SYSTEM AND METHOD FOR EFFICIENT DATA COLLECTION IN DISTRIBUTED SENSOR MEASUREMENT SYSTEMS

• US 20140019397A1
• Filed: 07/12/2012
• Published: 01/16/2014
• Est. Priority Date: 07/12/2012
• Status: Active Grant

• Current Assignee: Eaton Intelligent Power Limited (Eaton Corporation PLC.)
• Original Assignee: Eaton Intelligent Power Limited (Eaton Corporation PLC.)
• Inventors: Alexander, Roger K.
• Granted Patent: US 10,679,131 B2
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: H04W 4/029 Location-based management or tracking services
• CPC Class Codes: H04W 4/38 for collecting sensor information

Abstract

Endpoint device, central data collection point, and associated methods for collecting data over a communication network between endpoints and the central collection point. Actual measurements from a sensor are obtained by the endpoint device at a relatively fine time granularity. The endpoint device generates reports for receipt by a central data collection point. The reports include regular reports containing a portion of the actual measurements representing sensor measurements at a relatively coarse time granularity, and exception reports, containing information representing one or more of the actual measurements that differ in frequency or granularity of regular report measurements. Each of the exception reports is generated in response to a determination that at least one of the actual measurements differs from a predicted value for that at least one of the one or more actual measurements by an amount that exceeds a pre-established limit.

Claims

1. A method for operating a sensor data collection system for collecting data from a plurality of endpoint devices, the method comprising:

receiving, at a central data collection point, sensor data generated at each of the plurality of endpoint devices, the sensor data representing a series of actual measured values;

maintaining, at the central data collection point, a database containing records corresponding to the individual endpoint devices and including historic values of sensor data for those endpoint devices;

computing, at the central data collection point, predicted values representing presumed sensor data for at least one endpoint device, wherein the predicted values are computed based on the historic values of sensor data corresponding to the at least one endpoint device;

receiving, at the central data collection point, exception data generated by the at least one endpoint device, the exception data representing a discrepancy between at least one of the predicted values and a corresponding at least one actual measured sensor data value; and

in response to receiving the exception data, superseding, at the central data collection point, the at least one of the predicted values based on the exception data.

2. The method of claim 1, wherein the sensor data represents utility consumption.

3. The method of claim 1, wherein the series of actual measured values of the sensor data is measured at a first time granularity and the predicted values representing the presumed sensor data represent sensor data measured at a second time granularity, the second time granularity being a finer granularity than the first time granularity.

4. The method of claim 1, further comprising:

storing, at the central data collection point, actual measured values of sensor data received from a first endpoint device and predicted values corresponding to the first endpoint device in a common database record associated with the first endpoint device.

5. The method of claim 1, wherein computing the predicted values is performed in response to a call for instantaneous measurement.

6. The method of claim 1, wherein computing the predicted values is regularly performed at a condensed time interval that is shorter than a reporting interval at which the sensor data is regularly reported by each of the plurality of endpoint devices.

7. The method of claim 1, wherein the central data collection point computes the predicted values using a first computation algorithm;

and

wherein in receiving the exception data, the central data collection point obtains information about the discrepancy based on a prediction computed by the at least one endpoint device using a second computation algorithm that produces a substantially same result as the first computation algorithm and is performed based on the same historic values of sensor data corresponding to each of the at least one endpoint device.

8. The method of claim 1, wherein the central data collection point computes the predicted values using a first computation algorithm;

and further comprising;

updating, at the central data collection point, the first computation algorithm to a new computation algorithm and transmitting instructions to the at least one endpoint device to apply the new computation algorithm.

9. The method of claim 1, wherein computing the predicted values is performed for a large plurality of endpoint devices within a time window that is smaller than a regular reporting period of any one of the endpoint devices.

10. The method of claim 1, wherein in receiving the exception data, the central data collection point obtains information about the discrepancy only in response to a result of a comparison performed remotely from the central data collection point between predicted values and actual measured values exceeding limits of pre-established comparison criteria.

11. The method of claim 1, wherein in receiving the exception data, the central data collection point obtains information about the discrepancy based on a comparison between predicted values computed by the at least one endpoint device and actual measured values as measured by that at least one endpoint device exceeding limits of pre-established comparison criteria.

12. The method of claim 1, wherein receiving the exception data includes receiving a report containing the one or more actual measured sensor data values based upon which the discrepancy was determined by the at least one endpoint device.

13. The method of claim 1, wherein superseding the at least one of the predicted values includes replacing that at least one predicted value with actual measured sensor data.

14. The method of claim 1, wherein the sensor data represents utility consumption and further comprising:

generating, at the central data collection point, a determination of a current system-wide instantaneous consumption based on the actual measured values and on the predicted values for the plurality of endpoint devices.

15. A central data collection point for use in a sensor data collection system for collecting data from a plurality of endpoint devices, comprising:

a collection engine module configured to receive sensor data generated at each of the plurality of endpoint devices, the sensor data representing a series of actual measured values;

a database containing records corresponding to the individual endpoint devices and including historic values of sensor data for those endpoint devices;

a prediction module configured to compute predicted values representing presumed sensor data for at least one endpoint device, wherein the predicted values are computed based on the historic values of sensor data corresponding to the at least one endpoint device;

a data selector module configured to process received exception data generated by the at least one endpoint device, the exception data representing a discrepancy between at least one of the predicted values and a corresponding at least one actual measured sensor data value;

wherein the data selector module is further configured such that, in response to receiving the exception data, the data selector module supersedes the at least one of the predicted values based on the exception data.

