SYSTEM AND METHOD FOR ADVERSE VEHICLE EVENT DETERMINATION

• US 20180365772A1
• Filed: 06/18/2018
• Published: 12/20/2018
• Est. Priority Date: 06/16/2017
• Status: Active Grant

• Current Assignee: Nauto, Inc.
• Original Assignee: Nauto Global Limited
• Inventors: Thompson, Alex, Sankaran, Shobana, Hornbaker, Charles, Heck, Stefan
• Granted Patent: US 10,453,150 B2
• Time in Patent Office: Days
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: G06Q 50/40 Business processes related to the transportation industry shipping G06Q10/083

Abstract

A method for determining an adverse vehicle event, including: sampling sensor data an onboard vehicle system coupled to an ego-vehicle; at the onboard vehicle system, extracting a set of event parameters from the sensor data, wherein the vehicle event data is associated with a vehicle event occurring within the time interval; computing a loss probability based on the set of event parameters in response to the loss probability exceeding a threshold probability, transforming the set of event parameters into insurance claim data; and automatically transmitting the insurance claim data to an endpoint, wherein the endpoint is determined based on the participant identifier.

Claims

1. A method for automatically determining insurance claim data, comprising:

sampling interior image data over a time interval using an interior-facing camera of an onboard vehicle system rigidly coupled to an ego-vehicle;

contemporaneously sampling exterior image data over the time interval using an exterior-facing camera of the onboard vehicle system;

at a processor of the onboard vehicle system, extracting vehicle event data from the interior image data and the exterior image data, wherein the vehicle event data is associated with a vehicle event occurring within the time interval, wherein extracting the vehicle event data comprises extracting a set of event parameters, including extracting;

a first event parameter comprising a plurality of participant identifiers, wherein each event participant identifier is associated with insurance policy data corresponding to a participant in the vehicle event;

a second event parameter comprising a peak acceleration of the ego-vehicle during the time interval, and

a third event parameter comprising a peak attitudinal shift of the ego-vehicle during the time interval;

calculating a loss probability based on the peak acceleration and the peak attitudinal shift;

in response to the loss probability exceeding a threshold probability;

determining a fault status for each of the plurality of participant identifiers, and

transforming the set of event parameters and the fault status for each of the plurality of participant identifiers into insurance claim data;

prompting a user corresponding to one participant identifier to provide a transmission instruction at a mobile device associated with the user; and

in response to receiving the transmission instruction, automatically transmitting the insurance claim data to an endpoint at a remote computing system.

2. The method of claim 1, wherein the onboard vehicle system further comprises an accelerometer, and further comprising sampling an accelerometer signal from the accelerometer at a time point prior to the time interval, and sampling the interior image data and the exterior image data in response to the accelerometer signal exceeding a threshold value.

3. The method of claim 2, wherein, at the time point, the ego-vehicle is parked.

4. The method of claim 1, wherein calculating the loss probability based on the peak acceleration and the peak attitudinal shift comprises calculating the loss probability using an adverse vehicle event model, wherein the adverse vehicle event model is trained based on a corpus of event parameters extracted from a population of ego-vehicles and corresponding insurance claim data associated with the population of ego-vehicles.

5. The method of claim 1, wherein extracting the peak attitudinal shift of the ego-vehicle comprises:

determining a first position of each object depicted in a first frame of the exterior image data at a first time point within the time interval, wherein the first position is defined in two dimensions relative to a boundary of the frame;

determining a second position of each object depicted in a second frame of the exterior image data at a second time point within the time interval subsequent to the first time point, wherein the second position is defined in two dimensions relative to the boundary of the frame;

calculating the peak attitudinal shift of the ego-vehicle based on a comparison between the first position of each object and the second position of each object.

6. The method of claim 5, wherein extracting the peak acceleration of the ego-vehicle comprises:

determining a first velocity of each object depicted in the first frame of the exterior image data;

determining a second velocity of each object depicted in the second frame of the exterior image data, wherein the first and second velocity are defined in three dimensions relative to the ego-vehicle; and

calculating the peak acceleration of the ego-vehicle based on a comparison between the first velocity of each object and the second velocity of each object.

