Method of determining the cause of an error state in an apparatus

• US 7,600,671 B2
• Filed: 11/29/2005
• Issued: 10/13/2009
• Est. Priority Date: 12/17/2004
• Status: Active Grant

• Current Assignee: NCR Corporation
• Original Assignee: NCR Corporation
• Inventors: Forrest, Simon J., Timmis, Jon, Ayara, Modupe, de Lemos, Rogerio
• Primary Examiner(s): Le; Thien M.
• Assistant Examiner(s): Mai; Thien T
• Application Number:
  US11/288,777
• Publication Number:
  US 20060131380A1
• Time in Patent Office: 1,414 Days
• Field of Search: 235/379, 235/376, 235/383, 714/47, 700/29, 700/79, 700/76, 700/83, 707/102, 702/183, 702/184, 435/6, 717/135
• US Class Current: 235/376
• CPC Class Codes: G07D 11/235 for monitoring or indicating operating conditions; for detecting malfunctions
• CPC Class Codes: G07F 19/20 Automatic teller machines [ATMs]
• CPC Class Codes: G07F 19/207 Surveillance aspects at ATMs
• CPC Class Codes: G07F 9/026 for alarm, monitoring and auditing in vending machines or means for indication, e.g. when empty

Abstract

A method is described for determining the cause of an error state for one or more components within an apparatus. The apparatus comprises a plurality of sensors arranged to monitor the operation of components of the apparatus and a control means arranged to receive said information from said plurality of sensors. The method comprises analysing said sensor information in the form of an error log to ascertain sensor patterns from said sensor information comparing said sensor patterns with detectors, which are predefined patterns, indicative of the condition of said one or more components within the apparatus and classifying said sensor patterns as being indicative of said error state of a component or not based upon a comparison of sensor patterns with said detectors.

Claims

1. A method of determining the cause of an error state for one or more components within an apparatus comprising a plurality of sensors arranged to monitor the operation of components of the apparatus and a control means arranged to receive information from said plurality of sensors, the method comprising:

a) during operation of the apparatus, electronically analyzing said sensor information to ascertain sensor patterns from said sensor information;

b) electronically comparing said sensor patterns with detectors comprising an adaptable error prediction system, each detector being a previously defined pattern, indicative of the condition of said one or more components within the apparatus and classifying a matching present state of the apparatus as being a precursor to a fatal state of as not being a precursor to a fatal state;

c) upon detection during operation of the apparatus of a sensor pattern unable to be classified as indicating a precursor to a fatal or nonfatal state, electronically creating an immature detector to classify the sensor pattern, tolerizing the immature detector to evaluate if the detector correctly classifies a condition and discarding the immature detector if it fails to achieve tolerization existing detectors and known healthy states;

d) upon validation of an immature detector, electronically adding the detector to the adaptable error prediction system; and

e) electronically classifying said sensor patterns as being indicative of a precursor to a fatal or nonfatal state based upon a comparison of sensor patterns with said validated detectors comprising the adaptable error prediction system.

2. A method as claimed in claim 1, wherein the use of specific detectors is tailored to a known situation based on characteristics including one or more of:

apparatus type, apparatus make, operational environment and patterns of specific usage of the apparatus.

3. A method as claimed in claim 1, wherein time stamps within an error log are used to provide information on a sequence of states which leads to a final fatal error state.

4. A method as claimed in claim 2, wherein information relating to the behavior of the apparatus is stored in a log file and wherein graphical methods of displaying the information in the log file provide a high level overview of the behaviour of the apparatus.

5. A method as claimed in claim 1, wherein one or more detectors are determined prior to the operation of the apparatus.

6. A method as claimed in claim 5, wherein one or more detectors are created from off-line analysis of logs of previous sensor information.

7. A method as claimed in claim 5, wherein one or more detectors are created during the design of an apparatus.

8. A method as claimed in claim 5, wherein one or more detectors are refined during are refined during operation of the apparatus.

9. A method as claimed in claim 1, wherein detectors include information from two or more sensors received over a period of time.

10. A method as claimed in claim 1, wherein detectors are given a weighting determined by their rate of previous success in correctly predicting the failure of a component within/associated with the apparatus.

11. A method as claimed in claim 1, wherein detectors are continually evaluated during operation of the apparatus to optimise the selection of detectors for the apparatus.

12. A method as claimed in claim 11, wherein an immature detector is validated only when it has proven to successfully predict the future state of health/condition of said one or more components to a pre-determined level of acceptability.

13. A machine readable medium storing a program of instructions executable by a computer, the computer operating under the control of the program of instructions to perform a method comprising the steps of:

a) during operation of the apparatus, analyzing sensor information to ascertain sensor patterns from said sensor information;

b) compare said sensor patterns with detectors comprising an adaptable error prediction system, which are previously defined patterns, indicative of the condition of said one or more components within the apparatus and classifying a matching present state of the apparatus as being a precursor to a fatal state or as not being a precursor to a fatal state;

c ) upon detection during operation of the apparatus of a sensor pattern unable to be classified as indicating a precursor to a fatal or nonfatal state, creating an immature detector to classify the sensor pattern, tolerizing the immature detector to evaluate if the detector correctly classifies a condition and discarding the immature detector if it fails to achieve tolerization within a prescribed lifespan, and validating the detector by one or more of evaluation against existing detectors and known healthy states;

d) upon validation of an immature detector, adding the detector to the adaptable error prediction system; and

e) classifying said sensor patterns as being indicative of a precursor to a fatal or nonfatal state based upon a comparison of sensor patterns with said detectors.

14. A machine readable medium as claimed in claim 13, wherein the use of specific detectors is tailored to a known situation based on characteristics including one or more of;

apparatus type, apparatus make, operational environment and patterns of specific usage of the apparatus.

15. A machine readable medium as claimed in claim 13, wherein time stamps within an error log are used to provide information on a sequence of states which leads to a final fatal error state.

16. A machine readable medium as claimed in claim 14, wherein graphical methods of displaying the information in the log file provide a high level overview of the apparatus behaviour.

17. A machine readable medium as claimed in claim 13, wherein the detectors are determined prior to the operation of the apparatus.

18. A machine readable medium as claimed in claim 17, wherein the detectors are created from off-line analysis of logs of previous sensor information.

19. A machine readable medium as claimed in claim 17, wherein the detectors are created during the design of an apparatus.

20. A machine readable medium as claimed in claim 17, wherein the detectors are refined during operation of the apparatus.

21. A machine readable medium as claimed in claim 13, wherein detectors include information from two or more sensors received over a period of time.

22. A machine readable medium as claimed in claim 13, wherein detectors are given a weighting determined by their rate of previous success in correctly predicting the failure of a component within/associated with the apparatus.

23. A machine readable medium as claimed in claim 13, wherein a newly created detector is used more frequently if it proves to be successful in correctly predicting the failure of a component within/associated with the apparatus.

24. A machine readable medium as claimed in claim 13, wherein detectors are continually evaluated during operation of the apparatus to optimise the selection of detectors for the apparatus.

25. A machine readable medium as claimed in claim 24, wherein an immature detector is validated only when it has proven to successfully determine the cause of an error state in said one or more components to a pre-determined level of acceptability.