PREDICTIVE PEER-BASED TIRE HEALTH MONITORING

• US 20140277910A1
• Filed: 03/14/2013
• Published: 09/18/2014
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Goodyear Tire & Rubber Company
• Original Assignee: Goodyear Tire & Rubber Company
• Inventors: Suh, Peter J., Xu, Su
• Granted Patent: US 9,079,461 B2
• Time in Patent Office: Days
• US Class Current: 701/31.4
• CPC Class Codes: B60C 23/02 Signalling devices actuated by tyre pressure hand-held tyre pressure gauges G01L17/00
• CPC Class Codes: B60C 23/0445 Means for changing operating mode, e.g. sleep mode, factory mode or energy saving mode
• CPC Class Codes: B60C 23/0454 Means for changing operation mode, e.g. sleep mode, factory mode or energy save mode
• CPC Class Codes: B60C 23/0474 Measurement control, e.g. setting measurement rate or calibrating of sensors; Further processing of measured values, e.g. filtering, compensating or slope monitoring
• CPC Class Codes: B60C 23/0476 Temperature compensation of measured pressure values
• CPC Class Codes: B60C 23/0479 Communicating with external units being not part of the vehicle, e.g. tools for diagnostic, mobile phones, electronic keys or service stations
• CPC Class Codes: B60C 23/0481 System diagnostic, e.g. monitoring battery voltage, detecting hardware detachments or identifying wireless transmission failures

Abstract

A method, apparatus and program product determine a health condition for the tires on one or more vehicles by performing an individual self-comparison and a peer-based comparison for each tire, where the individual self-comparison reduces variances between the plurality of tires on the vehicle, and the peer-based comparison reduces variances within each tire. Anomalies, such as severe leakage and tire inflations may be detected, and slow leakages may be predicted based upon historical data associated with the health condition.

Claims

1. A method of monitoring vehicle tires, the method comprising:

receiving tire data associated with a plurality of tires on a vehicle; and

determining a health condition for each of the plurality of tires by performing an individual self-comparison and a peer-based comparison of the tire data for each tire, wherein the individual self-comparison reduces variances between the plurality of tires on the vehicle, and wherein the peer-based comparison reduces variances within each tire.

2. The method of claim 1, wherein the tire data includes temperature and pressure data for each of the plurality of tires and taken at a plurality of times.

3. The method of claim 2, wherein the tire data for a first tire among the plurality of tires and taken at a first time includes a first pressure, a first temperature and a first timestamp associated with the first time.

4. The method of claim 1, wherein performing the individual self-comparison includes applying an assessment model to the tire data.

5. The method of claim 4, wherein the assessment model comprises a distance-based assessment model.

6. The method of claim 4, wherein the assessment model comprises a logistic regression model or a self-organizing map model.

7. The method of claim 4, wherein performing the individual self-comparison further includes generating an individual health value (IHV) for each tire based upon the assessment model, wherein the IHV for each tire compares a current performance of such tire with a normal performance of such tire determined during training of the assessment model.

8. The method of claim 1, wherein performing the peer-based comparison includes comparing tire data for a first tire against that of at least a subset of the plurality of tires.

9. The method of claim 1, wherein comparing the tire data for the first tire against that of the at least a subset of the plurality of tires includes comparing the tire data for the first tire against a median determined from the tire data for at least a subset of the plurality of tires.

10. The method of claim 1, wherein comparing the tire data for the first tire against that of the at least a subset of the plurality of tires includes comparing the tire data for the first tire against a mean determined from the tire data for at least a subset of the plurality of tires.

11. The method of claim 1, wherein the received tire data includes, for each tire, a raw pressure and a temperature, wherein the method further comprises determining a compensated pressure for each tire by compensating the raw pressure based upon the temperature.

12. The method of claim 1, further comprising removing at least one out-of range outlier from the tire data.

13. The method of claim 1, further comprising:

compensating pressure data in the tire data based upon temperature data in the tire data; and

removing at least one impossible data combination outlier from the tire data based on the compensated pressure data.

14. The method of claim 1, further comprising:

training a tire-specific model for each tire using a first portion of the tire data to establish a normal operating condition for each tire; and

applying a second portion of the tire data to the tire-specific model for each tire to determine the health condition of each tire.

15. The method of claim 14, wherein training the tire-specific model for each tire includes determining a mean pressure value for each tire, and wherein applying the second portion of the tire data includes:

determining an individual health value (IHV) for each tire based upon individual-self comparison, wherein determining the IHV for each tire includes determining a difference between a compensated pressure value and the mean pressure value; and

determining a peer health value (PHV) for each tire based upon peer-based comparison, wherein determining the PHV for each tire includes determining a difference between the IHV for each tire and a median of IHV'"'"'s for at least a subset of the plurality of tires.

