US 12,014,423 B1
Using vehicle data, geographic area type data, and vehicle collision data in determining an indication of whether a vehicle in a vehicle collision is a total loss
Christina P. Mullen, Normal, IL (US); Jeffrey S. Myers, Normal, IL (US); Andrew Karl Pulkstenis, Mahomet, IL (US); Stephen Russell Prevatt, Normal, IL (US); and Robert T. Trefzger, Bloomington, IL (US)
Assigned to STATE FARM MUTUAL AUTOMOBILE INSURANCE COMPANY, Bloomington, IL (US)
Filed by STATE FARM MUTUAL AUTOMOBILE INSURANCE COMPANY, Bloomington, IL (US)
Filed on Jun. 20, 2017, as Appl. No. 15/628,367.
Claims priority of provisional application 62/358,979, filed on Jul. 6, 2016.
Claims priority of provisional application 62/353,409, filed on Jun. 22, 2016.
Int. Cl. G06Q 40/02 (2023.01); G06Q 40/08 (2012.01); G06F 11/00 (2006.01); G06Q 40/04 (2012.01)
CPC G06Q 40/08 (2013.01) 20 Claims
OG exemplary drawing
 
1. A computer-implemented method of determining an indication of whether a vehicle in a crash is a total loss, the method comprising:
in a first instance:
receiving, by one or more processors, geolocation data, image data re presenting at least one image of an exterior of a vehicle, the vehicle having been in a crash, the image data being image data captured after the crash, and at least one of (i) sensor data associated with the crash, the sensor data including acceleration data associated with the crash or (ii) data indicative of a direction of a crash force exerted upon the vehicle during the crash;
determining, based at least upon the at least one of (i) the sensor data or (ii) the data indicative of the direction of the crash force, to analyze whether the vehicle is a total loss using a total loss machine learning algorithm;
determining, by the one or more processors, an indication of population density in an area in which the crash occurred using a population machine learning algorithm and based upon at least the geolocation data;
determining, by the one or more processors, that the image data should be analyzed using object recognition techniques;
determining, by the one or more processors and using the image machine learning algorithm that utilizes object recognition techniques, at least one of (i) a make of the vehicle, (ii) a model of the vehicle, or (iii) a year of the vehicle based at least upon the image data in response to determining that the image data should be analyzed using object recognition techniques; and
determining, by the one or more processors and using the total loss machine learning algorithm, the indication of whether the vehicle is a total loss as a result of the crash based upon (i) the at least one of (a) the sensor data associated with the crash or (b) the data indicative of the direction of the crash force exerted upon the vehicle during the crash, (ii) the indication of population density in the area in which the crash occurred, and (iii) the at least one of (a) the make of the vehicle, (b) the model of the vehicle, or (c) the year of the vehicle;
wherein the total loss machine learning algorithm is trained, by the one or more processors and using at least depersonalized historical claim data, to determine whether the vehicle is a total loss; and
in a second instance:
receiving, by one or more processors, second geolocation data and second image data representing at least one second image of a second exterior of a second vehicle, the second vehicle having been in a second crash, the second image data being image data captured after the second crash;
determining, based at least upon one of (i) an indication that one or more sensors associated with the second vehicle are not operative or (ii) an indication that a second direction of a second crash force exerted upon the second vehicle cannot be retrieved, to transmit data associated with the second crash to a user; and
transmitting the data associated with the second crash to the user.