LARGE-SCALE HUMAN MOBILITY PATTERN RECOGNITION

• US 20140195470A1
• Filed: 01/08/2013
• Published: 07/10/2014
• Est. Priority Date: 01/08/2013
• Status: Abandoned Application

• Current Assignee: Telefonaktiebolaget LM Ericsson
• Original Assignee: Telefonaktiebolaget LM Ericsson
• Inventors: Xing, Bo, Li, Du
• Application Number:
  US13/736,589
• Publication Number:
  US 20140195470A1
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G08G 1/012 from other sources than vehicle or roadside beacons, e.g. mobile networks
• CPC Class Codes: G08G 1/0137 for specific applications

Abstract

A method for grouping trips of a mobile device, a trip including a set of trip traces from a plurality of traces collected from base stations. Each base station includes a unique station-ID. Each trace includes a station-ID and an association time indicating when the mobile device was associated with the base station. The mobile device is associated with at most one base station at any point in time. The method to group two trips by calculating an edit distance using an approximate string matching algorithm. The method includes identifying traces for the two trips, extracting ordered sets of station-IDs from the traces, calculating the edit distance between the two trips by applying the approximate string matching algorithm, to the ordered sets of station-IDs, checking if the edit distance is less than a preset edit distance threshold, and grouping the two trips into a group of trips.

Claims

1. A method executed by a computer processor for grouping a plurality of trips of a mobile device, where each trip in the plurality of trips includes a set of trip traces from a plurality of traces for the mobile device, where the plurality of traces is collected from a plurality of base stations and stored in a trace storage, where each base station in the plurality of base stations includes a unique station-ID, where each trace in the plurality of traces includes a station-ID of a base station in the plurality of base stations and an association time indicating when the mobile device was associated with the base station, and where the mobile device is associated with at most one base station at any point in time, the method to group a first trip in the plurality of trips and a second trip in the plurality of trips by calculating an edit distance between the first trip and the second trip using an approximate string matching algorithm, the method comprising the steps of:

identifying a first set of trip traces for the first trip and a second set of trip traces for the second trip in the plurality of traces for the mobile device;

extracting a first ordered set of station-IDs from the first set of trip traces and a second ordered set of station-IDs from the second set of trip traces, wherein each ordered set of station-IDs is ordered by association time;

calculating the edit distance between the first trip and the second trip by applying the approximate string matching algorithm to the first ordered set of station-IDs and the second ordered set of station-IDs;

checking that the edit distance between the first trip and the second trip is less than a preset edit distance threshold; and

grouping the first trip and the second trip into a group of trips in response to determining that the edit distance is less than the preset edit distance threshold.

2. The method of claim 1, wherein identifying the first set of trip traces for the first trip further comprises the steps of:

identifying a first stay including a first stay association time and a second stay including a second stay association time for the mobile device, wherein a stay includes a set of stay traces from the plurality of traces;

identifying a trace having an association time between the first stay association time and the second stay association time; and

adding the trace to the first set of trip traces of the first trip.

3. The method of claim 2, wherein identifying the first stay further comprises the steps of:

checking whether the mobile device was in an area of sparse base station coverage or an area of dense base station coverage during a particular association time of a particular trace;

in response to determining that the mobile device was in the area of sparse base station coverage;

checking that the particular association time of the particular trace lasted longer than a preset sparse threshold, and

setting the particular trace as the set of stay traces for the first stay; and

in response to determining that the mobile device was in the area of dense base station coverage;

checking that the mobile device had oscillating associations between a set of base stations in the plurality of base stations during at least one of before and after the particular association time of the particular trace,

checking that a total association time for the oscillating associations lasted longer than a preset dense threshold, and

adding the particular trace to the set of stay traces for the first stay.

4. The method of claim 3, wherein identifying the first stay further comprises the step of:

checking that a geographical distance between any two base stations in the set of base stations is less than a preset geographical distance threshold in response to determining that the mobile device was in the area of dense base station coverage.

5. The method of claim 1, further comprising the step of:

associating a user demographic information with the first trip.

6. The method of claim 1, further comprising the steps of:

associating a context with the first trip based on at least one of an association time of the first trip and a geographical location of associated base stations of the first trip, the context providing at least one of temporal and geographical information for the first trip; and

checking that the second trip is associated with the context before grouping the first trip and the second trip into the group of trips.

7. The method of claim 1, further comprising the steps of:

checking that a number of trips in the group of trips has reached a preset count threshold;

identifying a first mobility pattern for the mobile device based on the group of trips in response to determining that the number of trips has reached the preset count threshold, wherein a mobility pattern includes a set of pattern traces from the plurality of traces;

checking that the first trip has a shortest average edit distance to all other trips in the group of trips; and

setting the first set of trip traces of the first trip as the set of pattern traces for the first mobility pattern in response to determining that the first trip has the shortest average edit distance.

8. The method of claim 7, further comprising the steps of:

associating a first context with the first mobility pattern, wherein the first context is associated with every trip in the group of trips; and

indexing the first mobility pattern by the first context when the first mobility pattern is stored in a mobility pattern storage.

9. The method of claim 8, further comprising the steps of:

checking that an edit distance between the first mobility pattern of the mobile device and a second mobility pattern of the mobile device is less than a preset pattern distance threshold;

checking that the first context of the first mobility pattern is different than a second context of the second mobility pattern; and

combining the first mobility pattern and the second mobility pattern into a new mobility pattern for the mobile device, wherein the new mobility pattern is not associated with the first context or the second context.

10. The method of claim 7, further comprising the step of:

calculating an occurrence rating for the first mobility pattern based on the number of trips in the group of trips and a total number of trips for the mobile device.