16. The method of claim 14, wherein training the tire-specific model is performed in a training mode for the tire-specific model and applying the second portion to the tire data is performed in a testing mode for the tire-specific model, the method further comprising, in response to detecting an anomaly for a tire, returning the tire-specific model to the training mode to retrain the tire-specific model for the tire.

17. The method of claim 16, wherein the anomaly is a tire inflation.

18. The method of claim 1, further comprising:

determining an anomaly detection feature; and

detecting an anomaly based on the anomaly detection feature.

19. The method of claim 18, wherein the anomaly detection feature is a standard deviation of peer health values (PHVs) or a mean of temporal differences.

20. The method of claim 18, further comprising determining a dynamic threshold based on a historical trend of the anomaly detection feature, wherein detecting the anomaly is based on the anomaly detection feature exceeding the dynamic threshold.

21. The method of claim 18, further comprising, in response to detecting an anomaly:

determining an anomaly diagnosis feature; and

determining an anomaly type based on the anomaly diagnosis feature.

22. The method of claim 21, wherein the anomaly diagnosis feature is a difference between individual health value (IHV) and a minimum or standard deviation.

23. The method of claim 21, further comprising determining a dynamic threshold based on a historical trend of the anomaly diagnosis feature, wherein determining the anomaly type is based a number of tire for which the anomaly diagnosis feature exceeds the dynamic threshold.

24. The method of claim 23, wherein the anomaly type is selected from the group consisting of a severe leakage, a single tire inflation, a subset of tires inflation, and an all-tires inflation, and wherein for the single tire inflation, a tire-specific model for the single tire is retrained, for the subset of tires inflation, tire-specific models for the subset of tires are retrained, and for the all-tires inflation, tire-specific models for all tires are retrained.

25. The method of claim 21, further comprising:

determining the anomaly detection feature excluding a tire with a minimum individual health value (IHV); and

detecting a severe leakage in response to a difference between the anomaly detection feature with the tire with the minimum IHV and the anomaly detection feature excluding the tire with the minimum IHV.

26. The method of claim 1, further comprising predicting leakage in a tire based on historical peer health values (PHVs) of the tire.

27. The method of claim 26, wherein predicting leakage includes performing at least one of linear, exponential, cubic spline and piecewise linear regression.

28. The method of claim 26, wherein predicting leakage includes performing piecewise linear regression to predict a future PHV.

29. The method of claim 26, wherein predicting leakage includes building a prediction model using Multivariate Adaptive Regression Splines (MARS) regression analysis to minimize root mean square error (RMSE) in the prediction model.

30. The method of claim 1, further comprising detecting a slow leakage in a tire based on peer health values (PHVs) of the tire over a plurality of iterations, wherein detecting the slow leakage includes, during each iteration:

comparing a PHV of the tire against a first threshold criterion;

accumulating a count for the tire if the PHV of the tire meets a first threshold criterion; and

resetting the count for the tire if the PHV of the tire does not meet the first threshold criterion.

31. The method of claim 30, wherein detecting the slow leakage further comprises:

comparing the count against a second threshold criterion;

determining an initial slow leak level from a PHV of the tire if the count meets the second threshold criterion; and

triggering a slow leak event in response to meeting a third threshold criterion relative to the initial slow leak level.

32. The method of claim 1, further comprising detecting recovery of a tire based on historical peer health values (PHVs) of the tire, wherein detecting recovery of the tire comprises:

comparing a current PHV of the tire against a threshold criterion; and

triggering a tire recovery event in response to the current PHV of the tire meeting the threshold criterion.

33. An apparatus, comprising:

at least one processor; and

program code configured upon execution by the at least one processor to monitor vehicle tires by receiving tire data associated with a plurality of tires on a vehicle, and determining a health condition for each of the plurality of tires by performing an individual self-comparison and a peer-based comparison of the tire data for each tire, wherein the individual self-comparison reduces variances between the plurality of tires on the vehicle, and wherein the peer-based comparison reduces variances within each tire.

34. The apparatus of claim 33, wherein the tire data is generated locally on the vehicle, and wherein the at least one processor is disposed in a remote monitoring service configured to receive the tire data over a network.

35. A program product, comprising:

a non-transitory computer readable medium; and

program code stored on the computer readable medium and configured upon execution to monitor vehicle tires by receiving tire data associated with a plurality of tires on a vehicle, and determining a health condition for each of the plurality of tires by performing an individual self-comparison and a peer-based comparison of the tire data for each tire, wherein the individual self-comparison reduces variances between the plurality of tires on the vehicle, and wherein the peer-based comparison reduces variances within each tire.