TECHNOLOGIES FOR PREDICTIVE MONITORING OF A CHARACTERISTIC OF A SYSTEM

• US 20180189659A1
• Filed: 12/29/2016
• Published: 07/05/2018
• Est. Priority Date: 12/29/2016
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Granted Patent: US 10,339,026 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 11/0709 in a distributed system consisting of a plurality of standalone computer nodes, e.g. clusters, client-server systems
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/3089 Monitoring arrangements determined by the means or processing involved in sensing the monitored data, e.g. interfaces, connectors, sensors, probes, agents software debugging using additional hardware using a specific debug interface G06F11/3656; performance evaluation by tracing or monitoring G06F11/3466
• CPC Class Codes: G06N 3/044 Recurrent networks, e.g. Hopfield networks

Abstract

Technologies for monitoring a characteristic of a monitored system include determining a measured value of the primary characteristic of the monitored system sensed by a primary sensor and a measured value of the secondary characteristic of the monitored system sensed by a secondary sensor, and predicating a predicted value of the primary characteristic based on the measured value of the secondary characteristic. The measured and predicted values of the primary characteristic are used to determine whether the primary sensor is properly functioning.

Claims

1. A predictive sensor module to monitor a characteristic of a monitored system, the predictive sensor module comprising:

a primary sensor to produce primary sensor data indicative of a primary characteristic of the monitored system;

one or more secondary sensors to produce secondary sensor data indicative of a secondary characteristic, different from the primary characteristic, of the monitored system; and

a sensor controller to (i) determine a measured value of the primary characteristic based on the primary sensor data, (ii) determine a measured value of the secondary characteristic based on the secondary sensor data, (iii) predict a predicted value of the primary characteristic using a predictive model with the measured value of the secondary characteristic as an input to the predictive model, and (iv) determine whether to update the predicted model based on a comparison of the measured value and the predicted value of the primary characteristic.

2. The predictive sensor module of claim 1, wherein to determine whether to update the predicted model comprises to determine a difference between the measured value of primary characteristic and the predicted value of primary characteristic.

3. The predictive sensor module of claim 2, wherein to determine whether to update the predicted model further comprises to determine whether the difference is within a first reference threshold.

4. The predictive sensor module of claim 3, wherein the sensor controller is further to transmit, in response to a determination that the difference is within the first reference threshold, a heartbeat signal to a model management system that indicates a predictive model managed by the model management system does not need updated.

5. The predictive sensor module of claim 3, wherein to determine whether to update the predicted model further comprises to determine whether the difference is within a second reference threshold in response to a determination that the difference is not within the first reference threshold.

6. The predictive sensor module of claim 5, wherein the sensor controller is further to update the predictive model in response to a determination that the difference is within the second reference threshold and not within first reference threshold, and

wherein to update the predictive model comprises to update a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.

7. The predictive sensor module of claim 5, wherein the sensor controller is further to perform, in response to a determination that the difference is not within the second reference threshold, a sensor operation analysis to determine whether the primary sensor is operating correctly.

8. The predictive sensor module of claim 7, wherein to perform the sensor operation analysis comprises to transmit the measured value of the primary characteristic, the measured value of secondary characteristic, and a plurality of model parameters to a model management system, wherein the plurality of model parameters comprise the predictive model used to predict the predicted value of the primary characteristic based on the measured value of the secondary characteristic.

9. The predictive sensor module of claim 7, wherein the sensor controller is further to update the predictive model in response to a determination that the difference is not within the second reference threshold, and

wherein to update the predictive model comprises to update a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.

10. A method for monitoring a characteristic of a monitored system, the method comprising:

producing, by a primary sensor, primary sensor data indicative of a primary characteristic of the monitored system;

producing, by a secondary sensor, secondary sensor data indicative of a secondary characteristic, different from the primary characteristic, of the monitored system;

determining, by a sensor controller, a measured value of the primary characteristic based on the primary sensor data;

determining, by the sensor controller, a measured value of the secondary characteristic based on the secondary sensor data;

predicting, by the sensor controller, a predicted value of the primary characteristic using a predictive model with the measured value of the secondary characteristic as an input to the predictive model;

determining, by the sensor controller, whether to update the predicted model based on a comparison of the measured value and the predicted value of the primary characteristic.

