ESSENTIAL: an efficient self-learning test pattern generation algorithm for sequential circuits

The authors present ESSENTIAL, deterministic automatic test pattern generation algorithm for sequential circuits. By combining reverse time processing over time frames and forward processing within time frames, ESSENTIAL avoids the detrimental a priori determination of a topological path to be sensitized or of a primary output, to which the fault effects have to be propagated. Moreover, the proposed test generation approach fully exploits the beneficial techniques that have successfully been used for combinational circuits by the automatic test pattern generation system SOCRATES. In particular, the authors discuss a learning procedure for global implications not only over reconvergent fanout, but also over time frames as well as static and dynamic unique sensitization techniques. After introducing a couple of intelligent heuristics employed for guiding and supporting the decision-making process, the authors report some preliminary but encouraging experimental results