Updating the sets of target faults during test generation for multiple fault models

Abstract

A comprehensive test set targets the detection of several fault models. As an example, in this study, stuck-at faults, transition faults and four-way bridging faults are targeted. Bridging faults represent a fault model where it is necessary to select a subset of target faults from all the possible faults. After test generation for single stuck-at faults, undetectable single stuck-at faults can be used for identifying undetectable transition and four-way bridging faults. These faults can be removed from the sets of target faults to reduce the test generation effort. The new contribution of the study is related to the possibility of updating the set of target bridging faults again after test generation for transition faults. The analysis performed in the study leads to the premise that the presence of an undetectable or aborted transition fault on a line g makes bridging faults that are associated with line g less likely to be detected. As a result, line g may be covered by fewer bridging faults than selected for it, creating a test hole. To address this issue, the study suggests that more bridging faults should be selected for line g in this case. Experimental results are presented to support the discussion.