Vehicle-based intervention based on modeling driver performance
A vehicle-based safety intervention system receives sensor data collected or generated by an onboard computing system of a vehicle. The sensor data is divided into a plurality of blocks, each of the blocks having a duration. A driver behavioral model is applied to one or more of the plurality of blocks to generate one or more driver behavioral parameters. A trend of the one or more driver behavioral parameters is extracted from the plurality of blocks. Based on the extracted trend, it is determined that a driver's performance when operating the vehicle is unsatisfactory or will be unsatisfactory in the future. A vehicle-based intervention is generated based on the determination that the driver's performance is unsatisfactory or will be unsatisfactory in the future.
1. A method of performing a vehicle-based safety intervention, comprising:
receiving sensor data collected or generated by an onboard computing system of a vehicle;
dividing the sensor data into a plurality of blocks, each of the blocks having a duration;
applying a driver behavioral model to one or more of the plurality of blocks to generate a plurality of driver behavioral parameters for each block;
determining that a driver's performance when operating the vehicle is unsatisfactory or will be unsatisfactory in the future; and
generating a vehicle-based intervention based on the determination that the driver's performance is unsatisfactory or will be unsatisfactory in the future;
wherein the plurality of driver behavioral parameters includes a delay corresponding to a reaction time of the driver, a corrective force proportional value corresponding to a magnitude with which the driver takes corrective action and a damping value corresponding to a modulation of the corrective action; and
wherein the determination that the driver's performance is unsatisfactory or will be unsatisfactory in the future is based on a determination that a combination of the delay, the corrective force and the damping value exceeds a multidimensional bound.
2. The method of claim 1 , wherein the driver behavioral model comprises one or more of:
the driver's lateral behavior when operating the vehicle; or
the driver's longitudinal vehicle behavior when operating the vehicle.
3. The method of claim 2 , wherein the driver's longitudinal behavior comprises at least one of:
maintaining a predetermined distance from a leading vehicle; or
maintaining an approximately same velocity as the leading vehicle.
4. The method of claim 2 , wherein the driver's lateral behavior comprises:
maintaining position of the vehicle in a driving lane.
5. The method of claim 1 , wherein
the determining that a driver's performance when operating the vehicle is unsatisfactory or will be unsatisfactory in the future further comprises:
extracting a trend-over-time model of the plurality of driver behavioral parameters for each of the plurality of blocks;
extrapolating the trend-over-time model to identify a future plurality of blocks; and
comparing the future plurality of blocks with the predefined threshold.
6. The method of claim 1 , wherein the corrective action comprises one or more of:
controlling an accelerator of the vehicle;
controlling a brake of the vehicle; or
controlling a steering control of the vehicle.
7. The method of claim 1 , further comprising:
determining, based on comparing each of the plurality of behavioral parameters with a respective predefined threshold, that the driver's performance when operating the vehicle is satisfactory.
8. A system of performing a vehicle-based safety intervention, comprising:
a processor;
a memory in communication with the processor, the memory storing a plurality of instructions executable by the processor to cause the system to:
receive sensor data collected or generated by an onboard computing system of a vehicle;
divide the sensor data into a plurality of blocks, each of the blocks having a duration;
apply a driver behavioral model to one or more of the plurality of blocks to generate a plurality of driver behavioral parameters for each block;
determine that a driver's performance when operating the vehicle is unsatisfactory or will be unsatisfactory in the future; and
generate a vehicle-based intervention based on the determination that the driver's performance is unsatisfactory or will be unsatisfactory in the future;
wherein the plurality of driver behavioral parameters includes a delay corresponding to a reaction time of the driver, a corrective force proportional value corresponding to a magnitude with which the driver takes corrective action and a damping value corresponding to a modulation of the corrective action; and
wherein the determination that the driver's performance is unsatisfactory or will be unsatisfactory in the future is based on a determination that a combination of the delay, the corrective force and the damping value exceeds a multidimensional bound.
9. The system of claim 8 , wherein the driver behavioral model comprises one or more of:
the driver's lateral behavior when operating the vehicle; or
the driver's longitudinal vehicle behavior when operating the vehicle.
10. The system of claim 9 , wherein the driver's longitudinal behavior comprises at least one of:
maintaining a predetermined distance from a leading vehicle; or
maintaining an approximately same velocity as the leading vehicle.
11. The system of claim 9 , wherein the driver's lateral behavior comprises:
maintaining position of the vehicle in a driving lane.
12. The system of claim 8 , wherein the memory further comprises instructions executable by the processor to cause the system to:
identify, based on the sensor data, a decreasing variance in the driver's head yaw over time or an increasing average forward pitch angle.
13. The system of claim 8 , wherein the corrective action comprises one or more of:
controlling an accelerator of the vehicle;
controlling a brake of the vehicle; or
controlling a steering control of the vehicle.
14. The system of claim 8 , wherein
the memory further comprises instructions executable by the processor to cause the system to, during the determination that a driver's performance when operating the vehicle is unsatisfactory or will be unsatisfactory in the future:
extract a trend-over-time model of the plurality of driver behavioral parameters for each of the plurality of blocks;
extrapolate the trend-over-time model to identify a future plurality of blocks; and
compare the future plurality of blocks with the predefined threshold.