Method and apparatus for estimating a parameter based on a plurality of redundant signals

• US 20050228619A1
• Filed: 04/07/2004
• Published: 10/13/2005
• Est. Priority Date: 04/07/2004
• Status: Active Grant

• Current Assignee: Rtx Corporation
• Original Assignee: United Technologies Corporation (Rtx Corporation)
• Inventors: Burnet, Cheri S., Powell, Cary T.
• Granted Patent: US 7,130,772 B2
• Time in Patent Office: Days
• US Class Current: 702/189
• CPC Class Codes: G01D 3/08 with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown
• CPC Class Codes: G05B 13/0275 using fuzzy logic only
• CPC Class Codes: G05B 9/03 with multiple-channel loop, i.e. redundant control systems
• CPC Class Codes: G06N 5/048 Fuzzy inferencing

Abstract

In a method for estimating a parameter based on signals received from redundant sensors, at least a first sensed signal and a second sensed signal are received from at least corresponding first and second redundant sensors. The first sensed signal and the second sensed signal are indicative of the parameter, wherein the first sensed signal has associated therewith a first accuracy, wherein the second sensed signal has associated therewith a second accuracy. At least a reference signal indicative of the parameter is received, wherein the reference signal has associated therewith a reference accuracy. A weighting is determined based on at least the first sensed signal, the second sensed signal, and based on at least one of the first accuracy, the second accuracy, and the reference accuracy. An estimate of the parameter is generated as a weighted average, according to the weighting, of at least a value of the first sensed signal, a value of the second sensed signal, and a value of the reference signal.

Claims

1. A method for estimating a parameter based on signals received from redundant sensors, the method comprising:

receiving at least a first sensed signal and a second sensed signal from at least corresponding-first and second redundant sensors, the first sensed signal and the second sensed signal indicative of the parameter, wherein the first sensed signal has associated therewith a first accuracy, wherein the second sensed signal has associated therewith a second accuracy;

receiving at least a reference signal indicative of the parameter, wherein the reference signal has associated therewith a reference accuracy;

determining a weighting based on at least the first sensed signal, the second sensed signal, and based on at least one of the first accuracy, the second accuracy, and the reference accuracy; and

generating an estimate of the parameter as a weighted average, according to the weighting, of at least a value of the first sensed signal, a value of the second sensed signal, and a value of the reference signal.

2. A method as defined in claim 1, further comprising generating a first error magnitude signal based on the first sensed signal and the second sensed signal.

3. A method as defined in claim 2, wherein determining the weighting and generating the estimate of the parameter comprises determining the estimate of the parameter as an average of a value of the first sensed signal and a value of the second sensed signal if a value of the first error magnitude signal is less than a threshold, the threshold based on the first accuracy and the second accuracy.

4. A method as defined in claim 3, wherein the threshold is based on the sum of the first accuracy and the second accuracy.

5. A method as defined in claim 4, wherein the threshold equals the sum of the first accuracy and the second accuracy

6. A method as defined in claim 2, further comprising:

generating a second error magnitude signal based on the first sensed signal and the reference signal;

generating a third error magnitude signal based on the second sensed signal and the reference signal;

wherein determining the weighting and generating the estimate of the parameter comprises;

determining the estimate of the parameter as a value of the first sensed signal if a value of the first error magnitude signal is greater than a threshold and if a value of the second error magnitude signal is less than a value of the third error magnitude signal, wherein the threshold is based on the reference accuracy; and

determining the estimate of the parameter as a value of the second sensed signal if the value of the first error magnitude signal is greater than the threshold and if the value of the second error magnitude signal is more than the value of the third error magnitude signal.

7. A method as defined in claim 6, wherein the threshold equals twice the reference accuracy.

8. A method as defined in claim 2, wherein determining the weighting and generating the estimate of the parameter comprises determining the weighting and generating the estimate of the parameter according to a fuzzy rulebase if a value of the first error magnitude signal is greater than a first threshold and if the value of the first error magnitude signal is less than a second threshold, wherein the first threshold is based on the first accuracy and the second accuracy, and wherein the second threshold is based on the reference accuracy.

9. A method as defined in claim 8, further comprising:

generating a second error magnitude signal based on the first sensed signal and the reference signal;

generating a third error magnitude signal based on the second sensed signal and the reference signal;

wherein determining the weighting and generating the estimate of the parameter according to the fuzzy rulebase comprises;

determining degrees of membership of a value of the second error magnitude signal in a first plurality of fuzzy sets;

determining degrees of membership of a value of the third error magnitude signal in a second plurality of fuzzy sets; and

calculating the weighting based on the degrees of membership of the value of the second error magnitude signal and the degrees of membership of the value of the third error magnitude signal.

