Back to results
US 12,722,643 B1Granted
Drowsy sensor fusion
AssigneeLytx, Inc.
InventorVincent Nguyen
Life of this patent
Estimating the term…
Abstract
A system for indicating a drowsy driver state comprises one or more interfaces and a processor. The one or more interfaces is configured to receive a set of driver images and a group of lane images over a time period and one or more lane departure indications. The processor is configured to: 1) determine a closed eyes statistic using a classifier based at least in part on the set of driver images; 2) determine a lane departure statistic using a classifier based at least in part on the group of lane images one or more lane departure indications; 3) determine a drowsy indication based at least in part on the closed eyes statistic and the lane departure statistic; and 4) provide the drowsy indication.
Claims20
A system, comprising:
- a vehicle event recorder mounted to a vehicle, wherein the vehicle event recorder comprises:
- one or more interfaces configured to:
- receive over a time period a set of driver images from one or more internal cameras, a group of lane images from one or more external cameras, and sensor signals from one or more sensors;
- and
- a hardware processor configured to:
- determine a closed eyes statistic using a first classifier based on a driver image of the set of driver images;
- determine a lane departure statistic using a second classifier based at least in part on the group of lane images;
- determine a sensor statistic using a third classifier based at least in part on the sensor signals;
- determine a current probability of a driver being drowsy based at least in part on the closed eyes statistic, the lane departure statistic, and the sensor statistic;
- in response to the probability exceeding a first drowsiness threshold, provide a drowsy indication, wherein the drowsy indication causes a warning to the driver using an audio alert and/or a visual alert;
- in response to the probability not exceeding the first drowsiness threshold and exceeding a second drowsiness threshold, adjust a storage threshold for storing events on the vehicle event recorder to an adjusted storage threshold, wherein the storage threshold comprises at least one of a distance to vehicle ahead, a speed change, and a duration blinkers are kept on;
- and
- store the events on the vehicle event recorder based at least in part on the adjusted storage threshold.
10 more elements
A method, comprising:
- receiving over a time period, by one or more interfaces of a vehicle event recorder mounted to a vehicle, a set of images from one or more internal cameras, a group of lane images from one or more external cameras, and sensor signals from one or more sensors;
- determining, using a hardware processor of the vehicle event recorder, a closed eyes statistic using a first classifier based on a driver image of the set of driver images;
- determining a lane departure statistic using a second classifier based at least in part on the group of lane images;
- determining a sensor statistic using a third classifier based at least in part on the sensor signals;
- determining a current probability of a driver being drowsy based at least in part on the closed eyes statistic, the lane departure statistic, and the sensor statistic;
- in response to the probability exceeding a first drowsiness threshold, providing a drowsy indication, wherein the drowsy indication causes a warning to the driver using an audio alert and/or a visual alert;
- in response to the probability not exceeding the first drowsiness threshold and exceeding a second drowsiness threshold, adjusting a storage threshold for storing events on the vehicle event recorder to an adjusted storage threshold, wherein the storage threshold comprises at least one of a distance to vehicle ahead, a speed change, and a duration blinkers are kept on;
- and
- storing the events on the vehicle event recorder based at least in part on the adjusted storage threshold.
6 more elements
A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:
- receiving over a time period, by one or more interfaces of a vehicle event recorder mounted to a vehicle, a set of images from one or more internal cameras, a group of lane images from one or more external cameras, and sensor signals from one or more sensors;
- determining, using a hardware processor of the vehicle event recorder, a closed eyes statistic using a first classifier based on a driver image of the set of driver images;
- determining a lane departure statistic using a second classifier based at least in part on the group of lane images;
- determining a sensor statistic using a third classifier based at least in part on the sensor signals;
- determining a current probability of a driver being drowsy based at least in part on the closed eyes statistic, the lane departure statistic, and the sensor statistic;
- in response to the probability exceeding a first drowsiness threshold, providing a drowsy indication, wherein the drowsy indication causes a warning to the driver using an audio alert and/or a visual alert;
- in response to the probability not exceeding the first drowsiness threshold and exceeding a second drowsiness threshold, adjusting a storage threshold for storing events on the vehicle event recorder to an adjusted storage threshold, wherein the storage threshold comprises at least one of a distance to vehicle ahead, a speed change, and a duration blinkers are kept on;
- and
- storing the events on the vehicle event recorder based at least in part on the adjusted storage threshold.
6 more elements
Dependencies are inferred from the claim text.