Modeling when connections are the problem

• US 7,962,321 B2
• Filed: 07/10/2007
• Issued: 06/14/2011
• Est. Priority Date: 07/10/2007
• Status: Active Grant

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Palo Alto Research Center, Inc. (Xerox Holdings Corp.)
• Inventors: de Kleer, Johan
• Primary Examiner(s): Proctor; Jason
• Assistant Examiner(s): Jacob; Mary C
• Application Number:
  US11/775,650
• Publication Number:
  US 20090018802A1
• Time in Patent Office: 1,435 Days
• Field of Search: 703/7, 703/14
• US Class Current: 703/14
• CPC Class Codes: G06F 30/3323 using formal methods, e.g. equivalence checking or property checking

Abstract

A method and computer program product models connections for a system having a plurality of components, having component terminals, and connections. The connections are located to selectively connect the components of the system to each other. The method and computer product models signal levels for nodes of the system as a variable wherein the nodes of the system are representative of the connections. Each potential component terminal connection of the connections that could potentially influence the signal level on each of the nodes is modeled, as a qualitative variable. This results in each node of the system being modeled with n+1 variables.

Claims

1. A method for modeling connections of a system having a plurality of components and connections, the connections being located to connect the components to each other, the method comprising:

modeling components of the system;

modeling signal levels for nodes of the system as a variable, wherein the nodes of the system are representative of the connections; and

modeling each potential component terminal connection of the connections which could potentially influence the signal level on each of the nodes as a qualitative variable, wherein each node of the system is modeled with n+1 variables, and the components which are modeled include devices which explicitly determine whether drive outputs can be directly modeled.

2. The method according to claim 1, wherein each component is considered a causal driver of a corresponding node.

3. The method according to claim 1, wherein the system is an electrical circuit.

4. The method according to claim 1, wherein opens, shorts and bridge faults of an electrical circuit are directly modeled.

5. The method according to claim 1, wherein modeling of faults are done within a GDE/Sherlock framework.

6. The method according to claim 1, wherein candidate diagnoses of potential faults are generated in a best-first order search arrangement, avoiding potential computational complexity introduced by more detailed causal models.

7. The method according to claim 6, wherein the candidates are ordered by posterior probabilities, not a number of faults they contain.

8. The method according to claim 1, wherein the modeling of the connections, models connections whose faults are joining faults.

9. The method according to claim 1, wherein the components include causal inputs and causal outputs.

10. The method according to claim 9, wherein,

the causal inputs to the components are the signal levels at all the system nodes the components are connected to, and

the causal outputs of the components are the driving signals on all of the connections of the nodes to which the components connect.

11. The method according to claim 10, wherein,

a component is causally non-intermittent if all its driving outputs are a function of its sensed inputs.

12. The method according to claim 1, wherein the system is a mechanical system.

13. The method according to claim 12, wherein leaks in connections are directly modeled.

14. A method for modeling connections of a system having a plurality of components and connections, the connections being located to connect the components to each other, the method comprising:

modeling components of the system;

modeling signal levels for nodes of the system as a variable, wherein the nodes of the system are representative of the connections; and

modeling each potential component terminal connection of the connections which could potentially influence the signal level on each of the nodes as a qualitative variable, wherein each node of the system is modeled with n+1 variables and a change in causality is represented by an increased complexity of the model having a worst case linear increase in the number of modeled components.

15. A computer program product for use with a computing device, the computer product comprising:

a non-transitory computer readable medium that stores instructions, wherein when executed by a processor the instructions perform the following operations;

modeling components of a system,

modeling connections of the system having a plurality of components and connections, the connections being located to selectively connect the components to each other, the modeling of the connections including,

modeling signal levels for nodes of the system as a variable, wherein the nodes of the system are representative of the connections; and

modeling each potential component terminal connection of the connections which could potentially influence the signal level on each of the nodes as a qualitative variable, wherein each node of the system is modeled with n+1 variables, wherein the components which are modeled include devices which explicitly determine whether drive outputs can be directly modeled.

16. The computer product of claim 15, wherein each component is considered a causal driver of a corresponding node.

17. The computer product of claim 15, wherein the system is an electrical circuit.

18. The computer product of claim 15, wherein opens, shorts and bridge faults of an electrical circuit are directly modeled.

19. The computer product of claim 15, wherein the components include causal inputs and causal outputs.

20. The computer product of claim 19, wherein

the causal inputs to the components are the signal levels at all the system nodes the components are connected to, and

the causal outputs of the components are the driving signals on all of the connections of the nodes to which the components connect.

21. The computer product of claim 20, wherein,

a component is causally non-intermittent if all its driving outputs are a function of its sensed inputs.

22. The computer product of claim 15, wherein the system is a mechanical system.

23. The computer product of claim 22, wherein leaks in connections are directly modeled.