10. A method as defined in claim 9, further comprising:

normalizing the second error magnitude signal, based on the first accuracy and the reference accuracy, prior to determining the degrees of membership of the value of the second error magnitude signal; and

normalizing the third error magnitude signal, based on the second accuracy and the reference accuracy, prior to determining the degrees of membership of the value of the third error magnitude signal.

11. A method as defined in claim 10, wherein normalizing the second error magnitude signal comprises dividing the second error magnitude signal by the reference accuracy added with the first accuracy;

and

wherein normalizing the third error magnitude signal comprises dividing the third error magnitude signal by the reference accuracy added with the second accuracy.

12. A method as defined in claim 9, wherein determining the weighting and generating the estimate of the parameter further comprises determining a degree of fulfillment of each fuzzy logic rule in the fuzzy rulebase based on the degrees of membership of the value of the second error magnitude signal and the degrees of membership of the value of the third error magnitude signal.

13. A method as defined in claim 12, wherein generating the estimate of the parameter comprises:

determining weighted rule outputs by multiplying consequent portions of the fuzzy logic rules in the fuzzy rulebase by corresponding degrees of fulfillment of the fuzzy logic rules;

summing the weighted rule outputs;

summing the degrees of fulfillment of the fuzzy logic rules; and

dividing the summed weighted rule outputs by the summed degrees of fulfillment.

14. A method as defined in claim 1, wherein receiving the reference signal comprises receiving the reference signal from a model.

15. A method as defined in claim 1, wherein receiving the reference signal comprises receiving the reference signal from a third redundant sensor.

16. A method as defined in claim 1, wherein receiving the reference signal comprises receiving the reference signal from a processing device configured to generate the reference signal based on a signal, received from another sensor, indicative of a different parameter related to the parameter.

17. An apparatus for estimating a parameter based on signals received from redundant sensors, the apparatus comprising:

a weighting calculator to generate weighting information in response to a first sensed signal, a second sensed signal, a reference signal, the weighting information based on a first accuracy associated with the first sensed signal, a second accuracy associated with the second sensed signal, and a reference signal accuracy associated with the reference signal;

wherein the first sensed signal is indicative of the parameter and corresponds to a first redundant sensor, wherein the second sensed signal is indicative of the parameter and corresponds to a second redundant sensor, and wherein the reference signal is indicative of the parameter; and

a weighted average calculator to generate an estimate of the parameter based on the first sensed signal, the second sensed signal, the reference signal, and the weighting information.

18. An apparatus according to claim 17, wherein the weighting calculator comprises:

an error magnitude calculator to calculate a first error magnitude signal based on the first sensed signal and the second sensed signal, a second error magnitude signal based on the first sensed signal and the reference signal, and a third error magnitude signal based on the first sensed signal, the second sensed signal, and the reference signal; and

an estimate selector to generate an indication of one of a plurality of potential estimates of the parameter in response to the first error magnitude signal, the second error magnitude signal, and the third error magnitude signal, and based on the first accuracy, the second accuracy, and the reference signal accuracy.

19. An apparatus according to claim 18, wherein the weighting calculator and the weighted average calculator comprise a fuzzy logic estimate calculator, wherein the plurality of potential estimates comprises the first sensed signal, the second sensed signal, and a fuzzy estimate generated by the fuzzy logic estimate calculator.

20. An apparatus according to claim 19, wherein the weighted average calculator further comprises an average calculator to calculate an average of the first sensed signal and the second sensed signal, wherein the plurality of potential estimates further comprises the average of the first sensed signal and the second sensed signal.

21. An apparatus according to claim 19, wherein the weighted average calculator further comprises a normalizer to generate a first normalized error magnitude signal in response to the second error magnitude signal, and a second normalized error magnitudes signal in response to the third error magnitude signal;

wherein the first normalized error magnitude signal is based on the first accuracy and the reference signal accuracy, and wherein the second normalized error magnitude signal is based on the second accuracy and the reference signal accuracy

wherein the fuzzy logic estimate calculator generates the fuzzy estimate in response to the first normalized error magnitude signal and the second normalized error magnitude signal, and based on the first sensed signal, the second sensed signal, and the reference signal.

22. An apparatus according to claim 17, further comprising a model configured to generate the reference signal.

23. An apparatus according to claim 17, comprising a processor operatively coupled to a memory, the first redundant sensor, and the second redundant sensor, the memory having stored thereon:

first code to implement the weighting calculator; and

second code to implement the weighted average calculator.

