Systems and methods for probabilistic semantic sensing in a sensory network

• US 10,417,570 B2
• Filed: 03/05/2015
• Issued: 09/17/2019
• Est. Priority Date: 03/06/2014
• Status: Active Grant

• Current Assignee: Verizon Patent and Licensing Incorporated (Verizon Communications Inc.)
• Original Assignee: Verizon Patent and Licensing Incorporated (Verizon Communications Inc.)
• Inventors: Florence, Peter Raymond, Sachs, Christopher David, Ryhorchuk, Kent W.
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): Seck, Ababacar
• Application Number:
  US14/639,901
• Publication Number:
  US 20150254570A1
• Time in Patent Office: 1,657 Days
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: H05B 47/115 by determining the presence or movement of objects or living beings
• CPC Class Codes: H05B 47/12 by detecting audible sound
• CPC Class Codes: Y02B 20/40 Control techniques providing energy savings, e.g. smart controller or presence detection

Abstract

Systems and methods for probabilistic semantic sensing in a sensory network are disclosed. The system receives raw sensor data from a plurality of sensors and generates semantic data including sensed events. The system correlates the semantic data based on classifiers to generate aggregations of semantic data. Further, the system analyzes the aggregations of semantic data with a probabilistic engine to produce a corresponding plurality of derived events each of which includes a derived probability. The system generates a first derived event, including a first derived probability, that is generated based on a plurality of probabilities that respectively represent a confidence of an associated semantic datum to enable at least one application to perform a service based on the plurality of derived events.

Claims

1. A method comprising:

receiving, by one or more processors of a machine, raw sensor data from a plurality of sensors;

generating, by one or more processors of the machine, a set of classifiers for the automatic probabilistic semantic sensing using semantic data based on the raw sensor data, the generated semantic data including a plurality of parallel sensed event records that each indicate a corresponding event sensed by a corresponding sensor among the plurality of sensors, each sensed event record including a first classifier in the set of classifiers that indicates an event detected by the corresponding sensor, a second classifier in the set of classifiers that indicates a probability that the first classifier is true, a third classifier in the set of classifiers that indicates a sensor location including a spatial coordinate of the corresponding sensor, and a fourth classifier in the set of classifiers that indicates an event location at which the detected event occurred;

grouping, by one or more processors of the machine, the parallel sensed event records into intermediate groups of sensed event records based on a subset of the first through fourth classifiers included in each of the sensed event records;

generating, by one or more processors of the machine, a final derived event record based on a group of sensed event records selected from the intermediate groups of sensed event records, the generated final derived event record including a fifth classifier that includes a temporal coordinate associated with the third classifier and indicates an event detected by multiple sensors among the plurality of sensors and a sixth classifier that indicates a probability that the fifth classifier is true;

enabling, by one or more processors of the machine, at least one application to perform a service or sequence of events based on the generated final derived event record; and

detecting, by one or more processors of the machine, an object based on the generated final derived event record.

2. The method of claim 1, wherein the plurality of sensors includes a first sensor located on a first node in a light sensory network that includes a plurality of nodes.

3. The method of claim 1, wherein the raw sensor data includes visual data, audio data, and environmental data and wherein the event further includes detecting a person, detecting a vehicle, and detecting an empty parking space.

4. The method of claim 1, wherein the third classifier includes a spatial coordinate of a first sensor and wherein the plurality of classifiers includes a seventh classifier including a temporal coordinate that is associated with the third classifier.

5. The method of claim 4, wherein the fourth classifier includes a spatial coordinate of a first event represented by the semantic datum and wherein the plurality of classifiers includes an eighth classifier including a temporal coordinate that describes a time the first sensor was utilized to detect a first event represented by the semantic datum and wherein the first event is an occupancy state of empty for a parking spot.

6. The method of claim 1, wherein the first plurality of classifiers includes a further classifier including an application identifier that is utilized to identify the at least one application from a plurality of applications, wherein each application from the plurality of applications is utilized to perform a different service.

7. The method of claim 1, the grouping of the sensed event records includes at least one of correlating to generate abstract graphs based on classifiers that match or correlating to generate abstract graphs based on classifiers that fuzzy match.

8. The method of claim 1, wherein the grouping of the sensed event records includes correlating based on mathematical relationships between spatial and temporal coordinates and wherein the spatial and temporal coordinates involve at least one of a sensor itself, a detection of an event, and a combination of both.

9. The method of claim 1, wherein the plurality of sensed events include a first sensed event, a second sensed event and a third sensed event, and wherein the first sensed event includes a corresponding first classifier that describes an occupancy state of empty for a first parking spot, the second sensed event includes a corresponding first classifier that describes an occupancy state of empty for the first parking spot, and a third sensed event includes a corresponding first classifier that describes an occupancy state of empty for the first parking spot, and wherein the plurality of aggregated semantic data includes a first aggregated semantic data that aggregates the first, second and third sensed events.