11. The method of claim 7, further comprising the steps of:

comparing the first mobility pattern of the mobile device to a second mobility pattern of a different mobile device; and

combining the first mobility pattern and the second mobility pattern into a crowd mobility pattern based on at least one of a set of contexts associated with the first mobility pattern, a set of contexts associated with the second mobility pattern, and an edit distance between the first mobility pattern and the second mobility pattern.

12. The method of claim 11, further comprising the step of:

determining a number of mobile devices represented by the crowd mobility pattern.

13. A system for grouping a plurality of trips of a mobile device, wherein each trip in the plurality of trips includes a set of trip traces from a plurality of traces for the mobile device, wherein the plurality of traces is collected from a plurality of base stations, wherein each base station in the plurality of base stations includes a unique station-ID, wherein each trace in the plurality of traces includes a station-ID of a base station in the plurality of base stations and an association time indicating when the mobile device was associated with the base station, and wherein the mobile device is associated with at most one base station at any point in time, the system to group a first trip in the plurality of trips and a second trip in the plurality of trips by calculating an edit distance between the first trip and the second trip using an approximate string matching algorithm, the system comprising:

a trace storage configured to store the plurality of traces;

a trip finding module coupled to the trace storage and configured to identify a first set of trip traces for the first trip and a second set of trip traces for the second trip in the plurality of traces for the mobile device;

an edit distance calculation module configured to;

extract a first ordered set of station-IDs from the first set of trip traces and a second ordered set of station-IDs from, the second set of trip traces, wherein each ordered set of station-IDs is ordered by association time, and

calculate the edit distance between the first trip and the second trip by applying the approximate string matching algorithm to the first ordered set of station-IDs and the second ordered set of station-IDs; and

a trip grouping module configured to;

check that the edit distance between the first trip and the second trip is less than a preset edit distance threshold, and

group the first trip and the second trip into a group of trips in response to determining that the edit distance is less than the preset edit distance threshold.

14. The system of claim 13, further comprising:

a stay finding module coupled to the trace storage and configured to identify a first stay including a first stay association time and a second stay including a second stay association time for the mobile device, wherein a stay includes a set of stay traces from the plurality of traces,

wherein the trip finding module is further configured to;

identify a trace having an association time between the first stay association time and the second stay association time, and

add the trace to the first set of trip traces of the first trip.

15. The system of claim 14, wherein the stay finding module is further configured to:

check whether the mobile device was in an area of sparse base station coverage or an area of dense base station coverage during a particular association time of a particular trace;

in response to determining that the mobile device was in the area of sparse base station coverage;

check that the particular association time of the particular trace lasted longer than a preset sparse threshold, and

set the particular trace as the set of stay traces for the first stay; and

in response to determining that the mobile device was in the area of dense base station coverage;

check that the mobile device had oscillating associations between a set of base stations in the plurality of base stations during at least one of before and after the particular association time of the particular trace,

check that a total association time for the oscillating associations lasted longer than a preset dense threshold, and

add the particular trace to the set of stay traces for the first stay.

16. The system of claim 15, wherein the stay finding module is further configured to check that a geographical distance between any two base stations in the plurality of base stations is less than a preset geographical distance threshold in response to determining that the mobile device was in the area of dense base station coverage.

17. The system of claim 13, further comprising:

a demographic association module configured to associate a user demographic information with the first trip.

18. The system of claim 13, further comprising:

an ontology storage configured to store a plurality of contexts, wherein each context in the plurality of contexts provides at least one of temporal and geographical information; and

a context association module coupled to the ontology storage and configured to associate a context with the first trip based on at least one of an association time of the first trip and a geographical location of associated base stations of the first trip.

wherein the trip grouping module is further configured to check that the second trip is associated with the context before grouping the first trip and the second trip into the group of trips.

19. The system of claim 13, further comprising:

a mobility pattern module configured to;

check that a number of trips in the group of trips has reached a preset count threshold,

identify a first mobility pattern for the mobile device based on the group of trips in response to determining that the number of trips has reached the preset count threshold, wherein a mobility pattern includes a set of pattern traces from the plurality of traces,

check that the first trip has a shortest average edit distance to all other trips in the group of trips, and

set the first set of trip traces of the first trip as the set of pattern traces for the first mobility pattern in response to determining that the first trip has the shortest average edit distance.

20. The system of claim 19, further comprising:

a mobility pattern storage coupled to the mobility pattern module and configured to index the first mobility pattern by a first context when storing the first mobility pattern, wherein the first context is associated with every trip in the group of trips, and

wherein the mobility pattern module is further configured to associate the first context with the first mobility pattern.

21. The system of claim 20, further comprising:

a pattern aggregation module configured to;

check that an edit distance between the first mobility pattern of the mobile device and a second mobility pattern of the mobile device is less than a preset pattern distance threshold,

check that the first context of the first mobility pattern is different than a second context of the second mobility pattern, and

combine the first mobility pattern and the second mobility pattern into a new mobility pattern for the mobile device, wherein the new mobility pattern is not associated with the first context or the second context.

22. The system of claim 19, wherein the mobility pattern module is further configured to calculate an occurrence rating for the first mobility pattern based on the number of trips in the group of trips and a total number of trips for the mobile device.

23. The system of claim 19, further comprising:

a crowd pattern module configured to;

compare the first mobility pattern of the mobile device to a second mobility pattern of a different mobile device, and

combine the first mobility pattern and the second mobility pattern into a crowd mobility pattern based on at least one of a set of contexts associated with the first mobility pattern, a set of contexts associated with the second mobility pattern, and an edit distance between the first mobility pattern and the second mobility pattern.

24. The system of claim 23, wherein the crowd pattern module is further configured to determine a number of mobile devices represented by the crowd mobility pattern.