Vehicle action control
One or more target areas are identified proximate to a moving vehicle. The vehicle can be maneuvered to a target area selected according to a reinforcement learning reward function.
1. A computer comprising a processor and a memory, the memory storing instructions executable by the processor to:
identify one or more target areas proximate to a moving vehicle; and
maneuver the moving vehicle to a target area selected according to a reinforcement learning reward function;
wherein the reward function includes a difference between a current speed and a speed that would result from taking an action affecting movement of the moving vehicle, an amount of time that the moving vehicle has operated without collision in a simulation episode, and a penalty assessed if a lane change results from the action.
2. The computer of claim 1 , wherein the reward function includes a penalty for a safety risk.
3. The computer of claim 1 , wherein the reward function evaluates a relative velocity of a second vehicle to specify a reward.
4. The computer of claim 1 , wherein the reward function evaluates a length of the target area to specify a reward.
5. The computer of claim 1 , wherein the one or more target areas are specified according to a maximum longitudinal distance between vehicles on a roadway.
6. The computer of claim 1 , wherein the one or more target areas are specified according to a boundary of at least one second vehicle.
7. The computer of claim 1 , wherein the one or more target areas include target areas in a plurality of lanes on a roadway.
8. The computer of claim 1 , wherein the instructions further include instructions to determine that the moving vehicle is moving above a predetermined velocity threshold prior to selecting the target area.
9. A method, comprising:
identifying one or more target areas proximate to a moving vehicle; and
maneuvering the moving vehicle to a target area selected according to a reinforcement learning reward function;
wherein the reward function includes a difference between a current speed and a speed that would result from taking an action affecting movement of the moving vehicle, an amount of time that the moving vehicle has operated without collision in a simulation episode, and a penalty assessed if a lane change results from the action.
10. The method of claim 9 , wherein the reward function includes a penalty for a safety risk.
11. The method of claim 9 , wherein the reward function evaluates a relative velocity of a second vehicle to specify a reward.
12. The method of claim 9 , wherein the reward function evaluates a length of the target area to specify a reward.
13. The method of claim 9 , wherein the one or more target areas are specified according to a maximum longitudinal distance between vehicles on a roadway.
14. The method of claim 9 , wherein the one or more target areas are specified according to a boundary of at least one second vehicle.
15. The method of claim 9 , wherein the one or more target areas include target areas in a plurality of lanes on a roadway.
16. The method of claim 9 , further comprising determining that the moving vehicle is moving above a predetermined velocity threshold prior to selecting the target area.
17. A vehicle, comprising:
a sensor providing data about an area around the vehicle; and
a computer comprising a processor and a memory, the memory storing instructions executable by the processor to:
based on data from the sensor, identify one or more target areas proximate to a moving vehicle; and
maneuver the vehicle to a target area selected according to a reinforcement learning reward function;
wherein the reward function includes a difference between a current speed and a speed that would result from taking an action affecting movement of the moving vehicle, an amount of time that the moving vehicle has operated without collision in a simulation episode, and a penalty assessed if a lane change results from the action.
18. The vehicle of claim 17 , wherein the one or more target areas include target areas in a plurality of lanes on a roadway.