Vehicle driving system

Patent metadata

Field	Value
Publication number	US-10766487-B2
Application number	US-201816102678-A
Country	US
Kind code	B2
Filing date	Aug 13, 2018
Priority date	Aug 13, 2018
Publication date	Sep 8, 2020
Grant date	Sep 8, 2020

Abstract

Autonomous driving systems and methods include a controller of a first vehicle configured to detect an operating state of a second vehicle proximate the first vehicle, and to predict, based on the operating state of the second vehicle, a potential behavior for the second vehicle that optimizes a cost function from the perspective of the second vehicle. The controller then controls the first vehicle to avoid a collision with the second vehicle assuming the second vehicle operates according to the potential behavior.

First claim

What is claimed is:

1. An autonomous driving system of a first vehicle comprising: a controller configured to detect an operating state of a second vehicle traveling proximate the first vehicle; predict, at the first vehicle, a first potential behavior of the second vehicle based on the operating state of the second vehicle and from a perspective of the second vehicle, wherein the prediction comprises determining a plurality of characteristics of a cost function from a point of view of the second vehicle and optimizing the cost function from the point of view of the second vehicle, wherein the plurality of characteristics pertain to motion of the second vehicle; and control the first vehicle to avoid a collision with the second vehicle assuming the second vehicle operates according to the first potential behavior.

2. The autonomous driving system of claim 1 , wherein the first potential behavior includes a first probability based on the cost function, and the controller is configured to:
apply the operating state of the second vehicle to a trained behavior model to predict a second potential behavior for the second vehicle, the second potential behavior including a second probability based on the trained behavior model;
compare the first probability and the second probability; and
control the first vehicle to avoid colliding with the second vehicle based on the comparison.

3. The autonomous driving system of claim 2 , wherein the first potential behavior includes a first trajectory for the second vehicle, and the second potential behavior includes a second trajectory for the second vehicle that differs from the first trajectory.

4. The autonomous driving system of claim 2 , wherein the controller is configured to:
detect an operating state of a third vehicle traveling proximate the first vehicle;
apply the operating state of the third vehicle to the trained behavior model to predict a third potential behavior for the third vehicle; and
predict the first potential behavior for the second vehicle that optimizes the cost function from the perspective of the second vehicle based on the third potential behavior predicted for the third vehicle.

5. The autonomous driving system of claim 1 , wherein the controller is configured to:
generate, based on the operating state of the second vehicle, candidate trajectories for the second vehicle, each of the candidate trajectories having a cost according to the cost function; and
select, for the first potential behavior, the candidate trajectory having a lowest cost.

6. The autonomous driving system of claim 5 , wherein the controller is configured to:
identify a travel context for the first vehicle and the second vehicle;
retrieve a goal set based on the travel context; and
generate the candidate trajectories based on the goal set.

7. The autonomous driving system of claim 6 , wherein the controller is configured to, prior to generating the candidate trajectories, generate candidate trajectory endpoints for the second vehicle based on the goal set, wherein each of the candidate trajectories is to a different one of the candidate trajectory endpoints and optimizes the cost function for the candidate trajectory endpoint.

8. The autonomous driving system of claim 1 , wherein the controller is configured to:
retrieve map data relating to a geographic location of the first vehicle; and
predict the first potential behavior for the second vehicle that optimizes the cost function from the perspective of the second vehicle based on the map data.

9. The autonomous driving system of claim 8 , wherein the controller is configured to:
detect third vehicles traveling proximate the first vehicle;
select, based on the map data, a subset of the third vehicles including each of the third vehicles that is located less than a lane width from the first vehicle, the subset of the third vehicles including the second vehicle;
predict, for each third vehicle in the subset and not for each third vehicle outside the subset, a potential behavior for the third vehicle that optimizes the cost function from the perspective of the third vehicle; and
control the first vehicle to avoid colliding with each third vehicle assuming each third vehicle operates according to the potential behavior predicted for the third vehicle.

10. The autonomous driving system of claim 1 , wherein the controller is configured to:
determine a class of the second vehicle; and
predict the first potential behavior for the second vehicle that optimizes the cost function from the perspective of the second vehicle based on the class.

11. A controller for an autonomous driving system of a first vehicle, the controller comprising: at least one processor; and a memory storing instructions that, upon execution by the at least one processor, causes the at least one processor to; detect an operating state of a second vehicle traveling proximate the first vehicle; predict, at the first vehicle, a first potential behavior of the second vehicle based on the operating state of the second vehicle and from a perspective of the second vehicle, wherein the prediction comprises determining a plurality of characteristics of a cost function from a point of view of the second vehicle and optimizing the cost function from the point of view of the second vehicle, wherein the plurality of characteristics pertain to motion of the second vehicle; and control the first vehicle to avoid a collision with the second vehicle assuming the second vehicle operates according to the first potential behavior.

12. The controller of claim 11 , wherein the first potential behavior includes a first probability based on the cost function, and the instructions upon execution cause the at least one processor to:
apply the operating state of the second vehicle to a trained behavior model to predict a second potential behavior for the second vehicle, the second potential behavior including a second probability based on the trained behavior model;
compare the first probability and the second probability; and
control the first vehicle to avoid colliding with the second vehicle based on the comparison.

13. The controller of claim 12 , wherein the first potential behavior includes a first trajectory for the second vehicle, and the second potential behavior includes a second trajectory for the second vehicle that differs from the first trajectory.

14. The controller of claim 12 , wherein the instructions upon execution cause the at least one processor to:
detect an operating state of a third vehicle traveling proximate the first vehicle;
apply the operating state of the third vehicle to the trained behavior model to predict a third potential behavior for the third vehicle; and
predict the first potential behavior for the second vehicle that optimizes the cost function from the perspective of the second vehicle based on the third potential behavior predicted for the third vehicle.

15. The controller of claim 11 , wherein the instructions upon execution cause the at least one processor to:
generate, based on the operating state of the second vehicle, candidate trajectories for the second vehicle, each of the candidate trajectories having a cost according to the cost function; and
select, for the first potential behavior, the candidate trajectory having a lowest cost.

16. The controller of claim 15 , wherein the instructions upon execution cause the at least one processor to:
identify a travel context for the first vehicle and the second vehicle;
retrieve a goal set based on the travel context; and
generate the candidate trajectories

Assignees

• Denso Int America Inc
• RobotWits LLC

Inventors

• Likhachev Maxim
• Ratner Ellis
• Lyu Lu
• Bando Takashi
• Xu Yunfei

Classifications

• G06V20/58

  Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads · CPC title

• G06V20/588

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title

• B60W30/09

  Taking automatic action to avoid collision, e.g. braking and steering · CPC title

• B60W30/0956 Primary

  the prediction being responsive to traffic or environmental parameters · CPC title

• B60W2554/801

  Lateral distance · CPC title