Adaptive Test Vector Generation Under Zero-Defect Constraint

Le résumé fourni par la source

In order to meet the testing needs of high reliability and full area coverage in military electronic assembly, this paper proposes an adaptive test vector generation scheme for zerodefect constraint conditions. This method uses multi-source process parameters to build a defect probability response model, relies on a vector scoring mechanism to generate priority ranking, and achieves online real-time adjustment with a feedback optimization module. The system integrates reflow soldering perception data with model-driven strategies and establishes a closed-loop mechanism of test generation-execution-feedback. Experiments conducted with the SolDef_AI dataset show that the proposed method is superior to traditional strategies in terms of average coverage, redundancy control, and defect recognition accuracy. The number of vectors shows a significant compression trend, and the delay caused by feedback is significantly reduced. Experiments verify that the method has good deployability and performance stability in actual manufacturing environments. The research results provide key

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.