Enhanced occupant collision safety system
Various systems and methods for implementing an enhanced occupant collision safety system in a vehicle are described herein. A system for controlling deployment of vehicle safety mechanisms of a host vehicle, includes a sensor array interface to receive acceleration data from an impact sensor; an image processor to receive events from an event camera, the event camera configured to monitor a scene of an interior of the host vehicle that includes an occupant; and an occupant safety system controller to: cluster the events to discriminate between events associated with a background of the scene and events associated with a body part of the occupant; detect a collision based on the acceleration data; verify the collision based on the events associated with the background of the scene; and in response to verifying the collision, cause a collision safety mechanism of the host vehicle to deploy.
1. A system, comprising:
an image processor to receive events from an event camera, the event camera configured to monitor a scene of an interior of a host vehicle that includes an occupant; and
an occupant safety system controller to:
cluster the events to discriminate between events associated with a background of the scene and events associated with a body part of the occupant;
detect a collision based on acceleration data from an impact sensor;
verify the collision based on the events associated with the background of the scene by:
calculating a velocity of the events associated with the background, wherein the events associated with the background are events that are not associated with occupants in the host vehicle; and
determining that the velocity of the events associated with the background matches an axis and magnitude of the collision provided by the acceleration data; and
in response to verifying the collision, cause a collision safety mechanism of the host vehicle to deploy.
2. The system of claim 1 , wherein the acceleration data includes three-axis acceleration data.
3. The system of claim 1 , wherein an event from the event camera is triggered when a luminance change of a pixel of the event camera increases or decreases more than a threshold amount.
4. The system of claim 3 , wherein the event includes an address of the pixel that detected the luminance change, and a time stamp of the luminance change.
5. The system of claim 1 , wherein to cluster the events, the occupant safety system controller is to:
calculate a velocity vector of the events with respect to time;
use the velocity of each event as an error function to generate N clusters; and
remove clusters that represent the background of the scene.
6. The system of claim 1 , wherein the clustering is performed with a neural network that is configured to increase or decrease a number of neurons used depending on a number of roots detected in a derivative of an error function.
7. The system of claim 1 , wherein to detect a collision based on the acceleration data, the occupant safety system controller is to:
compute a jerk based on the acceleration data;
determine that the jerk exceeds a threshold; and
in response to determining that the jerk exceeds the threshold, determine the collision exists.
8. The system of claim 1 , wherein to cause a collision safety mechanism of the host vehicle to deploy, the occupant safety system controller is to actuate an airbag.
9. The system of claim 8 , wherein to actuate the airbag, the occupant safety system controller is to:
calculate a velocity of the body part of the occupant;
determine a magnitude of the velocity of the body part of the occupant; and
actuate the airbag corresponding with the magnitude and velocity of the velocity.
10. A method, comprising:
receiving acceleration data from an impact sensor installed in a host vehicle;
receiving events from an event camera, the event camera configured to monitor a scene of an interior of the host vehicle that includes an occupant;
clustering the events to discriminate between events associated with a background of the scene and events associated with a body part of the occupant;
detecting a collision based on the acceleration data;
verifying the collision based on the events associated with the background of the scene by:
calculating a velocity of the events associated with the background, wherein the events associated with the background are events that are not associated with occupants in the host vehicle; and
determining that the velocity of the events associated with the background matches an axis and magnitude of the collision provided by the acceleration data; and
in response to verifying the collision, cause a collision safety mechanism of the host vehicle to deploy.
11. The method of claim 10 , wherein the acceleration data includes three-axis acceleration data.
12. The method of claim 10 , wherein an event from the event camera is triggered when a luminance change of a pixel of the event camera increases or decreases more than a threshold amount.
13. The method of claim 12 , wherein the event includes an address of the pixel that detected the luminance change, and a time stamp of the luminance change.
14. The method of claim 10 , wherein clustering the events comprises:
calculating a velocity vector of the events with respect to time;
using the velocity of each event as an error function to generate N clusters; and
removing clusters that represent the background of the scene.
15. The method of claim 10 , wherein the clustering is performed with a neural network that is configured to increase or decrease a number of neurons used depending on a number of roots detected in a derivative of an error function.
16. The method of claim 10 , wherein detecting a collision based on the acceleration data comprises:
computing a jerk based on the acceleration data;
determining that the jerk exceeds a threshold; and
in response to determining that the jerk exceeds the threshold, determining the collision exists.
17. The method of claim 10 , wherein causing a collision safety mechanism of the host vehicle to deploy comprises actuating an airbag.
18. The method of claim 17 , wherein actuating the airbag comprises:
calculating a velocity of the body part of the occupant;
determining a magnitude of the velocity of the body part of the occupant; and
actuating the airbag corresponding with the magnitude and velocity of the velocity.
19. At least one non-transitory machine-readable medium including instructions, which when executed by a machine in a host vehicle, cause the machine to perform operations comprising:
receiving acceleration data from an impact sensor;
receiving events from an event camera, the event camera configured to monitor a scene of an interior of the host vehicle that includes an occupant;
clustering the events to discriminate between events associated with a background of the scene and events associated with a body part of the occupant;
detecting a collision based on the acceleration data;
verifying the collision based on the events associated with the background of the scene by:
calculating a velocity of the events associated with the background, wherein the events associated with the background are events that are not associated with occupants in the host vehicle; and
determining that the velocity of the events associated with the background matches an axis and magnitude of the collision provided by the acceleration data; and
in response to verifying the collision, cause a collision safety mechanism of the host vehicle to deploy.
20. The at least one machine-readable medium of claim 19 , wherein an event from the event camera is triggered when a luminance change of a pixel of the event camera increases or decreases more than a threshold amount.
21. The at least one machine-readable medium of claim 20 , wherein the event includes an address of the pixel that detected the luminance change, and a time stamp of the luminance change.
22. The at least one machine-readable medium of claim 19 , wherein the instructions for clustering the events comprise instructions for:
calculating a velocity vector of the events with respect to time;
using the velocity of each event as an error function to generate N clusters; and
removing clusters that represent the background of the scene.
23. The at least one machine-readable medium of claim 19 , wherein the clustering is performed with a neural network that is configured to increase or decrease a number of neurons used depending on a number of roots detected in a derivative of an error function.