Adaptable state based control system

• US 6,745,089 B2
• Filed: 02/01/2001
• Issued: 06/01/2004
• Est. Priority Date: 02/01/2000
• Status: Expired due to Fees

• Current Assignee: California Institute of Technology
• Original Assignee: California Institute of Technology
• Inventors: Keller, Robert M., Gat, Erann, Starbird, Thomas W., Dvorak, Daniel L., Chien, Steve Ankuo, Rasmussen, Robert D., Gostelow, Kim P.
• Primary Examiner(s): Patel, Ramesh
• Assistant Examiner(s): BARNES-BULLOCK, CRYSTAL JOY
• Application Number:
  US09/776,573
• Publication Number:
  US 20020026249A1
• Time in Patent Office: 1,216 Days
• Field of Search: 700/28, 700/29, 700/30, 700/31, 700/51, 700/52, 700/75, 318/561, 703/2
• US Class Current: 700/51
• CPC Class Codes: G05B 13/02 electric
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 2219/23255 Object oriented programming, OOP
• CPC Class Codes: G05B 2219/23425 Selection of program, adaptive to process

Abstract

An autonomous controller, comprised of a state knowledge manager, a control executor, hardware proxies and a statistical estimator collaborates with a goal elaborator, with which it shares common models of the behavior of the system and the controller. The elaborator uses the common models to generate from temporally indeterminate sets of goals, executable goals to be executed by the controller. The controller may be updated to operate in a different system or environment than that for which it was originally designed by the replacement of shared statistical models and by the instantiation of a new set of state variable objects derived from a state variable class. The adaptation of the controller does not require substantial modification of the goal elaborator for its application to the new system or environment.

Claims

1. An autonomous system that is controlled based on temporal constraints placed upon the behavior of the autonomous system and statistically indeterminate knowledge about the behavior of the autonomous system, the autonomous system comprising:

multiple hardware proxies which provide an interface between a controller and system hardware;

a statistical state estimator, which provides statistical estimated values and estimated uncertainty values of the state of the system and the controller in response to temporal and behavior constraints and based upon measurements received from the hardware proxies;

a control executor that is configured to issue commands to the hardware proxies based upon temporal and behavioral constraints on the system, the statistical estimated values, and the estimated uncertainty values;

a state knowledge manager which is configured to coordinate the use the statistical estimates and uncertainties the state of the system; and

a set of common models for the behavior of the system, the controller, and the external environment, wherein the common models are used by the state estimator, the control executor, and the state knowledge manager.

2. The autonomous system of claim 1, wherein the controller may be adapted for use in a second system or in a second external environment, by substituting a second set of common models for the first set of common models.

3. The autonomous system of claim 1, wherein the control executor, state knowledge manager, and statistical state estimator are implemented as a first set of collaborating objects in an object-oriented software system.

4. The autonomous system of claim 3, wherein the control executor, the state knowledge manager, and the statistical state estimator are adapted for use in a controller operating in a second system or in a second external environment by substituting a second set of collaborating objects for the first set of collaborating objects.

5. An autonomous system that is controlled based on temporal constraints placed upon the behavior of the autonomous system and statistically indeterminate knowledge about the behavior of the autonomous system, the autonomous system comprising:

multiple hardware proxies which provide an interface between a controller and system hardware;

a statistical state estimator, which provides statistical estimated values and estimated uncertainty values of the state of the system and the controller in response to temporal and behavior constraints and based upon measurements received from the hardware proxies;

a control executor that is configured to issue commands to the hardware proxies based upon temporal and behavioral constraints on the system, the statistical estimated values, and the estimated uncertainty values;

a state knowledge manager which is configured to coordinate the use the statistical estimates and uncertainties the state of the system;

a set of common models for the behavior of the system, the controller, and the external environment, wherein the common models are used by the state estimator, the control executor, and the state knowledge manager; and

a goal elaborator, wherein the goal elaborator hierarchically decomposes goals on the behavior of a derived state of the system into goals on a set of state variables which define the derived state of the system.

6. The autonomous system of claim 5, wherein the goal elaborator converts goals expressed with temporal indeterminance into goals on the behavior of the state of the system.

7. The autonomous system of claim 6, wherein the set of common models may be substituted with a second set of common models without modifying the goal elaborator.

8. The autonomous system of claim 7, wherein the first and second set of common models are implemented as object instantiations of a class of state variables corresponding to, and having the values of the attributes of, the first and second set of state variables.

9. A method of autonomous control which shares information about a set of state variables with a goal elaborator and a controller, the method comprising:

generating with a goal elaborator, desired values associated with a set of state variables wherein the state variables define the condition of an autonomous device;

generating with a controller, estimated values associated with the set of state variables;

generating with the controller, estimated uncertainty values associated with the estimate values; and

sharing with the goal elaborator and the controller, the desired values, the estimated values, and the estimated uncertainty values.

10. The method of claim 9 wherein the act of generating the desired values further comprises hierarchically decomposing goals on the behavior of a derived state of the system.

11. The method of claim 9 wherein the act generating the desired values further comprises hierarchically decomposes goals on the behavior of a derived state of the system into goals on a set of state variables which define the;

derived state of the system.

12. The method of claim 9 wherein the act of generating the estimated values is based upon temporal and behavioral constraints on the system.

13. The method of claim 9 wherein the act of generating the estimated values is based upon statistical estimates.

14. The method of claim 9 wherein the act of generating the estimated uncertainty values is based upon uncertainties of the state of the system.

