Model-based diagnostic system with automated procedures for next test selection

• US 6,167,352 A
• Filed: 06/26/1997
• Issued: 12/26/2000
• Est. Priority Date: 06/26/1997
• Status: Expired due to Fees

• Current Assignee: Agilent Technologies, Inc.
• Original Assignee: Agilent Technologies, Inc.
• Inventors: Barford, Lee A., Kanevsky, Valery
• Primary Examiner(s): Peeso, Thomas R.
• Application Number:
  US08/882,791
• Time in Patent Office: 1,279 Days
• Field of Search: 702/81, 395/183.09
• US Class Current: 702/81
• CPC Class Codes: G01R 31/318371 Methodologies therefor, e.g. algorithms, procedures

Abstract

A model-based diagnostic system that enables selection of components for replacement and selection of the next test to apply to a device under test based on economic costs associated with replacing components and conducting tests.

Claims

1. A diagnostic system that enables selection of one or more components in the device under test for replacement and a selection of one or more next tests from among a set of diagnostic tests not yet applied to a device under test in response to a test result for each of a subset of the diagnostic tests that have been applied to the device under test wherein each diagnostic test covers a portion of one or more of the components, and in response to a set of patterns of test results and good/bad states of the components and a frequency of occurrence indication for each pattern.

2. The diagnostic system of claim 1, wherein the next tests are provided to a user in a table that includes an identifier of each next test and a goodness indication for each next test.

3. The diagnostic system of claim 2, wherein the table is sorted in response to the goodness indications.

4. The diagnostic system of claim 2, wherein the goodness indication is a numerical value that indicates the relative goodness among the next tests.

6. The diagnostic system of claim 2, wherein each goodness indication includes an economic cost associated with performing the corresponding next test.

7. The diagnostic system of claim 1, wherein data indicating the next tests are transmitted to an element that operates in response thereto, the data including an identifier of each next test and a goodness indication for each next test.

14. The diagnostic system of claim 1, wherein the patterns of test results and good/bad states of the components and the frequency of occurrence indication for each pattern represent actual test results and good/bad states obtained from historical data.

5. A diagnostic system that enables a selection of one or more next tests from among a set of diagnostic tests not yet applied to a device under test in response to a test result for each of a subset of the diagnostic tests that have been applied to the device under test wherein each diagnostic test covers a portion of one or more of the components, and in response to a set of patterns of test results and good/bad states of the components and a frequency of occurrence indication for each pattern wherein the next tests are provided to a user in a table that includes an identifier of each next test and a goodness indication for each next test and wherein the goodness indication is a graphical display element that indicates the relative goodness among the next tests.

8. A diagnostic system that enables a selection of one or more next tests from among a set of diagnostic tests not yet applied to a device under test in response to a test result for each of a subset of the diagnostic tests that have been applied to the device under test wherein each diagnostic test covers a portion of one or more of the components, and in response to a set of patterns of test results and good/bad states of the components and a frequency of occurrence indication for each pattern wherein for each diagnostic test not yet applied the selection of the next test includes the steps of:

determining a first probability that the diagnostic test not yet applied will fail given the test results;

determining a second probability that the diagnostic test not yet applied will pass given the test results;

determining a first proportion of diagnoses that will be correct if the diagnostic test not yet applied is applied next and fails;

determining a second proportion of diagnoses that will be correct if the diagnostic test not yet applied is applied next and passes;

determining an expected proportion of correct diagnoses if the diagnostic test not yet applied is applied next given the first and second probabilities and the first and second proportions.

9. The diagnostic system of claim 8, wherein the selection of the next test further includes the step of selecting the next test from among the diagnostic tests not yet applied to the device under test given the expected proportions and a set of costs associated with applying the diagnostic tests not yet applied to the device under test.

