Abstract:This paper proposes a real-time denoising algorithm for fiber optic gyroscopes based on redundant second-generation wavelet transforms. The algorithm directly utilizes predictors and updaters with zero-padding interpolation to decompose and reconstruct signals. This approach ensures that the lengths of the approximation and detail signals are the same as the original signal, thus avoiding signal distortion after reconstruction. Through simulation experiments on both simulated signals and actual output signals from fiber optic gyroscopes, the results demonstrate that the algorithm effectively preserves the temporal characteristics of the signals, ensuring the integrity and accuracy of the signals are not compromised by the denoising process. Moreover, it reduce quantization noise by approximately 50.23%, decrease angular random walk by about 19.14%, and mitigate bias instability by around 53.02%. This provides an effective method for signal processing in fiber optic gyroscopes.
