A reconfigurable online BIST for combinational hardware using digital neural networks
S. Behdad Hosseini, Ali Shahabi, Hassan Sohofi, Zainalabedin Navabi · 2010
Online testing, one of the most challenging issues in design for test domain, is intended for inspection of digital systems behavior during their working period. This paper presents a novel approach for simultaneous online testing of several combinational circuits using a reconfigurable neural network implemented along the original hardware. Automatic generation of the neural network to model the behavior of each design is proposed as well as required techniques to obtain optimum configuration for its hardware realization. Advantages and shortcomings of this approach in terms of area overhead, fault latency and reliability are discussed as well.