24. A method for estimating a parameter based on signals received from redundant sensors, the method comprising:

receiving at least a first sensed signal and a second sensed signal from at least corresponding first and second redundant sensors, the first sensed signal and the second sensed signal indicative of the parameter, wherein the first sensed signal has associated therewith a first accuracy, wherein the second sensed signal has associated therewith a second accuracy;

receiving at least a reference signal indicative of the parameter, wherein the reference signal has associated therewith a reference accuracy;

generating a first error magnitude signal based on at least the first sensed signal and the reference signal;

generating a second error magnitude signal based on at least the second sensed signal and the reference signal;

normalizing the first error magnitude signal, based on at least the first accuracy, the second accuracy, and the reference accuracy;

normalizing the second error magnitude signal, based on at least the first accuracy, the second accuracy, and the reference accuracy;

evaluating a fuzzy rulebase based on at least the normalized first error magnitude signal and the normalized second error magnitude signal; and

generating an estimate of the parameter based on at least the evaluated fuzzy rulebase.

25. A method as defined in claim 24, wherein normalizing the first error magnitude signal comprises dividing the first error magnitude signal by the reference accuracy added with the first accuracy;

and

wherein normalizing the second error magnitude signal comprises dividing the second error magnitude signal by the reference accuracy added with the second accuracy.

26. A method as defined in claim 24, further comprising:

determining degrees of membership of a value of the normalized first error magnitude signal in a first plurality of fuzzy sets; and

determining degrees of membership of a value of the normalized second error magnitude signal in a second plurality of fuzzy sets;

wherein evaluating the fuzzy rulebase comprises determining a degree of fulfillment of each fuzzy logic rule in the fuzzy rulebase based on the degrees of membership of the value of the normalized first error magnitude signal and the degrees of membership of the value of the normalized second error magnitude signal.

27. A method as defined in claim 26, wherein determining the degrees of membership of the value of the normalized first error magnitude signal comprises determining the degrees of membership of the value of the normalized first error magnitude signal according to a plurality of membership functions;

and

wherein determining the degrees of membership of the value of the normalized second error magnitude signal comprises determining the degrees of membership of the value of the normalized second error magnitude signal according to the plurality of membership functions.

28. A method as defined in claim 26, wherein determining degrees of membership of the value of the normalized first error magnitude signal in the first plurality of fuzzy sets comprises:

determining a degree of membership in a first fuzzy set based on the value of the normalized first error magnitude signal;

determining a degree of membership in a second fuzzy set based on the first degree of membership; and

determining a degree of membership in a third fuzzy set based on the first degree of membership.

29. A method as defined in claim 28, wherein determining the degree of membership in the second fuzzy set comprises determining the degree of membership in the second fuzzy set as a value indicative of full membership minus the degree of membership in the first fuzzy set if the value of the normalized first error magnitude signal is less than a first threshold;

and

wherein determining the degree of membership in the third fuzzy set comprises determining the degree of membership in the third fuzzy set as the value indicative of full membership minus the degree of membership in the first fuzzy set if the value of the normalized first error magnitude signal is greater than a second threshold.

30. A method as defined in claim 29, wherein determining the degree of membership in the second fuzzy set comprises determining the degree of membership in the second fuzzy set as zero if the value of the normalized first error magnitude signal is greater than the first threshold.

31. A method as defined in claim 29, wherein determining the degree of membership in the third fuzzy set comprises determining the degree of membership in the third fuzzy set as zero if the value of the normalized first error magnitude signal is less than the second threshold.

32. A method as defined in claim 26, wherein determining degrees of membership of the value of the normalized first error magnitude signal in the first plurality of fuzzy sets comprises:

determining full membership in a first fuzzy set if the value of the normalized first error magnitude signal is less than a first threshold;

determining less than full membership in the first fuzzy set if the value of the normalized first error magnitude signal is greater than the first threshold;

determining full membership in a second fuzzy set if the value of the normalized first error magnitude signal is greater than a second threshold and less than a third threshold;

determining less than full membership in the second fuzzy set if the value of the normalized first error magnitude signal is less than the second threshold;

determining less than full membership in the second fuzzy set if the value of the normalized first error magnitude signal is greater than the third threshold;

determining full membership in a third fuzzy set if the value of the normalized first error magnitude signal is greater than a fourth threshold; and

determining less than full membership in the third fuzzy set if the value of the normalized first error magnitude signal is less than the fourth threshold.

