperature drift sequence. Having the model at hand, a FOG temperature characteristic evaluation software is developed. This thesis analyzes Shupe effect of the FOG, and collects the temperature drift data to build the improved AR model, and validates the effectiveness of the model through the temperature experiments. The FOG output can be compensated by the predictive value of the model. A FOG temperature characteristic evaluation software is developed with Labview, marking it valuable for engineering application in rapid assessment and prediction of FOG temperature characteristic and improvement of FOG measurement accuracy.