Processing Technique On Fog Random Error And Its Application

Fiber optic gyro (FOG) is a new solid-state inertial instrument. With its unique advantages, FOG has been in the application of an increasing number of occasions. However, the error of FOG is an important factor on the precision of FOG output and navigation system. So establishing the model of FOG error, and then using the error compensation to process the FOG output signal, have become an important means to improve the precision of FOG output and navigation system.On the basis of studying the FOG principle and error characterization, the classification, meaning and test method of FOG major performance index are generalized. Besides, the Allan variance is utilized to analyze the FOG output, and some main noise of FOG can be extracted and confirmed.Then estimation approach which can estimate the parameter of FOG random drift is studied. Based on correlation function and ARMA model, a new effective approach which can estimate the parameter of first-order markov noise and gauss white noise is proposed. Furthermore, a prior knowledge of noise is not needed, which means the approach is suitable in most situations. The approach can offer exact parameter of inertial sensor in INS and is of great value in engineering applications.Meanwhile, the FOG random error compensation method is studied. According to the short of ARMA model, an ARIMA model which can model the nonzero output of FOG is established. Then, the random noise of FOG is processed by using Kalman filter. In addition, Kalman filter based on ARIMA model is applied to FOG signal processing in the rotation condition. The experimental results demonstrate the performance is feasible. The research extends the application range of ARIMA model. Based on this, a new improved ARIMA model filter method based on Gaussian particle filtering is put forward. This method combined FOG state estimation with ARIMA model's parametric estimation. Results indicated that, filtering accuracy of the improved ARIMA model filter method is higher than Kalman filter based on ARIMA model.For the purpose of reducing the influence caused by ambient temperature to FOG output, a two-step compensation method which can compensate both temperature noise and drift is put forward based on the static temperature test. On one hand, the temperature drift is compensated by polynomial approximation method. On the other hand, the temperature noise is compensated by Kalman filter based on ARMA model. Through this two-step compensation, the precision of FOG output is improved in the temperature changeable condition. Finally, a FIMU integrated navigation system improved scheme is put forward. The algorithm proposed in this paper was embedded into the FIMU Integrated Navigation System in the improved scheme. Then based on the micro-navigation computer, the semi-physical simulation system that is used to establish FIMU integrated navigation system is constructed. The semi-physical simulation experiments indicate that the methods proposed in this paper are of great value in engineering applications, as it can effectively improve the accuracy of FOG and FIMU integrated navigation system.