Granted Patent B1

US 12,391,256 · App. 18/455,167 · Granted Aug 19, 2025

Inventors: Evaline Shin-Tin Tsai (Cupertino, CA); Kieran K. Gupta (San Francisco, CA); Margaret Irene Finch (San Francisco, CA); Matthew Lee Basham (Oakland, CA); Bodecker John DellaMaria (Crystal Lake, IL); Brian Tuan (Cupertino, CA); James Chen (San Francisco, CA); Jason Noah Laska (San Francisco, CA); Jason Frederic Symons (Dublin, CA); John Charles Bicket (Burlingame, CA); Kavya Joshi (San Francisco, CA); Ryan Reading (San Francisco, CA); Sabrina Quinn Shemet (Felton, CA); Sean Kyungmok Bae (San Francisco, CA); Ishaan Kansal (San Francisco, CA)

Assignee: Samsara Inc.

B60W40/04G06V10/82G06V20/46G06V20/56G06F18/214G06V20/44G06V20/582G06V20/597

Abstract

Claims (76)

1. A method comprising:

detecting a start of a trip of a vehicle, the trip comprising a first portion and a second portion;

accessing a sensor data stream generated by a sensor device associated with the vehicle in response to the detecting the start of the trip, the sensor data stream comprising image data;

identifying a user account based on the image data from the sensor device;

accessing a baseline sensor data stream generated by the sensor device in response to the detecting the start of the trip, the baseline sensor data stream comprising baseline image data corresponding to the first portion of the trip;

generating baseline data associated with the user account based on the baseline image data;

generating a training data set based on the baseline data associated with the user account;

training a machine-learned model based on the training data set;

detecting an event, using the machine-learned model, based on a second sensor data stream corresponding to the second portion of the trip; and

causing a notification at a client device based on the event.

2. The method of claim 1 , wherein the detecting the start of the trip includes:

accessing accelerometer data; and

detecting the start of the trip based on the accelerometer data.

3. The method of claim 1 , wherein the detecting the start of the trip includes:

accessing geolocation data; and

detecting the start of the trip based on the geolocation data.

4. The method of claim 1 , wherein the sensor device includes a dash-cam, and the image data comprises a set of image features that depict a gaze and a head pose.

5. The method of claim 4 , wherein the identifying a user account based on the image data from the sensor device comprises:

performing one or more facial recognition techniques on the set of image features.

6. The method of claim 1 , further comprising:

generating a user score associated with the user account based on the event; and

adjusting the user score in response to detecting a subsequent event.

7. The method of claim 6 , further comprising:

determining the user score transgresses a threshold value; and

causing display of a second notification including a user identifier and the user score at the client device based on the user score transgressing the threshold value.

8. A system comprising:

at least one vehicle;

at least one sensor device to generate sensor data comprising a plurality of data streams;

a memory; and

at least one hardware processor to perform operations comprising:

detecting a start of a trip of a vehicle, the trip comprising a first portion and a second portion;

accessing a sensor data stream generated by a sensor device associated with the vehicle in response to the detecting the start of the trip, the sensor data stream comprising image data;

identifying a user account based on the image data from the sensor device;

accessing a baseline sensor data stream generated by the sensor device in response to the detecting the start of the trip, the baseline sensor data stream comprising baseline image data corresponding to the first portion of the trip;

generating baseline data associated with the user account based on the baseline image data;

generating a training data set based on the baseline data associated with the user account;

training a machine-learned model based on the training data set;

detecting an event, using the machine-learned model, based on a second sensor data stream corresponding to the second portion of the trip; and

causing a notification at a client device based on the event.

9. The system of claim 8 , wherein the detecting the start of the trip includes:

accessing accelerometer data; and

detecting the start of the trip based on the accelerometer data.

10. The system of claim 8 , wherein the detecting the start of the trip includes:

accessing geolocation data; and

detecting the start of the trip based on the geolocation data.

11. The system of claim 8 , wherein the sensor device includes a dash-cam, and the image data comprises a set of image features that depict a gaze and a head pose.

12. The system of claim 11 , wherein the identifying a user account based on the image data from the sensor device comprises:

performing one or more facial recognition techniques on the set of image features.

13. The system of claim 8 , wherein the operations further comprise:

generating a user score associated with the user account based on the event; and

adjusting the user score in response to detecting a subsequent event.

14. The system of claim 13 , wherein the operations further comprise:

determining the user score transgresses a threshold value; and

causing display of a second notification including a user identifier and the user score at the client device based on the user score transgressing the threshold value.

15. A non-transitory machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform operations comprising:

detecting a start of a trip of a vehicle, the trip comprising a first portion and a second portion;

accessing a sensor data stream generated by a sensor device associated with the vehicle in response to the detecting the start of the trip, the sensor data stream comprising image data;

identifying a user account based on the image data from the sensor device;

accessing a baseline sensor data stream generated by the sensor device in response to the detecting the start of the trip, the baseline sensor data stream comprising baseline image data corresponding to the first portion of the trip;

generating baseline data associated with the user account based on the baseline image data;

generating a training data set based on the baseline data associated with the user account;

training a machine-learned model based on the training data set;

detecting an event, using the machine-learned model, based on a second sensor data stream corresponding to the second portion of the trip; and

causing a notification at a client device based on the event.

16. The non-transitory machine-readable storage medium of claim 15 , wherein the detecting the start of the trip includes:

accessing accelerometer data; and

detecting the start of the trip based on the accelerometer data.

17. The non-transitory machine-readable storage medium of claim 15 , wherein the detecting the start of the trip includes:

accessing geolocation data; and

detecting the start of the trip based on the geolocation data.

18. The non-transitory machine-readable storage medium of claim 15 , wherein the sensor device includes a dash-cam, and the image data comprises a set of image features that depict a gaze and a head pose.

19. The non-transitory machine-readable storage medium of claim 18 , wherein the identifying a user account based on the image data from the sensor device comprises:

performing one or more facial recognition techniques on the set of image features.

20. The non-transitory machine-readable storage medium of claim 15 , wherein the operations further comprise:

generating a user score associated with the user account based on the event; and

adjusting the user score in response to detecting a subsequent event.

Assignments (2)

Continuity (2)

References Cited (341)

Cited By (1)

US 12,682,753