Next Generation Test Generator (NGTG) for digital circuits

Abstract

The process outlined in this paper describes the system developed to meet the goals of the Next Generation Test Generator program, funded by the Office of Naval Research. This system takes advantage of an unsupervised pattern classification algorithm (Adaptive Resonance Theory (ART)) and a Genetic Algorithm (GA) that is combined to form an optimizing control system. The GA generates a population of test patterns (individuals). Each individual is provided as a set of timed inputs to behavior based simulations representing good and faulty systems. The response of each model (good and faulty) is recombined in the form of an image matrix with each row representing a signature of each of the different circuits. FuzzyART (Fuzzy Logic Based ART) provides a method of image recognition, extracting those images that are distinctly different from any other. Each individual generated by the GA is provided as input to the list of models, then evaluated by FuzzyART and a fitness representing the number of separate classes is formed. New test sequences evolve with increasing fault isolation and detection. The process is repeated until a maximum number of models have been identified and separated. A selective breading algorithm was included to reduce the need for large populations, thus increasing the speed to converge to the "best test". The process was demonstrated using a commercial simulator based on Verilog HDL with a simple master/slave flip-flop and a moderately complex digital circuit (real UUT).