Method and system for identifying vehicle collisions using sensor data

• US 10,829,071 B1
• Filed: 07/08/2019
• Issued: 11/10/2020
• Est. Priority Date: 07/13/2015
• Status: Active Grant

• Current Assignee: State Farm Mutual Automobile Insurance Company (State Farm)
• Original Assignee: State Farm Mutual Automobile Insurance Company (State Farm)
• Inventors: Christensen, Scott T., Dosher, David James, Menon, Sunish Shreenarayan, Dunham, Susan M.
• Primary Examiner(s): Lau, Hoi C
• Application Number:
  US16/504,658
• Time in Patent Office: 491 Days
• CPC Class Codes: B60Q 5/006 indicating risk of collision between vehicles or with pedestrians
• CPC Class Codes: B60R 2021/0027 Post collision measures, e.g. notifying emergency services
• CPC Class Codes: B60R 2021/01322 comprising variable thresholds, e.g. depending from other collision parameters
• CPC Class Codes: B60R 21/0132 responsive to vehicle motion parameters , e.g. to vehicle longitudinal or transversal deceleration or speed value
• CPC Class Codes: B60R 21/0136 responsive to actual contact with an obstacle , e.g. to vehicle deformation, bumper displacement or bumper velocity relative to the vehicle monitoring crash strength in combination with passenger seat detection B60R21/01558
• CPC Class Codes: B60W 2040/0818 Inactivity or incapacity of driver
• CPC Class Codes: G07C 5/008 communicating information to a remotely located station transmission systems for measured values G08C
• CPC Class Codes: G08G 1/205 Indicating the location of the monitored vehicles as destination, e.g. accidents, stolen, rental
• CPC Class Codes: H04L 51/046 Interoperability with other network applications or services
• CPC Class Codes: H04W 4/90 Services for handling of emergency or hazardous situations, e.g. earthquake and tsunami warning systems [ETWS]
• CPC Class Codes: H04W 76/50 for emergency connections

Abstract

A method and system may identify vehicle collisions in real-time or at least near real-time based on statistical data collected from previous vehicle collisions. The statistical data may be used to train a machine learning model for identifying whether a portable computing device is in a vehicle collision based on sensor data from the portable computing device. The machine learning model may be trained based on a first subset of sensor data collected from vehicle trips involved in vehicle collisions and a second subset of sensor data collected from vehicle trips not involved in vehicle collisions. When a current set of sensor data is obtained from a portable computing device in a vehicle, the current set of sensor data is compared to the machine learning model to determine whether the portable computing device is in a vehicle involved in a vehicle collision.

Claims

1. A server device for automatically identifying vehicle collisions using sensor data, the server device comprising:

one or more processors; and

a non-transitory computer-readable memory coupled to the one or more processors, and the one or more sensors, and storing thereon instructions that, when executed by the one or more processors, cause the server device to;

obtain a set of training data including a plurality of sets of sensor data from portable computing devices collected during vehicle trips, the plurality of sets of sensor data including a first subset of sensor data corresponding to vehicle trips involved in vehicle collisions and a second subset of sensor data corresponding to vehicle trips not involved in vehicle collisions, each set of sensor data including at least one of;

position data, speed data, acceleration data, rotation data, pressure data, or sound data, the second subset of sensor data including sensor data indicative of at least one of the portable computing devices being manipulated by a user to change an orientation of the at least one portable computing device;

generate, using the training data, a statistical model for identifying whether a portable computing device is in a vehicle involved in a vehicle collision based on sensor data from the portable computing device, wherein the statistical model is adjusted in accordance with the sensor data indicative of the at least one portable computing device being manipulated by the user;

obtain a current set of sensor data from a portable computing device;

compare the current set of sensor data from the portable computing device to the statistical model to determine whether the portable computing device is in a vehicle involved in a vehicle collision; and

determine that a vehicle collision has occurred involving the vehicle including the portable computing device based on the comparison.

