Detection of travel mode associated with computing devices

• US 10,401,181 B2
• Filed: 08/09/2017
• Issued: 09/03/2019
• Est. Priority Date: 08/09/2017
• Status: Active Grant

• Current Assignee: Mapbox, Inc.
• Original Assignee: Mapbox, Inc.
• Inventors: Chai, Lily, Herlocker, Morgan
• Primary Examiner(s): Marc, McDieunel
• Application Number:
  US15/673,160
• Publication Number:
  US 20190049255A1
• Time in Patent Office: 755 Days
• 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: G08G 1/123 indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, tramstransmission of navigation instructions to vehicles G08G1/0968
• CPC Class Codes: H04W 4/02 Services making use of location information

Abstract

Digital data corresponding to a trace by a device is received. The digital data includes a set of values for each of a plurality of metrics. A set of derived values for the plurality of metrics are calculated from the sets of values corresponding to the metrics, and the set of derived values are compared to a machine learning model trained using a plurality of travel modes. A plurality of confidence scores for the derived values are calculated. The plurality of confidence scores correspond to the plurality of travel modes. In response to a confidence score for a first travel mode of the plurality of confidence scores for the plurality of travel modes meeting a threshold value for the first travel mode, the received digital data corresponding to the trace is labelled with the first travel mode.

Claims

1. A data processing method providing an improvement in classification of inferred activities of mobile computing devices based upon data representing movement of the devices, the method comprising:

receiving digital data corresponding to a trace by a device, the digital data comprising a set of values for each of a plurality of metrics;

calculating a set of derived values for the plurality of metrics from the sets of values corresponding to the metrics;

comparing the set of derived values to a machine learning model trained using a plurality of travel modes;

calculating a plurality of confidence scores for the derived values, the plurality of confidence scores corresponding to the plurality of travel modes; and

in response to a confidence score for a first travel mode of the plurality of confidence scores for the plurality of travel modes meeting a threshold value for the first travel mode, labeling the received digital data corresponding to the trace with the first travel mode.

2. The method of claim 1, wherein:

a first metric of the plurality of metrics is a speed metric; and

a second metric of the plurality of metrics is one of a directional metric, an acceleration metric, and a deceleration metric.

3. The method of claim 1, wherein the set of derived values comprises a set of vectors of quantile values, each vector corresponding to a respective metric of the plurality of metrics, and calculating the set of derived values for the plurality of metrics comprises calculating at least a first vector of quantile values for a first metric of the plurality of metrics and calculating a second vector of quantile values for a second metric of the plurality of metrics.

4. The method of claim 3, wherein calculating the first vector of quantile values for the first metric comprises:

ordering the set of values for the first metric;

determining, for each of a plurality of predetermined quantiles, a quantile value by identifying a value at a position within the ordered set of values that corresponds to the predetermined quantile; and

dividing each quantile value by a maximum value for the first metric to obtain a normalized quantile value; and

storing the normalized quantile values in the first vector of quantile values.

5. The method of claim 4, wherein calculating the first confidence score for the first travel mode comprises weighting at least one quantile value higher than a second quantile value.

6. The method of claim 1, further comprising:

receiving raw trace data comprising location data and corresponding time stamps; and

generating the digital data corresponding to the trace for each metric of the plurality of metrics by calculating the set of values corresponding to the metric from the location data and the corresponding time stamps in the raw trace data.

7. The method of claim 1, further comprising training the machine learning model based on a plurality of training sets of digital data each comprising an identification of one of the plurality of travel modes.

8. The method of claim 7, wherein training the machine learning model comprises:

determining a set of rules for calculating a confidence score for a particular travel mode of the plurality of travel modes based on the training sets of digital data and the identified travel modes.

9. The method of claim 1, wherein the first travel mode is a driving mode, the method further comprising:

in response to labeling the received digital data corresponding to the trace with the driving mode, adding data associated with the trace to a driving database.

10. A non-transitory computer-readable storage containing computer program code executable on a processor classifying inferred activities of mobile computing devices based upon data representing movement of the devices, the computer program code for:

receiving digital data corresponding to a trace by a device, the digital data comprising a set of values for each of a plurality of metrics;

calculating a set of derived values for the plurality of metrics from the sets of values corresponding to the metrics;

comparing the set of derived values to a machine learning model trained using a plurality of travel modes;

calculating a plurality of confidence scores for the derived values, the plurality of confidence scores corresponding to the plurality of travel modes; and

in response to a confidence score for a first travel mode of the plurality of confidence scores for the plurality of travel modes meeting a threshold value for the first travel mode, labeling the received digital data corresponding to the trace with the first travel mode.

11. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 10, wherein:

a first metric of the plurality of metrics is a speed metric; and

a second metric of the plurality of metrics is one of a directional metric, an acceleration metric, and a deceleration metric.

12. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 10, wherein:

the set of derived values comprises a set of vectors of quantile values, each vector corresponding to a respective metric of the plurality of metrics, and

calculating the set of derived values for the plurality of metrics comprises calculating at least a first vector of quantile values for a first metric of the plurality of metrics and calculating a second vector of quantile values for a second metric of the plurality of metrics.

13. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 12, wherein calculating the first vector of quantile values for the first metric comprises:

ordering the set of values for the first metric;

determining, for each of a plurality of predetermined quantiles, a quantile value by identifying a value at a position within the ordered set of values that corresponds to the predetermined quantile; and

dividing each quantile value by a maximum value for the first metric to obtain a normalized quantile value; and

storing the normalized quantile values in the first vector of quantile values.

14. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 13, wherein calculating the first confidence score for the first travel mode comprises weighting at least one quantile value higher than a second quantile value.

15. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 10, wherein the computer program code is further for:

receiving raw trace data comprising location data and corresponding time stamps; and

generating the digital data corresponding to the trace for each metric of the plurality of metrics by calculating the set of values corresponding to the metric from the location data and the corresponding time stamps in the raw trace data.

16. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 10, wherein the computer program code is further for training the machine learning model based on a plurality of training sets of digital data each comprising an identification of one of the plurality of travel modes.

17. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 16, wherein training the machine learning model comprises:

determining a set of rules for calculating a confidence score for a particular travel mode of the plurality of travel modes based on the training sets of digital data and the identified travel modes.

18. The non-transitory computer-readable storage containing computer program code executable on a processor of claim 10, wherein the first travel mode is a driving mode, the computer program code further for:

in response to labeling the received digital data corresponding to the trace with the driving mode, adding data associated with the trace to a driving database.

19. A computer system comprising:

a computer processor; and

a non-transitory computer-readable storage containing computer program code for classifying inferred activities of mobile computing devices based upon data representing movement of the devices, the computer program code when executed by the computer processor performing actions comprising;

receiving digital data corresponding to a trace by a device, the digital data comprising a set of values for each of a plurality of metrics;

calculating a set of derived values for the plurality of metrics from the sets of values corresponding to the metrics;

comparing the set of derived values to a machine learning model trained using a plurality of travel modes;

calculating a plurality of confidence scores for the derived values, the plurality of confidence scores corresponding to the plurality of travel modes; and

in response to a confidence score for a first travel mode of the plurality of confidence scores for the plurality of travel modes meeting a threshold value for the first travel mode, labeling the received digital data corresponding to the trace with the first travel mode.

20. The computer system of claim 19, the actions further comprising:

receiving raw trace data comprising location data and corresponding time stamps; and

generating the digital data corresponding to the trace for each metric of the plurality of metrics by calculating the set of values corresponding to the metric from the location data and the corresponding time stamps in the raw trace data.