7. The method of claim 1, further comprising training a claims adjustment module based on vehicle event data extracted from interior image data and exterior image data sampled by one or more onboard vehicle systems of a population of ego-vehicles, and wherein transforming the set of event parameters and the fault status for each of the plurality of participant identifiers into insurance claim data is performed using the claim adjustment module.

8. A method for determining an adverse vehicle event, comprising:

sampling sensor data over a time interval using an onboard vehicle system rigidly coupled to an ego-vehicle;

at a processor of the onboard vehicle system, extracting a set of event parameters from the sensor data, wherein the vehicle event data is associated with a vehicle event occurring within the time interval, wherein extracting the set of event parameters comprises extracting;

a first event parameter comprising a participant identifier,

a second event parameter comprising an acceleration of the ego-vehicle during the time interval, and

a third event parameter comprising an attitudinal shift of the ego-vehicle during the time interval;

computing a loss probability based on the acceleration and the attitudinal shift;

in response to the loss probability exceeding a threshold probability, transforming the set of event parameters into insurance claim data; and

automatically transmitting the insurance claim data to an endpoint, wherein the endpoint is determined based on the participant identifier.

9. The method of claim 8, wherein the sensor data comprises image data, and wherein sampling the image data is performed at a monocular camera of the onboard vehicle system.

10. The method of claim 9, wherein sampling the image data is performed in response to detecting a trigger.

11. The method of claim 10, further comprising sampling an accelerometer signal from an accelerometer of the onboard vehicle system at a time point prior to the time interval, wherein the trigger comprises the accelerometer signal exceeding a threshold value.

12. The method of claim 9, wherein extracting vehicle event data further comprises:

extracting driver behavior data corresponding to a driver of the ego-vehicle from the image data;

determining a fault status associated with the driver based on the driver behavior data; and

in response to the loss probability exceeding the threshold probability, transforming the set of event parameters and the fault status into insurance claim data.

13. The method of claim 12, wherein generating insurance claim data comprises estimating a claim cost based on the loss probability and the fault status.

14. The method of claim 8, wherein extracting the plurality of participant identifiers comprises determining a number of occupants in the ego-vehicle during the time interval.

15. The method of claim 8, further comprising training a claims adjustment module based on vehicle event data extracted from sensor data sampled by one or more onboard vehicle systems of a population of ego-vehicles, and wherein transforming the set of event parameters into the insurance claim data is performed using the claim adjustment module.

16. The method of claim 8, further comprising prompting a user corresponding to one participant identifier of the plurality of participant identifiers to provide a transmission instruction at a mobile device associated with the user, and wherein automatically transmitting the insurance claim data is performed in response to receiving the transmission instruction.

17. The method of claim 9, wherein extracting the attitudinal shift of the ego-vehicle from the image data comprises:

determining a first position of each object depicted in a first frame of the image data at a first time point within the time interval, wherein the first position is defined in two dimensions relative to a boundary of the frame;

determining a second position of each object depicted in a second frame of the image data at a second time point within the time interval subsequent to the first time point, wherein the second position is defined in two dimensions relative to the boundary of the frame;

calculating the attitudinal shift of the ego-vehicle based on a comparison between the first position of each object and the second position of each object.

18. The method of claim 17, wherein extracting the acceleration of the ego-vehicle from the image data comprises:

determining a first velocity of each object depicted in the first frame of the image data;

determining a second velocity of each object depicted in the second frame of the image data, wherein the first and second velocity are defined in three dimensions relative to the ego-vehicle;

calculating the acceleration of the ego-vehicle based on a comparison between the first velocity of each object and the second velocity of each object.

19. The method of claim 18, wherein calculating the loss probability based on the acceleration of the ego-vehicle and the attitudinal shift of the ego-vehicle comprises calculating the loss probability using an adverse vehicle event model, wherein the adverse vehicle event model is trained based on a set of event parameters extracted from a population of ego-vehicles and corresponding insurance claim data.

20. The method of claim 18, wherein the acceleration comprises a single peak acceleration value, and wherein the attitudinal shift comprises a single peak attitudinal shift value.