EVENT-DETECTION IN MULTI-CHANNEL SENSOR-SIGNAL STREAMS

• US 20090099820A1
• Filed: 10/15/2007
• Published: 04/16/2009
• Est. Priority Date: 10/27/2006
• Status: Active Grant

• Current Assignee: Sony France SA (Sony Group Corp.)
• Original Assignee: Sony France SA (Sony Group Corp.)
• Inventors: KAPLAN, Frederic
• Granted Patent: US 7,996,189 B2
• Time in Patent Office: Days
• US Class Current: 702/189
• CPC Class Codes: H04M 1/72451 according to schedules, e.g. using calendar applications
• CPC Class Codes: H04M 1/72454 according to context-related or environment-related conditions
• CPC Class Codes: H04M 2250/12 including a sensor for measuring a physical value, e.g. temperature or motion telephones including GPS signal receivers H04M2250/10
• CPC Class Codes: H04W 24/08 Testing, supervising or monitoring using real traffic

Abstract

Multiple parallel signal streams are produced by the sensors (20, 30, 40, 50) associated with a device (10). Events corresponding to real-world activities, contexts, environments, internal-states, uses, interactions and the like involving the device (10) can be detected in the multi-channel sensor-signal stream by evaluating how the pattern of coordination between the different sensors varies over time. The pattern of coordination between the sensors can be evaluated for a particular time period by computing an information distance matrix (IDM) for the sensor-signals generated during that time period. The IDM summarizes the normalized information distance between each possible pairing of the sensors.

Claims

1. A signal processing method for identifying events in a multi-channel sensor-signal stream, comprising the steps of:

inputting a section of said multi-channel sensor-signal stream, said section consisting of respective signal streams output by a plurality of sensors during a common time period;

evaluating the pattern of coordination that exists between the sensor-signal streams during respective time intervals within said common time period,

comparing the evaluated patterns of coordination, and

detecting events in the multi-channel sensor-signal stream by determining when the patterns of coordination between the different sensor-signal streams evaluated for consecutive time intervals are similar to each other.

2. The event-detection method according to claim 1, wherein the evaluating step comprises the step of computing the information distance between signals from pairs of said sensors during said time interval.

3. The event-detection method according to claim 1 or 2, wherein the plurality of sensors comprises sensors adapted to measure values of parameters of the environment, use, context and/or internal state of a device and/or parameters of interactions occurring between said device and one or more external agents, wherein said external agents can include the environment, one or more users, or one or more devices.

4. The event-detection method according to claim 1 or 2, wherein the plurality of sensors comprises sensors adapted to measure values of parameters of the environment, use, context and/or internal state of a set of devices and/or parameters of interactions occurring between one or more devices of said set and one or more external agents, wherein said external agents can include the environment, one or more users, or one or more devices.

5. The event-detection method according to claim 1 or 2, wherein the plurality of sensors comprises sensors adapted to measure values of parameters of the environment and/or physical state of a human or animal and/or parameters of activities engaged in by said human or animal and/or parameters of interactions occurring between said human or animal and one or more external agents, wherein said external agents can include the environment, one or more users, or one or more devices.

6. The event-detection method according to claim 1 or 2, wherein the correspondence between the different channels of the multi-channel sensor-signal stream and the individual sensors within said plurality of sensors is unknown.

7. Event-detection apparatus (100) adapted to detect identifying events in a multi-channel sensor-signal stream, the apparatus comprising:

input means adapted to receive as input a section of said multi-channel sensor-signal stream, said section consisting of respective signal streams output by a plurality of sensors (20-50) during a common time period;

computing means adapted to evaluate the pattern of coordination that exists between the sensor-signal streams during respective different time intervals within said common time period;

comparing means adapted to compare the patterns of coordination evaluated by the computing means and to detect events in the multi-channel sensor-signal stream by determining when the patterns of coordination between the different sensor-signal streams evaluated for consecutive time intervals are similar to each other; and

output means adapted to output event data identifying at least the timings of the detected events.

8. Event-detection apparatus (100) according to claim 7, wherein the computing means is adapted to evaluate the pattern of coordination between sensor streams during a give time interval by computing the information distances between pairs of said sensor streams during said time interval.

9. Event-detection apparatus (100) according to claim 7 or 8, and comprising classification means adapted to assign categories to events detected by the comparing means, wherein the output means is adapted to output event data identifying the categories of the detected events.

10. An event-detection system comprising an event-detection apparatus (100) according to any one of claims 7 to 9 and a set of sensors (20-50), said set of sensors (20-50) being adapted to produce said multi-channel sensor-signal stream.