US 11,884,225 B2
Methods and systems for point of impact detection
Yuting Qi, Lexington, MA (US); Rizki Syarif, Lexington, MA (US); Edward J. Gramer, Dover, MA (US); and Cornelius F. Young, Needham, MA (US)
Assigned to Cambridge Mobile Telematics Inc., Cambridge, MA (US)
Filed by Cambridge Mobile Telematics Inc., Cambridge, MA (US)
Filed on Oct. 20, 2021, as Appl. No. 17/506,566.
Claims priority of provisional application 63/165,659, filed on Mar. 24, 2021.
Claims priority of provisional application 63/159,948, filed on Mar. 11, 2021.
Claims priority of provisional application 63/094,824, filed on Oct. 21, 2020.
Prior Publication US 2022/0118931 A1, Apr. 21, 2022
Int. Cl. B60R 21/0132 (2006.01); G08G 1/01 (2006.01); G06N 3/08 (2023.01)
CPC B60R 21/01338 (2014.12) [G06N 3/08 (2013.01); G08G 1/0112 (2013.01); G08G 1/0133 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving a first set of sensor measurements, wherein sensor measurements of the first set of sensor measurements are received from a mobile device that includes an accelerometer sensor and a global positioning service (GPS) sensor, and wherein the sensor measurements of the first set of sensor measurements are associated with vehicle collisions;
extracting, using a pattern analysis model, a set of motifs from the first set of sensor measurements, each motif of the set of motifs corresponding to a pattern of sensor measurements that occur during a vehicle collision;
training a machine-learning model using the set of motifs;
collecting, during a collision involving a vehicle, a second set of sensor measurements using a mobile device including at least one of accelerometer sensor or a GPS sensor, wherein the mobile device is positioned within the vehicle;
executing the machine-learning model using the second set of sensor measurements to predict a point of impact on the vehicle; and
transmitting an indication of the point of impact to a remote device.