Temperature compensation for fiber optic gyroscope based on dual models

In FOG-based strapdown inertial navigation system (SINS), the temperature variation's influence on zero bias of FOG is one of the key factors that degrades the performance, which consists of two parts: a deterministic part and a random part. In this paper, a polynomial model based on correlation analysis is proposed to compensate the deterministic effect of temperature on FOG bias output, and the remaining drift of FOG is compensated based on Kalman filter (KF) model combined with time series analysis to reduce the random drift. This novel dual temperature modeling and compensation method for FOG is established respectively according to the different characteristics of the deterministic and random parts. Traditional single polynomial model is also investigated to provide a comparison with the proposed approach. Experimental results show that the bias stability of FOG output is increased to 0.01 (°)/h from 0.05 (°)/h after compensation by this novel method, while it is increased only to 0.04 (°)/h from 0.05 (°)/h by the traditional single polynomial model. The new method is shown to be more effective in compensating FOG temperature drift and improving FOG accuracy and has significant practical values in engineering.