15. A method of control comprising:

generating with an elaborator, a set of desired values that are associated with a set of state variables;

generating with a controller, a set of estimated values that are associated with the set of state variables;

generating with a controller, a set of estimated uncertainty values that are associated with the estimate values; and

sharing with both the elaborator and controller, the desired values, the estimated values, and the estimated uncertainty values.

16. The method of claim 15 wherein the act of sharing is performed by a state knowledge manager that is in communication with the elaborator and controller.

17. The method of claim 15 wherein the act of generating the set of estimated values is in response to temporal constraints.

18. The method of claim 15 wherein the act of generating the set of estimated values is in response to temporal and behavior constraints.

19. The method of claim 15 wherein the act of generating the set of estimated values is in response to temporal constraints, behavior constraints and upon measured data.

20. The method of claim 15 wherein the act of generating the set of estimated uncertainty values is in response to temporal constraints.

21. The method of claim 15 wherein the act of generating the set of estimated uncertainty values is in response to temporal and behavior constraints.

22. The method of claim 15 wherein the act of generating the set of estimated uncertainty values is in response to temporal constraints, behavior constraints and upon measured data.

23. A method of autonomous control that combines temporal elaboration with statistical modeling, the method comprising:

generating with a controller statistically estimated values associated with a set of state variables;

generating with a controller statistically estimated uncertainty values associated with the statistically estimated values; and

generating with a goal elaborator a set of desired values associated with a set of state variables wherein the act of generating the desired values is based on the statistically estimate values, statistically estimated uncertainty values and a set of temporally indeterminate goals.

24. The method of claim 23 wherein the act of generating the statistically estimated uncertainty values is based on statistical models associated with the controller and the desired values generated by the elaborator.

25. A method of autonomous control that combines temporal elaboration with statistical modeling, the method comprising:

generating with an elaborator based on a set of temporally indeterminate goals, desired values associated with a set of state variables; and

generating with a controller statistically estimated values associated with the set of state variables wherein the act of generating the statistically estimated values is based on statistical models associated with the controller and the desired values generated by the elaborator.

26. The method of claim 25 further comprising generating with the controller statistically estimated uncertainty values based on statistical models associated with the controller, the statistically estimated values, and the desired values generated by the elaborator.

27. A control apparatus comprising:

an elaborator which is configured to generate a set of desired values, wherein the set of desired values are associated with a set of state variables;

a controller which is configured to generate a set of estimated values that are associated with the set of state variables, the controller further configured to generate a set of estimated uncertainty values that are associated with the estimate values; and

wherein the controller and elaborator are configured to share the desired values, the estimated values, and the estimated uncertainty values.

28. The control apparatus of claim 27 further comprising multiple hardware proxies which provide measurement data to the controller.

29. The control apparatus of claim 27 wherein the controller further comprises a statistical state estimator, which provides the estimated values and the estimated uncertainty values in response to temporal and behavior constraints.

30. The control apparatus of claim 27 further comprising a set of common models for the behavior of the system, the controller, or an external environment.

31. A method of reconfiguring a controller and an elaborator associated with a first autonomous system so that the controller and elaborator can direct a second autonomous system, the method comprising:

replacing a first set of state variables associated with a first autonomous system with a second set of state variables associated with a second autonomous system;

replacing a first set of statistical models associated with the first set of state variables with a second set of statistical models associated with the second set of state variables; and

automatically sharing with the elaborator and controller, data associated with the second set of state variables.

32. An method of adapting the use of an autonomous controller configured to operate within a first system for a second system, wherein the adaptation comprises:

(A) replacing a first set of state variables associated with a first system and with a second set of state variables associated with a second system;

(B) identifying attribute values associated with members of the second set of state variables;

(C) replacing a first set of statistical models that estimate and predict the behavior of the first system with a second set of statistical models that estimate the behavior of the second system;

(D) replacing a first set of models for the behavior of sensors and actuators in the first system with a second set of models of the behavior of sensors and actuators in the second system; and

(E) sharing with an elaborator and a controller, the second set of state variables and the behavior of the second system predicted by the statistical models.

33. The method of claim 32, wherein (A) and (B) are achieved by instantiating a set of software objects of a state variable class corresponding to, and having the values of the attributes of, the set of state variables of the second system.

34. The method of claim 32, wherein the attributes of the set of state variables of the second system include the definition of the probability distribution of the state variables of the second system.

35. The method of claim 32 wherein (B) comprises:

(B-1) defining a member of the set of state variables of the second system as a discrete or continuous variable;

(B-2) defining the probability distribution of the member of the set of state variables of the second system;

(B-3) identifying a time interval over which the member of the set of state variables of the second system is defined; and

(B-4) producing a mathematical model for the time varying nature of the member of the set of state variables for the second system during the time interval over which the member of the set of state variables of the second system is defined.

36. The method of claim 35, wherein (B) further comprises:

(B-5) describing a set of monitoring criteria which delineate conditions when the value of the member of the set of state variables of the second system will result in a notification of the condition being met being transmitted to a control executor within the controller.

37. The method of claim 32, wherein (C) comprises:

(C-1) identifying a source of evidence relating to a member of the set of state variables of the second system;

(C-2) identifying a measurement from the source of evidence relating to a member of the set of state variables of the second system;

(C-3) defining the mathematical relationship between the member of the set of state variables of the second system and the measurement from the source of evidence;

(C-4) defining the statistical uncertainty distribution for the measurement from the source of evidence;

(C-5) defining the effect of a command issued to an actuator on the member of the set of state variables of the second system; and

(C-6) defining the estimation filter algorithm to be used estimate the value of, and the uncertainty in the knowledge of, the member of the set of state variables of the second system.