LEARNING AND REASONING ABOUT THE CONTEXT-SENSITIVE RELIABILITY OF SENSORS

• US 20090002148A1
• Filed: 06/28/2007
• Published: 01/01/2009
• Est. Priority Date: 06/28/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric J.
• Granted Patent: US 7,696,866 B2
• Time in Patent Office: Days
• US Class Current: 340/514
• CPC Class Codes: G06Q 10/047 Optimisation of routes or paths, e.g. travelling salesman problem
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 50/06 Energy or water supply
• CPC Class Codes: G08G 1/0112 from the vehicle, e.g. floating car data [FCD]
• CPC Class Codes: G08G 1/0116 from roadside infrastructure, e.g. beacons
• CPC Class Codes: G08G 1/0133 for classifying traffic situation
• CPC Class Codes: G08G 1/097 Supervising of traffic control systems, e.g. by giving an alarm if two crossing streets have green light simultaneously

Abstract

A sensor monitoring system is described that identifies inaccurate or degraded sensor data and corresponding sensors. The system analyzes a plurality of sensors to determine reliable sensors and isolate or reinterpret unreliable sensor data. An analysis component analyzes a traffic representation that varies as context varies and determines the probability that current sensor data is invalid. Learning and reasoning are performed from a case library of sensor failure data to build predictive models of sensor health that can provide inferences about the output of sensors. Resulting inferences about sensor reliability can be employed in base-level inferences and predictions about traffic flow.

Claims

1. A system for context-sensitive monitoring of sensor performance for an arterial flow system comprising the following computer-executable components:

a sensor interface component that obtains sensor data from a plurality of sensors;

a context analyzer component that analyzes contextual data associated with the plurality of sensors;

a sensor analyzer component that identifies degraded sensor data based at least in part upon analysis of the contextual data; and

an output component that provides output sensor data based at least in part upon identification of the degraded sensor data.

2. The system of claim 1, the plurality of sensors is heterogeneous.

3. The system of claim 1, the output sensor data is based at least in part upon the sensor data, filtered to remove the degraded sensor data or to appropriately use the sensor data in accordance with a sensor error model.

4. The system of claim 1, the output sensor data includes a probabilistic function that describes the degradation of the sensor data.

5. The system of claim 1, further comprising a creation component that employs machine learning to build at least one model used to predict and/or interpret output of sensors, including error models and/or failure models.

6. The system of claim 1, further comprising:

a notification component that generates an alert as a function of the degraded sensor data.

7. The system of claim 1, further comprising:

a sensor predictor component that predicts failure of a subset of the sensors; and

a notification component that generates an alert as a function of the predicted failure.

8. The system of claim 1, further comprising a sensor data store that maintains previous sensor data used in analysis of the sensor data.

9. The system of claim 1, the sensor analyzer component identifies degradation in the sensor data based at least in part upon sensor data from a physically proximate sensor.

10. The system of claim 1, the sensor analyzer component utilizes a traffic flow representation during identification of degraded sensor data.

11. A computer-implemented method for identifying questionable sensor data, comprising:

obtaining data from a data source;

analyzing context of the data source; and

identifying a subset of the data as suspect data based at least in part upon context.

12. The method of claim 11, further comprising outputting the suspect data and a probability that the data is incorrect.

13. The method of claim 11, the act of obtaining data includes receiving data from a collection of heterogeneous sensors.

14. The method of claim 11, further comprising:

predicting a failure of the data source as a function of identifying suspect data; and

generating a notification as a function of the predicted failure of the data source.

15. The method of claim 11, further comprising:

predicting a probability of failure of the data source within a specified time period; and

establishing a notification based upon the probability.

16. The method of claim 11, the context includes at least one of day of week, time of day, and/or weather conditions.

17. The method of claim 11, the context of a first data source includes data obtained from a physically proximate data source.

18. A system for context-sensitive monitoring of sensor performance for an arterial flow system comprising the following computer-executable components:

means for receiving sensor data from a plurality of sensors;

means for analyzing sensor data based at least in part upon sensor context; and

means for filtering output data based at least in part upon the sensor data analysis.

19. The system of claim 18, the sensor data analysis utilizes a traffic flow representation.

20. The system claim 18, further comprising:

means for predicting a failure of a subset of the sensors; and

means for notifying a user of the predicted sensor failure.