Real-Time, Model-Based Autonomous Reasoner and Method of Using the Same

• US 20110060569A1
• Filed: 09/07/2010
• Published: 03/10/2011
• Est. Priority Date: 09/04/2009
• Status: Abandoned Application

• Current Assignee: Ridgetop Group, Inc.
• Original Assignee: Ridgetop Group, Inc.
• Inventors: Kunst, Neil, Lynn, Christopher C., Kim, Byoung Uk
• Application Number:
  US12/876,647
• Publication Number:
  US 20110060569A1
• Time in Patent Office: Days
• US Class Current: 703/6
• CPC Class Codes: G06F 11/0739 in a data processing system embedded in automotive or aircraft systems
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06

Abstract

An apparatus and method for detecting and classifying in real-time characteristic of a system component is provided. A sensor senses the system component and outputs a first quantity of data corresponding to a characteristic of the system component. A modeler receives the first quantity of data, converts it to a numerical value and runs a computer model simulation to detect an anomalous behavior of the system component. The detected anomalous behavior is optimized and expressed as a second quantity of data. An autonomous reasoner collects the second quantity of data. A database has a plurality of signatures related to predominant modes of the system component. The autonomous reasoner compares the second quantity of data with the signatures and identifies a signature that matches the second quantity of data. An output indicative of a cause of the anomalous behavior of the system component is provided.

Claims

1. An apparatus for detecting and classifying in real-time a characteristic of a system component, the apparatus comprising:

at least one sensor positioned to sense the system component and output a first quantity of data corresponding to at least one characteristic of the system component;

a computer modeler in communication with at least one sensor and receiving the first quantity of data from the at least one sensor, wherein the computer modeler converts the first quantity of data into at least one numerical value and runs a computer model simulation, wherein the computer model simulation models a present performance of the system component based on the at least one numerical value and compares the modeled present performance to an actual performance of the system component to detect an anomalous behavior of the system component, wherein the detected anomalous behavior is optimized and then expressed as a second quantity of data;

an autonomous reasoner in communication with the computer modeler wherein the autonomous reasoner collects the second quantity of data;

a database in communication with the autonomous reasoner, the database having a plurality of signatures related to predominant modes of the system component, wherein the autonomous reasoner compares the second quantity of data with the signatures related to predominant modes of the system component in the database and identifies at least one signature related to predominant modes that substantially matches the second quantity of data; and

an output of the autonomous reasoner corresponding to the identified at least one signature, wherein the output is indicative of a cause of the at least one anomalous behavior of the system component.

2. The apparatus for detecting and classifying in real-time a behavior of the system component of claim 1, wherein the at least one sensor senses the system component by retrieving at least one signal within a bus.

3. The apparatus for detecting and classifying in real-time a behavior of the system component of claim 1, wherein the computer modeler optimizes the detected anomalous behavior by adjusting a quantity of pre-specified variables until at least one of a substantially suitable match for the second quantity of data is found and an iteration limit is reached.

4. The apparatus for detecting and classifying in real-time a behavior of the system component of claim 3, wherein the iteration limit is reached when no substantially suitable match is found.

5. The apparatus for detecting and classifying in real-time a behavior of the system component of claim 1, wherein the anomalous behavior of the system component is detected when the modeled present performance and the actual performance of the system component are not substantially matched.

6. The apparatus for detecting and classifying in real-time a behavior of the system component of claim 1, wherein communication between at least one of the computer modeler and the at least one sensor, the autonomous reasoner and the computer modeler, and the database and the autonomous reasoner is facilitated with at least one wired communication connection.

7. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, wherein communication between at least one of the computer modeler and the at least one sensor, the autonomous reasoner and the computer modeler, and the database and the autonomous reasoner is facilitated with at least one wireless communication connection.

8. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, further comprising a system computer programmed to at least partially control the system component, wherein the output of the autonomous reasoner is communicated to the system computer, whereby the system computer adjusts control of the system component based on the indicated cause of the at least one anomalous behavior of the system component.

9. The apparatus for detecting and classifying in real-time behavior of the system component of claim 8, wherein the system computer sends a communication to an operator of a system, wherein the communication includes at least one of a notification of a system error, a notification of a system component error, a notification of adjustment of control of the system component, a notification of the output of the autonomous reasoner and a recommendation for future operation of the system.

10. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, wherein the system component further comprises a plurality of system components positioned to be sensed by the at least one sensor and the first quantity of data corresponds to at least one characteristic for each of the plurality of system components.

11. The apparatus for detecting and classifying in real-time behavior of the system component of claim 10, wherein the plurality of system components are housed within at least two independent systems.

12. The apparatus for detecting and classifying in real-time behavior of a system component of claim 1, wherein each of the plurality of signatures related to predominant modes of the system component substantially match at least one state of degradation of the system component.

