METHOD AND SYSTEM FOR PROVIDING QUALITY CONTROLLED DATA FROM A REDUNDANT SENSOR SYSTEM

• US 20200133250A1
• Filed: 10/30/2018
• Published: 04/30/2020
• Est. Priority Date: 10/30/2018
• Status: Active Grant

• Current Assignee: University Corporation For Atmospheric Research
• Original Assignee: University Corporation For Atmospheric Research
• Inventors: GAYDOS, Andrew, Goodrich, Robert Kent, Landolt, Scott
• Granted Patent: US 10,996,663 B2
• Time in Patent Office: Days
• CPC Class Codes: G01W 1/18 Testing or calibrating meteorological apparatus
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
• CPC Class Codes: G06F 17/15 Correlation function computation including computation of convolution operations arithmetic circuits for sum of products per se, e.g. multiply-accumulators G06F7/5443; digital filters, e.g. FIR, IIR, adaptive filters H03H17/00
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06F 18/251 of input or preprocessed data
• CPC Class Codes: G06F 2218/12 Classification; Matching

Abstract

A method for determining a quality controlled sensor set from a redundant sensor set comprising calculating a first time correlation coefficient and a first autocorrelation coefficient based on a first sensor time series data, calculating a second time correlation coefficient and a second autocorrelation coefficient based on a second sensor time series data, calculating a first and second sensor correlation coefficient based on the first sensor time series data and the second sensor time series data, and determining the quality controlled sensor set with a highest confidence level.

Claims

1. A method for determining a quality controlled sensor set from a redundant sensor set, the method comprising:

calculating a first time correlation coefficient (timec1) and a first autocorrelation coefficient (autoc1) based on a first sensor time series data;

calculating a second time correlation coefficient (timec2) and a second autocorrelation coefficient (autoc2) based on a second sensor time series data;

calculating a first and second sensor correlation coefficient (crossc1,2) based on the first sensor time series data and the second sensor time series data;

determining the quality controlled sensor set with a highest confidence level based on the timec1, the autoc1, the timec2, the autoc2, and the crossc1,2.

2. The method of claim 1, wherein determining the quality controlled sensor set with the highest confidence level further comprises including the first sensor and the second sensor in the quality controlled sensor set upon determining that a first and second sensor confidence (conf1,2) is over a confidence threshold, wherein the

3. The method of claim 1, wherein determining the quality controlled sensor set with the highest confidence level further comprises:

determining a first sensor confidence (conf1) based on the timec1 and the autoc1;

determining a second sensor confidence (conf2) based on the timec2 and the autoc2; and

including the first sensor in the quality controlled sensor set upon determining that the conf1 is greater than the conf2.

4. The method of claim 1, further comprising:

calculating a third time-correlation coefficient (timec3) and a third autocorrelation coefficient (autoc3) based on a third sensor time series data;

calculating a second and third sensor correlation coefficient (crossc2,3) based on the second sensor time series data and the third sensor time series data; and

calculating a first and third sensor correlation coefficient (crossc1,3) based on the first sensor time series data and the third sensor time series data,

wherein determining the quality controlled sensor set with the highest confidence level is further based on the timec3, autoc3, and crossc1,3.

5. The method of claim 4, wherein determining the quality controlled sensor set with the highest confidence level further comprises including the first sensor, the second sensor, and the third sensor in the quality controlled sensor set upon determining that a first, second, and third sensor confidence (conf1,2,3) is over a confidence threshold and the third in the quality controlled sensor set, wherein the

6. The method of claim 4, further comprising:

calculating a third time-correlation coefficient (timec3) and a third autocorrelation coefficient (autoc3) based on a third sensor time series data; and

calculating a second and third sensor correlation coefficient (crossc2,3) based on the second sensor time series data and the third sensor time series data,

wherein determining the quality controlled sensor set with the highest confidence level further comprises including the third sensor in the quality controlled sensor set upon determining that a second and third sensor confidence (conf2,3) is over the confidence threshold, wherein the

7. The method of claim 1, further comprising:

determining a best-fit curve to the first sensor time series data to determine a first sensor value estimation.

