Generating Circuit Tests by Exploiting Designed Behavior
Abstract
Generating tests for sequential devices is one of the hardest problems in designing and manufacturing digital circuits. This task is difficult primarily because internal components are accessible only indirectly, forcing a test generator to use the surrounding components collectively as a probe for detecting faults. This in turn forces the test generator to reason about complex interactions between the behaviors of these surrounding components. Current automated solutions are becoming ineffective as designs grow larger and more complex. Yet, despite the complexity, human experts remain remarkably successful, in part, because they use knowledge from many sources and use a variety of reasoning techniques. This thesis exploits several kinds of expert knowledge about circuits and test generation not used by the current algorithms. First, many test generation problems can be solved efficiently using operation relations, a novel representation of circuit behavior that connects internal component operations with directly executable circuit operations. Operation relations can be computed efficiently for sequential circuits that provide few operations at their interfaces by searching traces of simulated circuit behavior. Second, experts write test programs rather than test vectors because programs are a more readable and compact representation for tests than vectors are. Test programs can be constructed automatically by merging test program fragments using expert-supplied goal-refinement rules and domain-independent planning techniques for AI.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1988
- Accession Number
- ADA206175
Entities
People
- Mark H. Shirley
Organizations
- Massachusetts Institute of Technology