Denoising Ring Laser Gyroscope Data Using Deep Learning Model

Ring Laser Gyroscopes are optical gyroscopes that work based on Sagnac effect. They are essential for high precision inertial navigation, due to their high accuracy, good scale factor stability and wide dynamic range. This work investigates the problem of noise in Ring Laser Gyroscope data, which significantly impacts the accuracy of inertial navigation systems. Traditional denoising methods, like Kalman filtering, Principal Component Analysis, and Empirical Mode Decomposition, were found to be inadequate in handling dynamic gyroscope data. In order to address this, a deep learning-based approach using a U-Net autoencoder was proposed and shown to be highly effective in reducing noise while preserving essential signal information. Furthermore, a hybrid approach combining U-Net with Kalman filtering was explored to enhance performance. The results demonstrate the superiority of deep learning methods in denoising gyroscope data, leading to improved accuracy and reliability in navigation systems.