11. The method of claim 10, wherein determining whether to update the predicted model further comprises determining a difference between the measured value of primary characteristic and the predicted value of primary characteristic.

12. The method of claim 11, wherein determining whether to update the predicted model further comprises determining whether the difference is within a first reference threshold.

13. The method of claim 12, further comprising transmitting, by the sensor controller, in response to a determination that the difference is within the first reference threshold, a heartbeat signal to the model management system that indicates a predictive model managed by the model management system does not need updated.

14. The method of claim 12, wherein determining whether to update the predicted model further comprises determining whether the difference is within a second reference threshold in response to a determination that the difference is not within the first reference threshold.

15. The method of claim 14, further comprising updating, by the sensor controller, the predictive model in response to a determination that the difference is within the second reference threshold and not within first reference threshold,

wherein updating the predictive model comprises updating a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.

16. The method of claim 14, further comprising transmitting, in response to a determination that the difference is not within the second reference threshold, the measured value of primary characteristic, the measured value of secondary characteristic, and a plurality of model parameters to a model management system, wherein the plurality of model parameters comprise a predictive model used to predict the predicted value of the primary characteristic based on the measured value of the secondary characteristic.

17. The method of claim 14, further comprising updating, by the sensor controller, the predictive model in response to a determination that the difference is not within the second reference threshold, and

wherein updating the predictive model comprises updating a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.

18. One or more machine-readable storage media comprising a plurality of instructions stored thereon that, when executed, causes sensor controller of a predictive sensor module to:

acquire, from a primary sensor, primary sensor data indicative of a primary characteristic of the monitored system;

acquire, from a secondary sensor, secondary sensor data indicative of a secondary characteristic, different from the primary characteristic, of the monitored system;

determine a measured value of the primary characteristic based on the primary sensor data;

determine a measured value of the secondary characteristic based on the secondary sensor data;

predict a predicted value of the primary characteristic using a predictive model with the measured value of the secondary characteristic as an input to the predictive model;

determine whether to update the predicted model based on a comparison of the measured value and the predicted value of the primary characteristic.

19. The one or more machine-readable storage media of claim 18, wherein to determine whether to update the predicted model further comprises to determine a difference between the measured value of primary characteristic and the predicted value of primary characteristic.

20. The one or more machine-readable storage media of claim 19, wherein to determine whether to update the predicted model further comprises to determine whether the difference is within a first reference threshold.

21. The one or more machine-readable storage media of claim 20, wherein the plurality of instructions, when executed, further cause the sensor controller to transmit, in response to a determination that the difference is within the first reference threshold, a heartbeat signal to the model management system that indicates a predictive model managed by the model management system does not need updated.

22. The one or more machine-readable storage media of claim 20, wherein to determine whether to update the predicted model further comprises to determine whether the difference is within a second reference threshold in response to a determination that the difference is not within the first reference threshold.

23. The one or more machine-readable storage media of claim 22, wherein the plurality of instructions, when executed, further cause the sensor controller to update the predictive model in response to a determination that the difference is within the second reference threshold and not within first reference threshold,

wherein to update the predictive model comprises to update a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.

24. The one or more machine-readable storage media of claim 22, wherein the plurality of instructions, when executed, further cause the sensor controller to transmit, in response to a determination that the difference is not within the second reference threshold, the measured value of primary characteristic, the measured value of secondary characteristic, and a plurality of model parameters to a model management system, wherein the plurality of model parameters comprise a predictive model used to predict the predicted value of the primary characteristic based on the measured value of the secondary characteristic.

25. The one or more machine-readable storage media of claim 22, wherein the plurality of instructions, when executed, further cause the sensor controller to update the predictive model in response to a determination that the difference is not within the second reference threshold, and

wherein to update the predictive model comprises to update a plurality of model parameters of the predictive model based on at least one of the measured value of primary characteristic, the measured value of secondary characteristic, or the difference.