Data validation

• US 10,540,225 B2
• Filed: 05/27/2015
• Issued: 01/21/2020
• Est. Priority Date: 05/27/2015
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Inventors: Kaur, Satwant, Makkinejad, Babak, Doshi, Parag, Wick, Corey
• Primary Examiner(s): Bonzo, Bryce P
• Assistant Examiner(s): Arcos, Jeison C
• Application Number:
  US15/574,578
• Publication Number:
  US 20180157552A1
• Time in Patent Office: 1,700 Days
• 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 23/0262 Confirmation of fault detection, e.g. extra checks to confirm that a failure has indeed occurred
• 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 11/0736 in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function testing or monitoring of automated control systems G05B23/02
• 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 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/16 using machine learning or artificial intelligence
• CPC Class Codes: H04L 67/12 specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks

Abstract

Examples associated with data validation are disclosed. One example includes a sensor attribute data store having information describing attributes of a set of distributed sensors. A pattern data store stores information describing patterns indicating anomalous sensor activity. A data aggregation module flags data received from a tested sensor as anomalous data when the anomalous data exceeds a variance level described by an attribute of the tested sensor. A data validation module validates the anomalous data by comparing the anomalous data to the patterns indicating anomalous sensor activity. A learning module updates the pattern indicating anomalous sensor activity based on a result received from the validation logic after the validation logic validates data received from the tested sensor.

Claims

1. A system, comprising:

a sensor attribute data store to store information describing attributes of a set of distributed sensors, wherein the attributes of the set of distributed sensors comprise attributes describing model information of the sensors, attributes describing expected data to be received from the sensors, and attributes describing expected anomalous behaviors of the sensors;

a pattern data store to store information describing patterns indicating anomalous sensor activity;

a data aggregation module to flag data received from a tested sensor as anomalous data when the anomalous data exceeds a variance level described by an attribute of the tested sensor;

a data validation module to validate the anomalous data by comparing the anomalous data to the patterns indicating anomalous sensor activity;

a learning module to update a pattern indicating anomalous sensor activity based on a result received from the validation logic after the validation logic validates data received from the tested sensor.

2. The system of claim 1, further comprising:

an error logging data store to store anomalous data that fails validation; and

an error checking module to re-validate data in the error logging data store when the learning module updates a pattern indicating anomalous sensor activity.

3. The system of claim 1, further comprising:

a validated data store to store data received from the tested sensor that falls within the variance level and to store anomalous data received from the tested sensor that passes validation by the data validation module; and

a distribution module to facilitate distribution of data in the validated data store.

4. The system of claim 1, where the patterns indicating anomalous sensor activity include:

specific scenarios that describe a sensor behavior to be treated as anomalous; and

learned patterns of sensor behaviors to be treated as anomalous.

5. The system of claim 1, where the learning logic updates the pattern by modifying a weight associated with the pattern, the weight indicating a likelihood that anomalous data satisfying the pattern is a result of a sensor malfunction.

6. A method, comprising:

receiving sensor data from a set of distributed sensors, the set of distributed sensors associated with attributes describing model information of the sensors, attributes describing expected data to be received from the sensors, and attributes describing expected anomalous behaviors of the sensors;

marking sensor data as anomalous data when the anomalous data exceeds a variance level associated with a sensor from which the anomalous data was received;

validating, based on a set of patterns describing anomalous sensor activity, whether the anomalous data is a result of one of a sensor malfunction and an event of significance;

updating the set of patterns based on whether the anomalous data is validated as the sensor malfunction and the event of significance; and

distributing a plurality of verified data comprising sensor data within the variance level and validated anomalous data.

7. The method of claim 6, further comprising:

logging anomalous data indicated as a sensor malfunction; and

revalidating logged anomalous data after an update to the set of patterns.

8. The method of claim 6, where distributing the plurality of verified data comprises one or more of, providing raw data associated with a portion of the plurality of verified data, providing query access to the plurality of verified data, and providing a result of a query of the plurality of verified data.

9. The method of claim 6, where variance levels for sensors are generated based on device specifications describing known sensor malfunctions for respective sensors.

10. The method of claim 6, where the set of patterns includes weights that indicate a likelihood that anomalous data is a result of a sensor malfunction and where updating the set of patterns includes modifying the weights.

11. A non-transitory computer-readable medium storing computer-executable instructions that when executed by a computer cause the computer to:

receive sensor data from a set of distributed sensors, the set of distributed

sensors associated with attributes describing model information of the sensors,

attributes describing expected data to be received from the sensors, and

attributes describing expected anomalous behaviors of the sensors;

mark sensor data as anomalous data when the anomalous data exceeds a variance level associated with a sensor from which the anomalous data was received;

store non-anomalous data in a validated data store;

store in the validated data store, anomalous data found, based on a set of patterns describing anomalous sensor activity, to be a result of an event of significance;

update the set of patterns; and

distribute data from the validated data store.

12. The non-transitory computer-readable medium of claim 11, where the instructions further cause the computer to:

store in an error logging data store, anomalous data found, based on the set of patterns, to be a result of a sensor malfunction; and

periodically evaluate whether data in the error logging data store is a result of an event of significance based on updated patterns.

13. The non-transitory computer-readable medium of claim 11, where the variance level for a sensor is based on accuracy information associated with the sensor and how recently the sensor was calibrated.