An Application of Artificial Neural Networks in Combinational Circuit Test

摘要

The so-called Artificial Neural Network means to regard the structure and function of the direct emulation person's brain as principle and make use of a great deal of processing parts to build up network system with the artificial methods. It can carry out the non-linear mapping of problem object by any precision without the confining of the network scale and trains time. Studying test pattern generation of combination electric circuit based on Artificial Neural Networks has some special meaning. Every nerve cell of the network has brief structure and function and we can do group operation through the parallel and cooperating work among a large number of nerve cells to reduce the difficulty of combinational circuit test and test time. The article expatiates how to build a neural network model of combination circuit, and tells the method of generating primal test vector: firstly describe the Hopfield neural network of combination electric circuit, then find the relation between the minimum value of Hopfield neural network energy function and the test sequence of combination circuit and generate the primal test vector.