Hierarchical diagnosis of sequential circuits

This paper is concerned with automated diagnosis of sequential logic circuits, where we define the task of diagnosis to be the identification of (minimal) sets of circuit components which account for observed circuit misbehavior. We also discuss test generation for distinguishing between several possible candidate diagnoses. Our particular approach to diagnosis, an instance of consistency-based diagnosis, requires no information concerning the possible modes or probabilities of component failure such as stuck-at faults or MTBF data, respectively. The approach computes diagnoses utilizing only a set of observations of actual circuit behavior and a model of specified circuit structure and behavior. This latter model is nothing more than a logic program (i.e. a pure Prolog program or constraint logic program) satisfying certain ontological constraints