2. The server device of claim 1, wherein the instructions cause the server device to generate the statistical model using one or more machine learning techniques.

3. The server device of claim 2, wherein the one or more machine learning techniques include at least one of naï

ve Bayes, random forests, boosting, decision trees, logistic regression, or k-nearest neighbors.

4. The server device of claim 1, wherein the instructions further cause the server device to generate, using the training data, another statistical model for determining a type or severity of a vehicle collision.

5. The server device of claim 1, wherein each of the plurality of sets of sensor data are collected at a plurality of predetermined time intervals during a particular vehicle trip.

6. The server device of claim 1, wherein each set of sensor data further includes a sample rate at which the sensor data is collected.

7. The server device of claim 1, wherein in response to determining that a vehicle collision has occurred involving the vehicle, the instructions further cause the server device to transmit an emergency notification to emergency personnel, wherein the emergency notification includes a location of the user and at least some of the sensor data.

8. The server device of claim 1, wherein in response to determining that a vehicle collision has occurred involving the vehicle, the instructions further cause the server device to transmit an emergency notification to one or more emergency contacts for the user.

9. The server device of claim 1, wherein in response to determining that a vehicle collision has occurred involving the vehicle, the instructions further cause the server device to provide a collision request for display on a user interface asking the user to verify that a vehicle collision occurred.

10. A computer-implemented method for automatically identifying vehicle collisions using sensor data, the method executed by one or more processors programmed to perform the method, the method comprising:

obtaining, by one or more processors, a set of training data including a plurality of sets of sensor data from portable computing devices collected during vehicle trips, the plurality of sets of sensor data including a first subset of sensor data corresponding to vehicle trips involved in vehicle collisions and a second subset of sensor data corresponding to vehicle trips not involved in vehicle collisions, each set of sensor data including at least one of;

position data, speed data, acceleration data, rotation data, pressure data, or sound data, the second subset of sensor data including sensor data indicative of at least one of the portable computing devices being manipulated by a user to change an orientation of the at least one portable computing device;

generating, by the one or more processors using the training data, a statistical model for identifying whether a portable computing device is in a vehicle involved in a vehicle collision based on sensor data from the portable computing device, wherein the statistical model is adjusted in accordance with the sensor data indicative of the at least one portable computing device being manipulated by the user;

obtaining, by the one or more processors, a current set of sensor data from a portable computing device;

comparing, by the one or more processors, the current set of sensor data from the portable computing device to the statistical model to determine whether the portable computing device is in a vehicle involved in a vehicle collision; and

determining, by the one or more processors, that a vehicle collision has occurred involving the vehicle including the portable computing device based on the comparison.

11. The method of claim 10, wherein the statistical model is generated using one or more machine learning techniques.

12. The method of claim 11, wherein the one or more machine learning techniques include at least one of naï

ve Bayes, random forests, boosting, decision trees, logistic regression, or k-nearest neighbors.

13. The method of claim 10, further comprising generating, using the training data, another statistical model for determining a type or severity of a vehicle collision.

14. The method of claim 10, wherein each of the plurality of sets of sensor data are collected at a plurality of predetermined time intervals during a particular vehicle trip.

15. The method of claim 10, wherein each set of sensor data further includes a sample rate at which the sensor data is collected.

16. The method of claim 10, wherein in response to determining that a vehicle collision has occurred involving the vehicle, transmitting, by the one or more processors, an emergency notification to emergency personnel, wherein the emergency notification includes a location of the user and at least some of the sensor data.

17. The method of claim 10, wherein in response to determining that a vehicle collision has occurred involving the vehicle, transmitting, by the one or more processors, an emergency notification to one or more emergency contacts for the user.

18. The method of claim 10, wherein in response to determining that a vehicle collision has occurred involving the vehicle, providing, by the one or more processors, a collision request for display on a user interface asking the user to verify that a vehicle collision occurred.