Systems and methods for estimating movements of a vehicle using a mobile device

• US 9,360,323 B2
• Filed: 02/13/2015
• Issued: 06/07/2016
• Est. Priority Date: 02/17/2014
• Status: Active Grant

• Current Assignee: Tourmaline Labs, Inc.
• Original Assignee: Tourmaline Labs, Inc.
• Inventors: Grokop, Leonard H.
• Primary Examiner(s): Anwari, Maceeh
• Application Number:
  US14/622,502
• Publication Number:
  US 20150233718A1
• Time in Patent Office: 480 Days
• Field of Search: 701/501
• US Class Current: 1/1
• CPC Class Codes: G01C 21/165 combined with non-inertial navigation instruments
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: H04W 4/027 using movement velocity, acceleration information
• CPC Class Codes: H04W 64/006 with additional information processing, e.g. for direction or speed determination

Abstract

A method of computing an orientation of a mobile device in a vehicle includes collecting accelerometer data from the mobile device. The accelerometer data are subdivided into frames, each frame being a quantity of time in which a sample data point is taken. Statistics are calculated for each frame, including (i) a mean and (ii) a standard deviation of the magnitude of the acceleration vectors. Device usage delimiters, marking start and end points of a coherent block, are computed, the coherent block being consecutive frames in which the mobile device stays in the same orientation relative to the vehicle. A gravity vector is estimated using the statistics in coherent blocks. A nullspace is computed from the gravity vector, the nullspace being a plane orthogonal to the gravity vector. The accelerometer data is projected onto the nullspace, resulting in an estimated orientation of the mobile device in the vehicle.

Claims

1. A method of computing an orientation of a mobile device in a vehicle, the method comprising:

collecting accelerometer data from the mobile device traveling in the vehicle, the accelerometer data comprising acceleration vectors with acceleration values for three orthogonal axes;

subdividing the accelerometer data into frames, each frame being a quantity of time in which a sample data point is taken;

calculating statistics for each frame, using a hardware computing processor, the statistics comprising (i) a mean of the acceleration vectors and (ii) a standard deviation of a magnitude of the acceleration vectors;

computing device usage delimiters, using a hardware computing processor, the device usage delimiters marking start and end points of a coherent block, the coherent block comprising consecutive frames in which the mobile device stays in the same orientation relative to the vehicle;

estimating a gravity vector, using the statistics from the accelerometer data in coherent blocks;

computing a nullspace from the gravity vector, the nullspace being a plane orthogonal to the gravity vector; and

projecting the accelerometer data onto the nullspace, the projecting resulting in an estimated orientation of the mobile device in the vehicle.

2. The method of claim 1 further comprising calibrating the accelerometer, using the statistics from the accelerometer data.

3. The method of claim 1 wherein the computing device usage delimiters comprises:

calculating a similarity metric, the similarity metric being a normalized inner product of the acceleration vectors in adjacent frames; and

determining if the similarity metric meets a predetermined criteria.

4. The method of claim 3 wherein the estimating a gravity vector comprises excluding incoherent blocks, an incoherent block being consecutive frames in which the similarity metric fails to meet the predetermined criteria and in which the incoherent block has a duration less than a predetermined threshold.

5. The method of claim 4 wherein for the incoherent blocks, other sensor data from the mobile device is used in addition to the accelerometer data to compute the orientation of the mobile device.

6. The method of claim 5 wherein the other sensor data comprises gyroscope or magnetometer data.

7. The method of claim 1 wherein the estimating a gravity vector comprises summing the mean accelerometer values over a window of frames, the mean accelerometer values being exponentially weighted according to the standard deviation of the magnitude of the acceleration vectors, wherein smaller standard deviations are given larger weightings.

8. The method of claim 1 further comprising collecting global positioning system (GPS) data during a plurality of brief segments of time to obtain speed and position values.