8. The method of claim 4, wherein determining the quality controlled sensor set with the highest confidence level further comprises:

determining a first sensor value estimation from the first sensor time series data;

determining a second sensor value estimation from the second sensor time series data;

determining a third sensor value estimation from the third sensor time series data;

upon determining that the first sensor value estimation minus the second sensor value estimation is greater than a delta sensor value threshold, and the first sensor value estimation minus the third sensor value estimation is also greater than the delta sensor value threshold, setting a first sensor confidence (conf1) to zero.

9. The method of claim 1, further comprising:

determining a first sensor value estimation;

upon determining that the first sensor value estimation is less than a minimum value threshold, setting the first sensor value estimation to zero and setting a first sensor confidence (conf1) to 1.

10. The method of claim 1, further comprising:

determining a first sensor value estimation; and

upon determining that the first sensor value estimation is greater than a maximum value threshold, setting a first sensor confidence (conf1) to 0.

11. The method of claim 1, further comprising:

determining a quality controlled system measurement value based on an average of two or more individual sensor value estimations for the quality controlled sensor set.

12. The method of claim 1, further comprising:

determining a non-zero first sensor value estimation for a time within a persistence period; and

upon determining that there are no other non-zero values for the first sensor during the persistence period, setting the first sensor value estimation to zero.

13. A system for determining a quality controlled sensor set from a redundant sensor set, the system comprising:

a time correlation calculation module configured to calculate a first time correlation coefficient (timec1) and a first autocorrelation coefficient (autoc1) based on a first sensor time series data, and calculate a second time correlation coefficient (timec2) and a second autocorrelation coefficient (autoc2) based on a second sensor time series data;

a sensor correlation calculation module configured to calculate a first and second sensor correlation coefficient (crossc1,2) based on the first sensor time series data and the second sensor time series data; and

a quality controlled sensor determination module configured to determine the quality controlled sensor set with a highest confidence level based on the timec1, the autoc1, the timec2, the autoc2, the crossc1,2, and the crossc2,3.

14. The system of claim 13, wherein the quality controlled sensor determination module is further configured to determine the quality controlled sensor set with the highest confidence level by including the first sensor and the second sensor in the quality controlled sensor set upon determining that a first and second sensor confidence (conf1,2) is over a confidence threshold, wherein the

15. The system of claim 13, wherein the quality controlled sensor determination module is further configured to determine the quality controlled sensor set with the highest confidence level further by determining a first sensor confidence (conf1) based on the timec1 and the autoc1, determining a second sensor confidence (conf2) based on the timec2 and the autoc2, and including the first sensor in the quality controlled sensor set upon determining that the conf1 is greater than the conf2.

16. The system of claim 13, wherein the time correlation calculation module is further configured to calculate a third time-correlation coefficient (timec3) and a third autocorrelation coefficient (autoc3) based on a third sensor time series data;

the sensor correlation calculation module is further configured to calculate a second and third sensor correlation coefficient (crossc2,3) based on the second sensor time series data and the third sensor time series data, and calculate a first and third sensor correlation coefficient (crossc1,3) based on the first sensor time series data and the third sensor time series data; and

the quality controlled sensor determination module is further configured to determine the quality controlled sensor set with the highest confidence level based on the timec3, autoc3, and crossc1,3.

17. The system of claim 16, wherein the quality controlled sensor determination module is further configured to include the first sensor, the second sensor, and the third sensor in the quality controlled sensor set upon determining that a first, second, and third sensor confidence (conf1,2,3) is over a confidence threshold and the third in the quality controlled sensor set, wherein the

18. The system of claim 16, wherein the time correlation calculation module is further configured to calculate a third time-correlation coefficient (timec3) and a third autocorrelation coefficient (autoc3) based on a third sensor time series data;

the sensor correlation calculation module is further configured to calculate a second and third sensor correlation coefficient (crossc2,3) based on the second sensor time series data and the third sensor time series data, and the quality controlled sensor set determination module is further configured to include the third sensor in the quality controlled sensor set upon determining that a second and third sensor confidence (conf2,3) is over the confidence threshold, wherein the

19. The system of claim 13, further comprising:

a quality controlled value determination module configured to determine a quality controlled system measurement value based on an average of two or more individual sensor value estimations for the quality controlled sensor set.

20. The system of claim 13, further comprising:

a value persistence module configured to determine a non-zero first sensor value estimation for a time within a persistence period, and upon determining that there are no other non-zero values for the first sensor during the persistence period, set the first sensor value estimation to zero.