Compact Dictionaries for Reducing Compute Time in Adaptive Diagnosis

Field testing and field diagnosis are effective ways for achieving high reliability of modern systems. Since they are executed during an idle mode or a start-up mode in a system, they must be completed within very short time. Adaptive diagnosis applies test patterns selectively according to a candidate faults set that is obtained during the fault diagnosis process. In this paper, we propose an adaptive fault diagnosis method using a compact dictionary in order to reduce compute time for deducing candidate faults. A compact dictionary is created by compacting some output values into one bit. Although the compute time is reduced using a compact dictionary, the number of applied test patterns for diagnosis may increase in some cases. We investigate the relation between the size of a compact dictionary, compute time and the number of test patterns by experiments for benchmark circuits.