Learning transportation modes from raw GPS data

• US 8,315,959 B2
• Filed: 08/01/2011
• Issued: 11/20/2012
• Est. Priority Date: 02/26/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Zheng, Yu, Wang, Longhao, Liu, Like, Xie, Xing
• Primary Examiner(s): Holmes, Michael B
• Application Number:
  US13/195,496
• Publication Number:
  US 20110289031A1
• Time in Patent Office: 477 Days
• Field of Search: 706/14
• US Class Current: 706/14
• CPC Class Codes: G01C 21/20 Instruments for performing navigational calculations G01C21/24, G01C21/26 take precedence
• CPC Class Codes: G01C 21/3423 Multimodal routing, i.e. combining two or more modes of transportation, where the modes can be any of, e.g. driving, walking, cycling, public transport
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G08G 1/0112 from the vehicle, e.g. floating car data [FCD]

Abstract

Described is a technology by which raw GPS data is processed into segments of a trip, with a predicted mode of transportation (e.g., walking, car, bus, bicycling) determined for each segment. The determined transportation modes may be used to tag the GPS data with transportation mode information, and/or dynamically used. Segments are first characterized as walk segments or non-walk segments based on velocity and/or acceleration. Features corresponding to each of those walk segments or non-walk segments are extracted, and analyzed with an inference model to determine probabilities for the possible modes of transportation for each segment. Post-processing may be used to modify the probabilities based on transitioning considerations with respect to the transportation mode of an adjacent segment. The most probable transportation mode for each segment is selected.

Claims

1. A method comprising:

determining a transportation mode for one or more segments of positioning data based at least in part on one or more features of the one or more segments.

2. The method of claim 1, comprising separating the positioning data into a plurality of trips, determining the transportation mode corresponding to one of the trips.

3. The method of claim 1, comprising merging a segment below a threshold length into another segment.

4. The method of claim 1, determining the transportation mode for the one or more segments comprising differentiating between walk segments and non-walk segments based on at least one of velocity or acceleration.

5. The method of claim 4, differentiating between the walk segments and the non-walk segments comprising distinguishing between certain segments and uncertain segments based on a length of respective segments.

6. The method of claim 5, comprising:

in response to determining that a threshold number of consecutive segments are uncertain segments corresponding to walk segments, merging the uncertain segments corresponding to the walk segments into a non-walk segment.

7. The method of claim 1, the one or more features comprising location-related information.

8. The method of claim 1, the one or more features comprising timing-related information.

9. A system comprising:

an inference component configured to determine a transportation mode for one or more segments of positioning data based at least in part on one or more features of the one or more segments.

10. The system of claim 9, comprising:

a segmentation component configured to separate GPS data into the one or more segments of positioning data; and

a feature extraction component configured to extract the one or more features for the one or more segments of positioning data.

11. The system of claim 9, the inference component comprising at least one of a conditional random field framework, a decision tree classifier, a Bayesian Net inference model, or a support vector machine inference model.

12. The system of claim 11, comprising a post-processing component coupled to the inference component, the post-processing component configured to recalculate probabilities provided by the inference component into final probabilities using transition probabilities.

13. The system of claim 9, the one or more features comprising at least one of segment length, velocity data or acceleration data.

14. The system of claim 10, the segmentation component configured to merge a segment below a threshold length into another segment.

15. The system of claim 10, the segmentation component configured to:

differentiate between walk segments and non-walk segments;

distinguish between certain segments and uncertain segments based on a length of respective segments; and

merge uncertain segments into a non-walk segment.

16. One or more computer-readable media comprising computer-executable instructions, which when executed perform actions, comprising:

determining a transportation mode for one or more segments of positioning data based at least in part on one or more features of the one or more segments.

17. The one or more computer-readable media of claim 16, the actions comprising separating the positioning data into trips based on time duration information determined from the positioning data.

18. The one or more computer-readable media of claim 16, determining the transportation mode comprising determining probabilities for candidate transportation modes for respective segments.

19. The one or more computer-readable media of claim 18, determining the transportation mode comprising:

modifying at least some of the probabilities for candidate transportation modes for at least one of the one or more segments based on transition probability data corresponding to at least one adjacent segment; and

selecting as the transportation mode for the at least one of the one or more segments a candidate transportation mode having a desired probability value after recalculation.

20. The one or more computer-readable media of claim 16, the actions comprising tagging the positioning data with information corresponding to the determined transportation mode for respective segments.