33. A method as defined in claim 32, wherein determining less than full membership in the first fuzzy set comprises determining no membership in the first fuzzy set if the value of the normalized first error magnitude signal is greater than the second threshold;

wherein determining less than full membership in the second fuzzy set if the value of the normalized first error magnitude signal is less than the second threshold comprises determining no membership in the second fuzzy set if the value of the normalized first error magnitude signal is less than the first threshold;

wherein determining less than full membership in the second fuzzy set if the value of the normalized first error magnitude signal is greater than the third threshold comprises determining no membership in the second fuzzy set if the value of the normalized first error magnitude signal is greater than the fourth threshold;

wherein determining less than full membership in the third fuzzy set comprises determining no membership in the third fuzzy set if the value of the normalized first error magnitude signal is less than the third threshold.

34. A method as defined in claim 33, wherein the first threshold is 0.5;

wherein the second threshold is 1.0;

wherein the third threshold is 1.5; and

wherein the fourth threshold is 2.0.

35. A method as defined in claim 26, wherein generating the estimate of the parameter comprises:

determining weighted rule outputs by multiplying consequent portions of the fuzzy logic rules in the fuzzy rulebase by corresponding degrees of fulfillment of the fuzzy logic rules;

summing the weighted rule outputs;

summing the degrees of fulfillment of the fuzzy logic rules; and

dividing the summed weighted rule outputs by the summed degrees of fulfillment.

36. A method as defined in claim 26, wherein the first plurality of fuzzy sets comprises a Small set, a Medium set, and a Large set, and wherein the second plurality of fuzzy sets comprises the Small set, the Medium set, and the Large set;

wherein A is a value of the first sensed signal corresponding to the value of the normalized first error magnitude signal, B is a value of the first sensed signal corresponding to the value of the normalized second error magnitude signal, and R is a value of the reference signal corresponding to the value of the normalized first error magnitude signal and corresponding to the value of the normalized second error magnitude signal;

wherein evaluating the fuzzy rulebase comprises evaluating at least some of the following fuzzy logic rules;

1) IF EAR_NORM is Small AND EBR_NORM is Small OR IF EAR_NORM is Medium and EBR_NORM is Medium THEN X_A is the average;

of A, B, and R;

2) IF EAR_NORM is Small AND EBR_NORM is Medium OR IF EAR_NORM is Medium and EBR_NORM is Large THEN X_B is the average of A and R;

3) IF EAR_NORM is Medium AND EBR_NORM is Small OR IF EAR_NORM is Large and EBR_NORM is Medium THEN X_C is the average of B and R;

4) IF EAR_NORM is Small AND EBR_NORM is Large THEN X_D is A;

5) IF EAR_NORM is Large AND EBR_NORM is Small THEN X_E is B;

6) IF EAR_NORM is Large AND EBR_NORM is Large THEN X_F is R;

wherein EAR_NORM is the value of the normalized first error magnitude signal and EBR_NORM is the value of the normalized second error magnitude signal, and wherein X_A, X_B, X_C, X_D, X_E, and X_F are corresponding outputs of the fuzzy logic rules.

37. A method as defined in claim 36, wherein evaluating the fuzzy rulebase further comprises:

generating a weighted X_A based on X_A and a degree of fulfillment of the rule corresponding to X_A;

generating a weighted X_B based on X_B and a degree of fulfillment of the rule corresponding to X_B;

generating a weighted X_C based on X_C and a degree of fulfillment of the rule corresponding to X_C;

generating a weighted X_D based on X_D and a degree of fulfillment of the rule corresponding to X_D;

generating a weighted X_E based on X_E and a degree of fulfillment of the rule corresponding to X_E;

generating a weighted X_F based on X_F and a degree of fulfillment of the rule corresponding to X_F;

summing the weighted X_A, the weighted X_B, the weighted X_C, the weighted X_D, the weighted X_E, and the weighted X_F;

summing the degree of fulfillment of the rule corresponding to X_A, the degree of fulfillment of the rule corresponding to X_B the degree of fulfillment of the rule corresponding to X_C, the degree of fulfillment of the rule corresponding to X_D, the degree of fulfillment of the rule corresponding to XE, and the degree of fulfillment of the rule corresponding to X_F;

dividing the sum of the weighted X_A, the weighted X_B, the weighted X_C, the weighted X_D, the weighted X_E, and the weighted X_F by the sum of the degree of fulfillment of the rule corresponding to X_A, the degree of fulfillment of the rule corresponding to X_B, the degree of fulfillment of the rule corresponding to X_C, the degree of fulfillment of the rule corresponding to X_D, the degree of fulfillment of the rule corresponding to X_E, and the degree of fulfillment of the rule corresponding to X_F.

