Improving functional/diagnostic testing using model-based reasoning
David Carey, P.L. Dussault · 2002
The most difficult and time-consuming portion of test program development is the diagnostic element. Because of this difficulty, diagnostics are often foregone in the production/maintenance test environment This results in a rework pile of test items that usually become scrapped because manual troubleshooting is too labor-intensive and available technician time is too scarce. A solution to this dilemma is to implement diagnostics in the production/maintenance test environment using a simplified approach: model-based diagnostic reasoning. This paper will describe how to use model-based diagnostics to uncouple the diagnostic logic from the test logic. It will describe how design and test data can be merged to form a Diagnostic Knowledge Base (DKB) The Diagnostician, a commercial off the shelf (COTS) software tool, works with the DKB as part of the test program. The Diagnostician provides the model-based reasoning at runtime. This will allow the unit under test (UUT) design changes to be easily implemented into the diagnostic test program without extensive software changes. By having a different DKB for each revision level, diagnostics can be performed by keying on the revision of the UUT with the same test program.