10. A system comprising:

a plurality of sensing engines, implemented by one or more processors, that are configured to generate a set of classifiers for the automatic probabilistic semantic sensing using raw sensor data received from a plurality of sensors, the plurality of sensing engines are further configured to generate semantic data based on the raw sensor data, the generated semantic data including a plurality of parallel sensed event records that each indicate a corresponding event sensed by a corresponding sensor among the plurality of sensors, each sensed event record including a first classifier in the set of classifiers that indicates an event detected by the corresponding sensor, a second classifier in the set of classifiers that indicates a probability that the first classifier is true, a third classifier in the set of classifiers that indicates a sensor location including a spatial coordinate of the corresponding sensor, and a fourth classifier in the set of classifiers that indicates an event location at which the detected event occurred; and

an correlation engine, implemented by one or more processors, that is configured to group the parallel sensed event records into intermediate groups of sensed event records based on a subset of the first through fourth classifiers included in each of the sensed event records;

a probabilistic engine, implemented by one or more processors, that is configured to generate a final derived event record based on a group of sensed event records selected from the groups of sensed event records, the generated derived event record including a fifth classifier that includes a temporal coordinate associated with the third classifier and indicates an event detected by multiple sensors among the plurality of sensors and a sixth classifier that indicates a probability that the fifth classifier is true, the probabilistic engine is further configured to enable at least one application to perform a service or sequence of events based on the generated derived event record; and

a light controller, communicatively coupled to the plurality of sensors and the probabilistic engine, wherein the probabilistic engine commands the light controller to adjust lighting based on the generated derived event record.

11. The system of claim 10, wherein the plurality of sensors includes a first sensor located on a first node in a light sensory network that includes a plurality of nodes.

12. The method of claim 10, wherein the raw sensor data includes visual data, audio data, and environmental data and wherein the event includes a detection of a person, a detection of a vehicle, a detection of an object, and a detection of an empty parking space.

13. The system of claim 10, wherein the third classifier includes a spatial coordinate of a first sensor and wherein the plurality of classifiers includes a seventh classifier including a temporal coordinate that is associated with the third classifier.

14. The system of claim 13, wherein the fourth classifier includes a spatial coordinate of a first event represented by the semantic datum and wherein the plurality of classifiers includes an eighth classifier including a temporal coordinate that describes a time the first sensor was utilized to detect a first event represented by the semantic datum and wherein the first event is an occupancy state of empty for a parking spot.

15. The system of claim 10, wherein the probabilistic engine is configured to analyze, based on user inputs, wherein the user inputs include a desired accuracy and a user preference.

16. The system of claim 10, wherein the probabilistic engine is configured to analyze based on a weight that is assigned to a second classifier and wherein the probabilistic engine alters the weight over time.

17. The system of claim 10, wherein the plurality of derived events includes a first derived event and wherein the probabilistic engine analyzes based on a first threshold and wherein the first threshold defines a minimum level of raw sensor data that is utilized to produce the first derived event.

18. The system of claim 17, wherein the probabilistic engine alters the first threshold over time.

19. The system of claim 10, wherein the at least one application includes a parking location application, a surveillance application, a traffic application, retail customer application, a business intelligence application, an asset monitoring application, an environmental application, an earthquake sensing application.

20. A non-transitory machine-readable medium storing a set of instructions that, when executed by a processor, cause a machine to perform operations in a method, the operations including, at least:

receiving, by one or more processors of a machine, raw sensor data from a plurality of sensors;

generating, by one or more processors of the machine, a set of classifiers for the automatic probabilistic semantic sensing using semantic data based on the raw sensor data, the generated semantic data including a plurality of parallel sensed event records that each indicate a corresponding event sensed by a corresponding sensor among the plurality of sensors, each sensed event record including a first classifier in the set of classifiers that indicates an event detected by the corresponding sensor, a second classifier in the set of classifiers that indicates a probability that the first classifier is true, a third classifier in the set of classifiers that indicates a sensor location including a spatial coordinate of the corresponding sensor, and a fourth classifier in the set of classifiers that indicates an event location at which the detected event occurred;

grouping, by one or more processors of the machine, the parallel sensed event records into intermediate groups of sensed event records based on a subset of the first through fourth classifiers included in each of the sensed event records;

generating, by one or more processors of the machine, a final derived event record based on a group of sensed event records selected from the intermediate groups of sensed event records, the generated final derived event record including a fifth classifier that includes a temporal coordinate associated with the third classifier and indicates an event detected by multiple sensors among the plurality of sensors and a sixth classifier that indicates a probability that the fifth classifier is true; and

enabling, by one or more processors of the machine, at least one application to perform a service or sequence of events based on the generated final derived event record.