Detecting Device, Detecting Method, and Program

• US 20160327417A1
• Filed: 08/29/2014
• Published: 11/10/2016
• Est. Priority Date: 09/06/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hara, Satoshi, Morimura, Tetsuro, Takahashi, Toshihiro
• Granted Patent: US 10,584,986 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: B60W 2400/00 Indexing codes relating to detected, measured or calculated conditions or factors
• CPC Class Codes: G01D 18/00 Testing or calibrating apparatus or arrangements provided for in groups G01D1/00 - G01D15/00
• CPC Class Codes: G05B 23/0208 characterized by the configuration of the monitoring system
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G05B 23/0235 based on a comparison with predetermined threshold or range, e.g. "classical methods", carried out during normal operation; threshold adaptation or choice; when or how to compare with the threshold

Abstract

Embodiments of the present invention specify a group of sensors detecting normal operation and a group of sensors detecting abnormal operation from time-series data on the sensors without using threshold values that are based on the experience of users. One aspect of an embodiment is a detecting device for detecting changes in the output of a plurality of sensors, in which the detecting device includes: a first output acquiring unit for acquiring a first relevance matrix; a second output acquiring unit for acquiring a second relevance matrix; a change calculating unit for calculating a change matrix representing the degree of change between the first relevance matrix and the second relevance matrix; and a specifying unit for specifying a group of sensors showing a degree of change greater than the others in the change matrix. Other aspects of the present invention include a detecting method and a program.

Claims

1. A device comprising:

a first output acquiring unit for acquiring a first relevance matrix representing the degree of relevance between a plurality of sensors during a first output from the plurality of sensors;

a second output acquiring unit for acquiring a second relevance matrix representing the degree of relevance between a plurality of sensors during a second output from the plurality of sensors;

a change calculating unit for calculating a change matrix representing the degree of change between the first relevance matrix and the second relevance matrix; and

a specifying unit for specifying, as a group of changing sensors showing a change between the first output and the second output, a group of sensors from the plurality of sensors showing a degree of change greater than the remaining sensors of the plurality of sensors in the change matrix.

2. The device according to claim 1, wherein the specifying unit specifies a submatrix of sensors from the plurality of sensors having a degree of changes smaller than the other sensors from the change matrix, and specifies the group of sensors not corresponding to the submatrix as the group of changing sensors.

3. The device according to claim 1, wherein the specifying unit further comprises:

a score calculating unit for calculating a score for each sensor of the plurality of sensors to minimize an objective function corresponding to a value obtained by multiplying scores for the sensors corresponding to their degree of change by each of the degrees of change included as elements in the change matrix; and

a sensor specifying unit for specifying a group of unchanging sensors not showing a change between the first output and the second output, and a group of changing sensors on the basis of the scores of the sensors.

4. The device according to claim 3, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on condition that the score of each sensor is equal to or greater than 0, and the sensor specifying unit includes sensors with a score of 0 in the group of changing sensors excluding calculation errors.

5. The device according to claim 4, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that a sum of the scores of the sensors is a predetermined value.

6. The device according to claim 5, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that a sum of the scores of the sensors is equal to the number of sensors in the plurality of sensors.

7. The device according to claim 5, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that the sum of the scores of the sensors is equal to a predetermined number of unchanging sensors.

8. The device according to claim 7, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that the score of each sensor is from 0 to 1 inclusive.

9. The device according to claim 7, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that the score of each sensor is 0 or 1.

10. The device according to claim 7, wherein the score calculating unit calculates scores for the plurality of sensors to minimize the objective function on further condition that the sum of the scores of the sensors equals a number of unchanging sensors for each of a different number of unchanging sensors, and

the sensor specifying unit specifies the group of unchanging sensors and the group of changing sensors on the basis of the scores of the sensors for each of a different number of unchanging sensors.

11. The device according to claim 5, wherein the score calculating unit calculates a first score value for the plurality of sensors to minimize the objective function on condition that the sum of the scores of the sensors equals the number of sensors,

the score calculating unit specifies a first group of unchanging sensors and a first group of changing sensors by including the sensors with a first score of 0 in the group of changing sensors excluding calculation errors,

the score calculating unit calculates a second score value for the plurality of sensors to minimize the objective function on condition that the sum of the scores of the sensors equals the number of sensors in the first group of unchanging sensors,

the score calculating unit specifies a second group of unchanging sensors and a second group of changing sensors by including the sensors with a second score of 0 in the group of changing sensors excluding calculation errors, and

a determining unit is further provided for determining the matching sensors in at least either the first group of unchanging sensors and the second group of unchanging sensors, or the first group of changing sensors and the second group of changing sensors.

12. The device according to claim 1, wherein the change calculating unit calculates the change matrix on the basis of the absolute values of the differences between the first relevance matrix and the second relevance matrix.

13. The device according to claim 12, wherein the change calculating unit calculates the change matrix by adding a unit matrix to the absolute values of the differences between the first relevance matrix and the second relevance matrix as a constant multiple.

14. The device according to claim 13, wherein the change calculating unit determines the multiplication factor of the constant multiple so the smallest eigenvalue in the change matrix is positive.

15. The device according to claim 1, wherein the change calculating unit further comprises a normalizing unit for normalizing the elements of one diagonal in the first relevance matrix and the second relevant matrix to 1, and normalizing the first relevance matrix to the second relevance matrix.

16. The device according to claim 1, wherein the first output acquiring unit acquires the first relevance matrix corresponding to the first output outputted by the sensors, and the second output acquiring unit acquires the second relevance matrix corresponding to the output outputted by the sensors during an inspection period for an object when the object provided with the sensors is in a normal state.

17. A method comprising:

acquiring a first relevance matrix representing a degree of relevance between a plurality of sensors during a first output from the plurality of sensors;

acquiring a second relevance matrix representing the degree of relevance between a plurality of sensors during a second output from the plurality of sensors;

calculating a change matrix representing the degree of change between the first relevance matrix and the second relevance matrix; and

specifying, as a group of changing sensors showing a change between the first output and the second output, a group of sensors from the plurality of sensors showing a degree of change greater than the remaining sensors of the plurality of sensors in the change matrix.

18. (canceled)

19. A computer program product comprising:

a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions readable by a processor system to cause the processor system to perform a method comprising;

acquiring a first relevance matrix representing a degree of relevance between a plurality of sensors during a first output from the plurality of sensors;

acquiring a second relevance matrix representing the degree of relevance between a plurality of sensors during a second output from the plurality of sensors;

calculating a change matrix representing the degree of change between the first relevance matrix and the second relevance matrix; and

specifying, as a group of changing sensors showing a change between the first output and the second output, a group of sensors from the plurality of sensors showing a degree of change greater than the remaining sensors of the plurality of sensors in the change matrix.