Fiber Optic Gyro North Finder

Inertial north-seeker that mainly depends on gyro is an important directional instrument in modern wars, it is the assurance of weapon system's speediness, maneuverability and precise hit. Compared with the conventional mechanical gyro, the fiber optic gyro (FOG) has many advantages, such as high sensitivity, short warm-up time, light weight, wide dynamic range, low energy consumption, low cost and so on. The main factor that affects the precision of FOG is its drift. As filtering can eliminate the drift effectively, it can improve the north-seeking precision.This paper mainly discusses two aspects of the FOG north-seeker in the view of engineering usage.One aspect is that studys on filtering of FOG's signals, the main filtering means are the Kalman Filter and the Wavelet Filter, which is discussed in chapter 3. Before discussing Kalman Filter, the method of building ARMA model is introduced, the AR model of FOG data (signals) is built using MATLAB. Then the AR model is transformed to state-space model so as to filter the drift of FOG. Before Wavelet Filter, the Wavelet Transform Theory and the Mallat fast-arithmetic are introduced. So the data of FOG straight can be co-producted with the coefficient of filter. The filter effect can be realized through decomposing and recomposing. Finally based on the math models of FOG's data built using different methods, the data of FOG are processed by Kalman Filter and Wavelet Filter respectively, and the effect of filtering is analyzed using the method of power spectral density, the result that Wavelet Transform Filter is more simply and more efficiently to filter high frequency noise than Kalman Filter is educed.The other aspect is that the validity of filtering is testified through north-seeking experiment. Multi-position north-seeking method is introduced and the errors of north-seeker are analyzed in chapter 2; In chapter 4, base on the method of linking MATLAB and VC++, the idea that software of north-seeker can be modularized and the project of experiments are introduced, The north-seeking precision is improved using the methods of deleting bad data, filtering and detrending. The detrending includes the least square algorithm and the mean-pick-point algorithm, and all algorithms are realized in MATLAB and Microsoft VC++. The complicated filtering algorithms are realized in VC++ with the math library of MATLAB through MATLAB engine. Compared an algorithm with another, precision is improved more through combining the Wavelet Transform Filter and the mean-pick-point algorithm. This method can be used in engineering.