Random drift modeling and compensation for fiber optic gyroscope under vibration

Discerning fiber optic gyro´s (FOG) drift and predicting the variation tendency of FOG´s drift under vibration environment are very important to improve the accuracy of the whole inertia navigation system (INS). In this paper, the vibrant signal of FOG is divided into three stages and analyzed by frequency spectrum analysis respectively, where, the three stages are the FOG signal before vibration, the FOG signal under vibration and the signal after vibration. Then the signal under vibration (stage 2) is filtered by low pass filter. At last the Wavelet Neural Network (WNN) is applied for modeling and compensating the random drift of FOG under vibration. To analyze the denoising and compensation effect, Allan variance method is introduced. Finally, the comparison of simulation results shows that the proposed method can effectively eliminate the noise and compensate the Fog´s random drift under vibration.