13. The apparatus for detecting and classifying in real-time behavior of the system component of claim 12, wherein the at least one state of degradation corresponds to an identifiable type of degradation of the system component.

14. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, wherein the modeled a present performance of the system component is based in part, on at least one of an environmental condition and a physical condition.

15. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, wherein the computer modeler optimizes the detected anomalous behavior in substantially real-time.

16. The apparatus for detecting and classifying in real-time behavior of the system component of claim 1, further comprising a sensor output from the at least one sensor, wherein the sensor output is stored, and wherein the computer modeler optimizes the detected anomalous behavior at least partly based on the stored sensor output.

17. A method for detecting and classifying in real-time a characteristic of a system component, the method comprising the steps of:

sensing the system component with at least one sensor;

outputting from the at least one sensor a first quantity of data corresponding to at least one characteristic of the system component;

receiving the first quantity of data at a computer modeler in communication with the at least one sensor;

converting the first quantity of data into at least one numerical value;

modeling a present performance of the system component based on the at least one numerical value;

detecting an anomalous behavior of the system component by comparing the modeled present performance to an actual performance of the system component;

optimizing the detected anomalous behavior and expressing it as a second quantity of data;

collecting the second quantity of data with an autonomous reasoner in communication with the computer modeler;

comparing the second quantity of data with a plurality of signatures related to predominant modes of the system component stored in a database in communication with the autonomous reasoner;

identifying at least one signature related to predominant modes that substantially matches the second quantity of data; and

outputting an output of the autonomous reasoner corresponding to the identified at least one signature, wherein the output is indicative of a cause of the at least one anomalous behavior of the system component.

18. The method for detecting and classifying in real-time characteristic of the system component 17, wherein the step of sensing the system component with at least one sensor further comprises retrieving at least one signal within a bus.

19. The method for detecting and classifying in real-time characteristic of the system component 17, wherein the step of optimizing the detected anomalous behavior with the computer modeler further comprises adjusting a quantity of pre-specified variables until at least one of a substantially suitable match for the second quantity of data is found and an iteration limit is reached.

20. The method for detecting and classifying in real-time characteristic of the system component 17, wherein the iteration limit is reached when no substantially suitable match is found.

21. The method for detecting and classifying in real-time characteristic of the system component 17, wherein the step of detecting the anomalous behavior of the system component further comprises detecting the anomalous when the modeled present performance and the actual performance of the system component are not substantially matched.

22. The method for detecting and classifying in real-time characteristic of the system component 17, further comprising the step of activating an autonomous reasoner when an anomalous behavior of the system component is detected.

23. The method for detecting and classifying in real-time characteristic of the system component 17, further comprising the steps of:

at least partially controlling the system component with a system computer;

communicating the output of the autonomous reasoner to the system computer; and

adjusting a control of the system component based the indicated cause of the at least one anomalous behavior of the system component.

24. The method for detecting and classifying in real-time characteristic of the system component 23, further comprising the step of sending a communication from the system computer to an operator of a system, wherein the communication includes at least one of a notification of a system error, a notification of a system component error, a notification of adjustment of control of the system component, a notification of the output of the autonomous reasoner and a recommendation for future operation of the system.

25. The method for detecting and classifying in real-time characteristic of the system component 17, further comprising the steps of:

sensing a plurality of system components with the at least one sensor; and

outputting the quantity of data corresponding to the at least one characteristic for each of the plurality of system components.

26. The method for detecting and classifying in real-time characteristic of the system component 25, further comprising the step of housing the plurality of system components within at least two independent systems.

27. The method for detecting and classifying in real-time characteristic of the system component 17, further comprising the step of substantially matching each of the plurality of signatures related to predominant modes of the system component to at least one state of degradation of the system component.

28. The method for detecting and classifying in real-time characteristic of the system component 27, wherein the at least one state of degradation corresponds to an identifiable type of degradation of the system component.

29. An apparatus for detecting and classifying in real-time a characteristic of a system component:

at least one sensor positioned to sense the system component and output a first quantity of data corresponding to at least one characteristic of the system component;

a computer modeler in communication with the at least one sensor and receiving the first quantity of data from the at least one sensor, wherein the computer modeler runs a computer model simulation of the system component to detect an anomalous behavior of the system component, wherein the detected anomalous behavior is optimized and expressed as a second quantity of data;

an autonomous reasoner in communication with the computer modeler wherein the autonomous reasoner collects the second quantity of data;

a database in communication with the autonomous reasoner, the database having a plurality of signatures related to predominant modes of the system component, wherein the autonomous reasoner compares the second quantity of data with the signatures related to predominant modes of the system component in the database and identifies at least one signature related to predominant modes that substantially matches the second quantity of data; and

an output of the autonomous reasoner corresponding to the identified at least one signature, wherein the output is indicative of a cause of the at least one anomalous behavior of the system component.