FOG Combination Of Calibration And Error Compensation Technology Research

The fiber-optic gyro inertial equipment is a newer inertial navigation system, which wasslowly developed, researched and continued to be matured in the last ten years. Such inertialnavigation system, relativing to the traditional navigation system, has many advantages, suchas the less cost, better linearity, no moving parts, and ease of mass production and so on. Dueto the rapid development of the world military equipment modernization and the broadapplication prospects needs of the future civilian market, the navigation system has got arapid development.In the technology of FOG combination, equipment made in certain process conditions,before you really use it, need to accomplish the calibration experiments to obtain the scalefactor, installation errors, and a series of datas and so on. If wanting to improve the accuracyof navigation, it will take the great cost to improve process of production equipment, and onthe contrary, improveing on through the certain calibration method as well as the errorcompensation method will reduce costs. In this paper, for the laboratory purchased fiberInertial Measurement Unit, it is firstly established that the five gyroscope mathematical modeland seven accelerometer error mathematical models, and based in three-axis turntablefocusing gyro error model, the rate experimental is designed; and for the acceleration model,position experiment is put forward; and for the position errors which there are in the locationtest, the zero experiments is designed; The data is collected by serial and calculated, and thenuse the least squares method to calculate the error coefficients of each model.Deterministic error terms are given above, while there are some uncertain error terms,generally mainly referring to the gyro drift and the zero bias instability. INS Gyro outputdata,by using Allan variance method, is analysised more systematicly, obtaining the fiverandom error terms of three gyroscopes and three accelerometers: the bias instability、anglerandom walk、quantization error、rate rampand the rate random walk.In this paper, for the existence of a large number of random errors, wavelet analysismethod is applied to filterring. In the process, the first is to do a detailed analysis to determinethe wavelet basis, then the samely, by comparing it is to select the method for determining thethreshold size, the threshold way and the decomposition scale. At the same time, the signal iscarried by filtering process on the basis; finally, for the lack of the ordinary wavelet filtering,the stationary wavelet method is used in the filtering process, and gets the better results thanthe ordinary wavelet. In addition to using wavelet filtering, the adaptive filtering method is put forward toprocess the gyro output signal. The first is the use of the standard LMS method and get betterresults, but the convergence speed of the method is slow, and then the normalized LMSmethod is applied to the signal filtering. At the last, for the contradictory both filtering smallerconvergence time and stable precisions, the paper is to propose the improved the LMS method,the variable step size LMS and prove the effectiveness of this improved method, by filteringthe data and Allan variance analysis.