9. The method of claim 8 further comprising:

obtaining longitudinal and lateral accelerometer data from the projecting of the accelerometer data onto the nullspace; and

correlating longitudinal and lateral accelerometer data with longitudinal G-forces estimates from the GPS data.

10. The method of claim 8 further comprising setting data unusable flags for frames that are delimited as incoherent blocks, or that have insufficient longitudinal G-force GPS data.

11. The method of claim 8 further comprising interpolating speed estimates for an interval between GPS time segments, by integrating longitudinal acceleration data over the interval, wherein the longitudinal acceleration data is relative to the nullspace.

12. The method of claim 8 further comprising:

computing a plurality of nearest route segments and associated distances, for each position value;

computing an optimal path of route segments such that each route segment is associated with one position value, and such that if a first position value is obtained from an earlier segment of time than a second position value, then the route segment associated with the first position value occurs earlier in the route than the route segment associated with the second position value; and

pinning each position value to the nearest point on its associate route segment.

13. The method of claim 1 further comprising:

storing the statistics, after the statistics are calculated; and

discarding the accelerometer data that was collected from the mobile device.

14. A method of inferring movements of a vehicle, the method comprising:

collecting accelerometer data from the mobile device traveling in the vehicle, the accelerometer data comprising acceleration vectors with acceleration values for three orthogonal axes;

subdividing the accelerometer data into frames, each frame being a quantity of time in which a sample data point is taken;

calculating statistics for each frame, using a hardware computing processor, the statistics comprising (i) a mean of the acceleration vectors and (ii) a standard deviation of a magnitude of the acceleration vectors;

computing device usage delimiters, using a hardware computing processor, the device usage delimiters marking start and end points of a coherent block, the coherent block comprising consecutive frames in which the mobile device stays in the same orientation relative to the vehicle;

estimating a gravity vector, using the statistics from the accelerometer data in coherent blocks;

computing a nullspace, using the gravity vector, the nullspace being a plane orthogonal to the gravity vector;

projecting the accelerometer data onto the nullspace, wherein the nullspace provides a forward direction of vehicle movement;

collecting global positioning system (GPS) data during a plurality of brief segments of time to obtain speed and position values;

correlating longitudinal and lateral accelerometer data with longitudinal G-forces estimates from the GPS data, the longitudinal and lateral accelerometer data being relative to the nullspace;

correcting for speed drift by (a) identifying periods of time when the vehicle is stationary, based on the GPS position values, and (b) setting G-force values to zero for those periods; and

determining vehicle speed estimates, comprising calculating vehicle speed estimates for intervals between the segments of collecting GPS data.

15. The method of claim 14 further comprising estimating a vehicle position by fusing the vehicle speed estimates with route and road geometry information.

16. The method of claim 15 further comprising:

computing a route taken by the vehicle using the estimated vehicle positions and the road geometry information; and

interpolating a position of the vehicle during segments of time in which the GPS position data is not collected on the mobile device, using the vehicle speed estimates, the estimated vehicle positions, and the computed route taken by the vehicle.

17. The method of claim 16 wherein the interpolating the position of the vehicle comprises:

finding points on the computed route associated with each interpolated position; and

interpolating a distance of the vehicle along the route, between the found points.

18. The method of claim 17 wherein the finding points comprises:

computing a plurality of nearest route segments and associated distances, for each position value;

computing an optimal path of route segments such that each route segment is associated with one position value, and such that if a first position value is obtained from an earlier segment of time than a second position value, then the route segment associated with the first position value occurs earlier in the route than the route segment associated with the second position value; and

pinning each position value to the nearest point on its associated route segment.

19. The method of claim 14, wherein the computing device usage delimiters comprises:

calculating a similarity metric, the similarity metric being a normalized inner product of the acceleration vectors in adjacent frames; and

determining if the similarity metric meets a predetermined criteria.

20. The method of claim 19, wherein the estimating a gravity vector comprises excluding incoherent blocks, an incoherent block being consecutive frames in which the similarity metric fails to meet the predetermined criteria and which have a duration less than a predetermined threshold.