A smart sensor-based sensing fabric surface texture identification method
By constructing a signal enhancement mechanism based on the micro-contact topology and symplectic geometric phase space of the interface, combined with an improved DeepWalk model and formal concept analysis, the problem of structural missing feature representation in texture recognition is solved, thereby improving the robustness and accuracy of texture recognition.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-08
- Publication Date
- 2026-07-03
AI Technical Summary
Existing texture recognition methods suffer from structural loss of feature representation when processing low signal-to-noise ratio signals. They neglect the complex topological structure of the contact interface of the sensing fabric and the microscopic geometric characteristics of the triboelectric signal, resulting in insufficient robustness and reliability of texture recognition.
By constructing a signal enhancement mechanism for the microscopic contact topology and symplectic geometric phase space of the interface, and using an improved DeepWalk model and formal concept analysis logic reasoning, combined with tensor ring decomposition and tangent space alignment, the interlocking degree and high-frequency component features are extracted, the high-frequency components of texture friction are reconstructed and adaptive gain control is performed, thus realizing a closed loop from microscopic physical contact modeling to deep geometric topology description.
It significantly improves the robustness and accuracy of texture recognition, realizes deep decoupling and structured characterization of triboelectric signals, and ensures the consistency of recognition results with the surface under test.
Smart Images

Figure CN122336412A_ABST