38. A method as defined in claim 24, wherein the first accuracy equals the second accuracy.

39. A method as defined in claim 24, wherein the reference signal comprises a model output.

40. An apparatus for estimating a parameter based on signals received from redundant sensors, the apparatus comprising:

an error magnitude calculator to generate a first error magnitude signal and a second error magnitude signal in response to a first sensed signal received from a first redundant sensor, a second sensed signal received from a second redundant sensor, and a reference signal, the first sensed signal, the second sensed signal, and the reference signal indicative of the parameter;

a normalizer to generate a first normalized error magnitude signal and a second normalized error magnitude signal in response to the first error magnitude signal and the second error magnitude signal, and based on a first accuracy associated with the first sensed signal, a second accuracy associated with the second sensed signal, and a reference signal accuracy associated with the reference signal;

a fuzzy logic estimate calculator to generate an estimate of the parameter in response to the first normalized error magnitude signal and the second normalized error magnitude signal, and according to a fuzzy rulebase.

41. An apparatus according to claim 40, wherein the normalizer generates the first normalized error magnitude signal as the first error magnitude signal divided by the reference signal accuracy added with the first accuracy;

and

wherein the normalizer generates the second normalized error magnitude signal as the second error magnitude signal divided by the reference signal accuracy added with the second accuracy.

42. An apparatus according to claim 41, wherein the fuzzy logic estimate calculator comprises a fuzzy membership evaluator to generate first degrees of membership in a first plurality of fuzzy sets for the first normalized error magnitude signal and to generate second degrees of membership in a second plurality of fuzzy sets for the second normalized error magnitude signal.

43. An apparatus according to claim 42, wherein the first plurality of fuzzy sets is the same as the second plurality of fuzzy sets.

44. An apparatus according to claim 42, wherein the fuzzy logic estimate calculator further comprises a rule fulfillment evaluator coupled to the fuzzy membership evaluator to generate degrees of fulfillment of a plurality of fuzzy rules based on the first degrees of membership and the second degrees of membership.

45. An apparatus according to claim 44, wherein the fuzzy logic estimate calculator further comprises a rule output synthesizer coupled to the rule fulfillment evaluator to generate the estimate of the parameter based on the degrees of fulfillment of the plurality of fuzzy rules.

46. An apparatus according to claim 40, wherein the first accuracy is the same as the second accuracy.

47. An apparatus according to claim 40, comprising a processor operatively coupled to a memory, the first redundant sensor, and the second redundant sensor, the memory having stored thereon:

first code to implement the error magnitude calculator;

second code to implement the normalizer; and

third code to implement the fuzzy logic estimate calculator.

48. An engine control system, comprising:

a first redundant sensor to generate a first sensed signal indicative of a parameter associated with an aircraft engine, the first sensed signal having a first accuracy associated therewith;

a second redundant sensor to generate a second sensed signal indicative of the parameter, the second sensed signal having a second accuracy associated therewith;

a reference signal generator to generate a reference signal indicative of the parameter, the reference signal having a reference signal accuracy associated therewith;

a weighting calculator operatively coupled to the first redundant sensor, the second redundant sensor, and the reference signal generator to generate weighting information in response to the first sensed signal, the second sensed signal, and the redundant signal, the weighting information based on the first accuracy, the second accuracy, and the reference signal accuracy;

a weighted average calculator to generate an estimate of the parameter in response to the first sensed signal, the second sensed signal, the redundant signal, and the weighting information.

49. An engine control system according to claim 48, wherein the engine control system comprises an aircraft engine control system.

50. An engine control system, comprising:

a first redundant sensor to generate a first sensed signal indicative of a parameter associated with an aircraft engine, the first sensed signal having a first accuracy associated therewith;

a second redundant sensor to generate a second sensed signal indicative of the parameter, the second sensed signal having a second accuracy associated therewith;

a reference signal generator to generate a reference signal indicative of the parameter, the reference signal having a reference signal accuracy associated therewith;

an error magnitude calculator operatively coupled to the first redundant sensor, the second redundant sensor, and the reference signal generator to generate a first error magnitude signal and a second error magnitude signal in response to the first sensed signal, the second sensed signal, and the reference signal;

a normalizer to generate a first normalized error magnitude signal and a second normalized error magnitude signal in response to the first error magnitude signal and the second error magnitude signal and based on the first accuracy;

, the second accuracy, and the reference signal accuracy;

a fuzzy logic estimate calculator to generate an estimate of the parameter in response to the first normalized error magnitude signal and the second normalized error magnitude signal, and according to a fuzzy rulebase.

51. An engine control system according to claim 50, wherein the engine control system comprises an aircraft engine control system.