10. A diagnostic system that enables a selection of one or more next tests from among a set of diagnostic tests not yet applied to a device under test in response to a test result for each of a subset of the diagnostic tests that have been applied to the device under test wherein each diagnostic test covers a portion of one or more of the components, and in response to a set of patterns of test results and good/bad states of the components and a frequency of occurrence indication for each pattern and wherein the patterns of test results and good/bad states of the components and the frequency of occurrence indication for each pattern are generated by a statistical simulation of the test results and probable good/bad states in response to a probability of failure of each component and a set of test coverage information for the diagnostic tests.

11. The diagnostic system of claim 10, wherein the probability of failure of each component is based upon an a priori engineering estimate.

12. The diagnostic system of claim 10, wherein the test coverage information includes a test coverage indication for each of the diagnostic tests such that each test coverage indication specifies the portions of the components that are tested by the corresponding diagnostic test.

13. The diagnostic system of claim 10, wherein the test coverage information includes a set of joint coverage indications each of which specify a known dependency among the diagnostic tests wherein each known dependancy occurs when two or more of the diagnostic tests covers a portion of one or more of the components in the same way.

15. A diagnostic system that enables a determination of one or more most-likely bad components from among a set of components in a device under test in response to a test result for each of a set diagnostic tests that have been applied to the device under test wherein each diagnostic test covers a portion of one or more of the components, and in response to a set of patterns of test results and good/bad states of the components and a frequency of occurrence indication for each pattern, and in response to an indication of economic costs associated with replacing each component.

16. The diagnostic system of claim 15, wherein the determination of one or more most-likely bad components includes:

generating a first sum of the frequency of occurrence indications that correspond to the test results;

for each of the components, generating a second sum of the frequency of occurrence indications that correspond to the test results and that indicate that the component is BAD, and writing a list of most likely bad components to indicate that the component is GOOD if the second sum is greater than the first sum and otherwise writing the list of most likely bad components to indicate that the component is BAD.

17. The diagnostic system of claim 16, wherein the determination of one or more most-likely bad components further includes selecting a component for replacement from among the components indicated as BAD in the list of most likely bad components in response to the indication of economic costs associated with replacing each component.

18. The diagnostic system of claim 17, wherein a user performs selecting a component for replacement from among the components indicated as BAD in the list of most likely bad components in response to the indication of economic costs associated with replacing each component.

19. The diagnostic system of claim 15, wherein the patterns of test results and good/bad states of the components and the frequency of occurrence indication for each pattern are generated by a statistical simulation of the test results and probable good/bad states in response to a probability of failure of each component and a set of test coverage information for the diagnostic tests.

20. The diagnostic system of claim 19, wherein the probability of failure of each component is based upon an a priori engineering estimate.

21. The diagnostic system of claim 19, wherein the test coverage information includes a test coverage indication for each of the diagnostic tests such that each test coverage indication specifies the portions of the components that are tested by the corresponding diagnostic test.

22. The diagnostic system of claim 19, wherein the test coverage information includes a set of joint coverage indications each of which specify a known dependency among the diagnostic tests wherein each known dependancy occurs when two or more of the diagnostic tests covers a portion of one or more of the components in the same way.

23. The diagnostic system of claim 15, wherein the patterns of test results and good/bad states of the components and the frequency of occurrence indication for each pattern represent actual test results and good/bad states obtained from historical data.

24. The diagnostic system of claim 15, wherein the determination of one or more most-likely bad components includes:

generating a sum N of the frequency of occurrence indications of the patterns that match the test results;

for each pattern that matches the test results, if the pattern exists in a list of most likely bad components, then adding the frequency of occurrence indication of the pattern to a number of times indication for the pattern in the list of most likely bad components, and otherwise adding the pattern to the list of most likely bad components and writing the frequency of occurrence indication of the pattern to a number of times indication for the pattern in the list of most likely bad components;

for each pattern in the list of most likely bad components, replacing the number of times indication with the number of times indication divided by N.

25. The diagnostic system of claim 24, further comprising sorting the list of most likely bad components by the number of times indications.