Analysis And Processing Of Laser Gyro Random Error

As an independent navigation system, Inertial Navigation System(INS) has a lot of characters,such as high precision in short time, continuous output, good ability about anti-interferenceand so on, is characterized by its fully independent navigation systems in any other irreplaceable.With the the development of new solid-state gyro, such as laser gyro, Strapdown Inertial System(SINS) gradually replace the Gimbaled Inertial System in low-precision applications,become a new direction of inertial technology.Reserch on the error characteristics of laser gyro and compensate gyro output with the reliable error model, which could effectively improve the system accuracy of SINS. This paper deeply studied on the random error of laser gyro, the main contents and conclusions were as follows:The common analysis methods of laser gyro error were discussed. The principles and characteristics of Fourier transform, Power Spectral Density(PSD) estimation and the Allan variance were discussed. After in-depth researched on the Allan variance about random drift of laser gyro, the conversion formula about its PSD and Allan variance was derived, and the common log-log s (t )- t graph was given.Making static drift test with a certain type of laser gyro, fitting the test data with estimated value of Allan variance, the random noise of experimental gyro was obtained. The evaluation method about the estimation accuracy of Allan variance in the practical work was given.The analytic relation about the the time length of test samples between fitting accuracy was discussed.The typical time series model and the gyro random drift model were introduced. Focusing on the Box-Jenkins modeling method and Pandit-Wu modeling method to study the time series modeling methods, combined with BIC criterion, derived the improvement of the DDS modeling method.According to a certain type of laser gyro static drift test data, the reliable time series model about random drift of the laser gyro-ARMA(2,1) model was given by using an improved method of Pandit-Wu model combined with Allan variance analysis. Generality of the model structure is analyzed,and the conclusion that: the general reliable time series model of the experimental laser gyro drift can be established through a lot of static drift test.Based on the time series model , the Kalman filter algorithm of laser gyro random drift was established, the fading filter was introduced. Real data examples demonstrate: filtering algorithm can effectively suppress the various laser gyro random noise, the result of fading filter slightly better than conventional filter in the same conditions; after filtering,the quantization noise reduced about 60%, the angle random walk reduced about 30%, the bias instability,the rate random walk and the drift rate ramp were reduced about 70%.The influence of laser gyro random drift on the navigation precision of SINS was analyzed, which through comparing the position error static experiment and dynamic experiment before and after filtering. Experimental results show that after filtering the static navigation position error significantly reduced, the 900s position error reduced about 50%; As the complex changes in the external environment, the900s position error of sports experiment reduce was reduced about 20%.