A deep learning-based fiber-optic gyroscope vibration data detection method
By constructing a deep learning network model with an adaptive time query module, advanced residual blocks, and a decomposition-fusion prediction architecture, the problems of low detection accuracy and poor robustness in fiber optic gyroscope vibration data detection are solved, achieving higher detection accuracy and generalization ability.
CN122389622APending Publication Date: 2026-07-14BEIHANG UNIV
View PDF 0 Cites 0 Cited by
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2026-05-06
- Publication Date
- 2026-07-14
Smart Images
Figure CN122389622A_ABST
Abstract
The present application relates to a kind of deep learning-based fiber-optic gyroscope vibration data detection method, belong to deep learning algorithm technical field, the present application is enhanced data characteristics by dimension expansion processing, constructs the network model including adaptive time query module, senior residual block and decomposition-fusion prediction architecture, realizes the reconstruction and abnormal detection of vibration sequence, improves the robustness and precision of detection. Enhanced data characteristics by dimension expansion, adaptive time query dynamically captures periodicity, senior residual block improves extraction efficiency, decomposition-fusion architecture processes high-dimensional sequence, the higher detection accuracy and robustness of this method is realized on a variety of model fiber-optic gyroscope vibration test data.
Need to check novelty before this filing date? Find Prior Art