A high-precision product assembly factor extraction method embedded with a causal mechanism and an electronic device
By using a high-precision product assembly factor extraction method embedded with causal mechanisms, the problem of causal feature extraction in phased array radar assembly was solved, achieving high-precision assembly quality prediction and process parameter optimization, and improving the model's stability and cross-condition adaptability.
Patent Information
- Application Number
- CN202610555984.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-24
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to accurately characterize the quality impact mechanism during phased array radar assembly, exhibiting spurious correlation characteristics. Furthermore, the models lack generalization ability across different operating conditions, suffer from severe issues of sample scarcity and data bias, making assembly quality prediction and process parameter optimization difficult.
A high-precision product assembly factor extraction method with causal mechanism embedding is proposed. Data augmentation is performed through diffusion model, a progressive neural network is constructed, and techniques such as interval weighting mechanism, bias loss, maximum average difference and Jeffreys divergence are used to achieve causal feature extraction and separation of non-causal features, thereby enhancing the discriminative ability of causal representation.
It significantly improves the interpretability and cross-domain generalization ability of feature extraction, enhances the accuracy and stability of assembly quality prediction, reduces the dependence on large-scale labeled data, and has good engineering promotion value.
Smart Images
Figure CN122412910A_ABST