16. The central data collection point of claim 15, wherein the sensor data represents utility consumption.

17. The central data collection point of claim 15, wherein the series of actual measured values of the sensor data is measured at a first time granularity and the predicted values representing the presumed sensor data represent sensor data measured at a second time granularity, the second time granularity being a finer granularity than the first time granularity.

18. The central data collection point of claim 15, wherein the database is configured to store actual measured values of sensor data received from a first endpoint device and predicted values corresponding to the first endpoint device in a common database record associated with the first endpoint device.

19. The central data collection point of claim 15, wherein the prediction module is configured to compute the predicted values in response to a call for instantaneous measurement.

20. The central data collection point of claim 15, wherein the prediction module is configured to compute the predicted values at a condensed time interval that is shorter than a reporting interval at which the sensor data is regularly reported by each of the plurality of endpoint devices.

21. The central data collection point of claim 15, wherein the prediction module is configured to compute the predicted values using a first computation algorithm;

and further comprising;

a re-configuration module configured to update the first computation algorithm to a new computation algorithm and generate instructions to the at least one endpoint device to apply the new computation algorithm.

22. The central data collection point of claim 15, wherein the prediction module is configured to compute the predicted values for a large plurality of endpoint devices within a time window that is smaller than a regular reporting period of any one of the endpoint devices.

23. An endpoint device for use with a sensor data collection system for collecting data from a large plurality of sensors, the endpoint device comprising:

a sensor data input module configured to obtain actual measurements from a sensor at a relatively fine time granularity;

a reporting module operatively coupled to the sensor data input module and configured to generate reports for receipt by a central data collection point, the reports including;

regular reports containing a portion of the actual measurements representing sensor measurements at a relatively coarse time granularity; and

exception reports, each exception report containing information representing one or more of the actual measurements that differ in time granularity from the coarse granularity of the regular reports, each of the exception reports being generated in response to a determination that at least one of the one or more actual measurements differs from a predicted value for that at least one of the one or more actual measurements by an amount that exceeds a pre-established limit.

24. The endpoint device of claim 23, further comprising:

a prediction module operatively coupled to the reporting module and configured to compute predicted values representing presumed sensor data at a relatively fine time granularity, the predicted values being computed based on previously-reported actual measurements.

25. The endpoint device of claim 24, wherein the prediction module is configured to compute the predicted values using a computation algorithm that corresponds to a computation algorithm for predicting the same value at the central data collection point.

26. The endpoint device of claim 23, further comprising:

a comparator module operatively coupled to the sensor data input module and to the reporting module, and configured to compare the at least one of the one or more actual measurements with a predicted value for that at least one actual measurement, and to perform the determination of whether a difference between the at least one actual measurement and predicted value corresponding to that actual measurement exceeds the pre-established limit.

27. The endpoint device of claim 23, further comprising:

a data store operatively coupled to the sensor data input module and configured to maintain a historical record of at least a portion of the actual measurement data.

28. The endpoint device of claim 27, wherein the data store is configured to store primarily information relating to actual measurement data to be used in computing a next predicted value.

29. The endpoint device of claim 23, further comprising:

a utility meter interface operatively coupled to the sensor data input and wherein the sensor comprises a utility meter.

30. The endpoint device of claim 23, wherein the reporting module is configured to generate exception reports containing a series of measurements at a time granularity that is finer than the relatively coarse time granularity.

31. A method for operating an endpoint device in a sensor data collection system for collecting data from a large plurality of sensors, the method comprising:

obtaining actual measurements from a sensor at a relatively fine time granularity; and

generating reports for receipt by a central data collection point, the reports including;

regular reports containing a portion of the actual measurements representing sensor measurements at a relatively coarse time granularity; and

exception reports, each exception report containing information representing one or more of the actual measurements that differ in time granularity from the coarse granularity of the regular reports, each of the exception reports being generated in response to a determination that at least one of the one or more actual measurements differs from a predicted value for that at least one of the one or more actual measurements by an amount that exceeds a pre-established limit.

32. The method of claim 31, further comprising:

computing predicted values representing presumed sensor data at a relatively fine time granularity, the predicted values being computed based on previously-reported actual measurements.

33. The method of claim 32, wherein in computing the predicted values, a computation algorithm that corresponds to a computation algorithm for predicting the same value at the central data collection point is used.

34. The method of claim 31, further comprising:

comparing, by the endpoint device, the at least one of the one or more actual measurements with a predicted value for that at least one actual measurement; and

performing, by the endpoint device, the determination of whether a difference between the at least one actual measurement and predicted value corresponding to that actual measurement exceeds the pre-established limit.

35. The method of claim 31, further comprising:

maintaining a historical record of at least a portion of the actual measurement data.

36. The method of claim 35, wherein in maintaining the historical record, storing primarily information relating to actual measurement data to be used in computing a next predicted value.

37. The method of claim 31, wherein in generating the exception reports, providing in each of the exception reports a series of measurements at a time granularity that is finer than the relatively coarse time granularity.