基于物理-数据双驱动框架的气体识别方法和系统
By employing a physics-data dual-driven framework for gas identification, combining sensor arrays and deep learning networks, the challenge of identifying structurally similar volatile organic compounds has been solved. This approach achieves high-precision and interpretable gas classification and possesses the ability to generalize to unknown homologues.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- EAST CHINA NORMAL UNIV
- Filing Date
- 2026-03-19
- Publication Date
- 2026-07-17
AI Technical Summary
Existing gas identification technologies struggle to effectively distinguish between structurally similar volatile organic compounds, and purely data-driven models lack physical interpretability and generalization ability.
A gas identification method based on a physics-data dual-driven framework is adopted. It acquires macroscopic dynamic sensing signals and microscopic physical descriptors through a sensor array, and combines one-dimensional convolutional neural networks and multilayer perceptron networks to perform cross-modal feature fusion, thereby achieving accurate gas identification.
It significantly improves the classification accuracy of structurally similar volatile organic compounds, enhances the robustness and interpretability of the model, has the ability to generalize the identification of unknown homologues, and verifies the reliability of the decision logic through attribution analysis.
Smart Images
Figure CN121877972B_ABST