在此研究中,設計類神經網路,來進行邏輯IC的測試與分析。先用測試機台分析邏輯IC的電路特性與電性資料,並利用已知的部份測試資料,去建立各個邏輯IC所屬的類神經網路。 利用LVQ(learning vector quantization)網路及Hopfield網路設計所需的類神經網路,它可用以模擬測試過的邏輯IC,並估測元計未經測試時的電性資料。然後,將估測的結果實際應用到邏輯IC的故障分析上。如此,可以將測試機台的使用時間降低,並且使得邏輯IC的故障分析可以更有效地執行。
In this research, designing neural network to carry on the test and analysis of logic IC. This neuro network analyses the circuit characteristic of logic IC and electricity materials with the test machine platform first, and system utilize the known part to test the materials to set up each logic IC affiliated kind of neural networks. This research utilize LVQ (learning vector quantization ) network and Hopfield network to design neural networks, and neural networks can made a mock test for logic IC.System can estimate and examine the electricity materials to Logic IC. Then, this neuro network applies estimating the result examined to the accident analysis of logic IC. It can reduce the service time of the test machine platform, and make the accident analysis of logic IC can be carried out more effectively.