Optical fiber gyroscope temperature drift compensation method based on wavelet denoising and neural network

The invention relates to an optical fiber gyroscope temperature drift compensation method based on wavelet denoising and a neural network. In the prior art, a temperature drift suppression effect is limited. By using the method of the invention, the above problem is overcome. The method comprises the following steps of step (1) carrying out a temperature test experiment on an optical fiber gyroscope to obtain learning sample data; step (2) smoothing original data and temperature data of optical fiber gyroscope temperature drift, which is convenient for subsequent processing; step (3) processing the original data of the optical fiber gyroscope through wavelet decomposition, reducing a random noise in the original data and retaining a drift distance caused by a temperature; step (4) taking the temperature data obtained in the step (2) and optical fiber gyroscope data filtered by the step (3) as learning samples, using a particle swarm algorithm to optimize a BP neural network model to obtain an optical fiber gy