Adjusting lateral clearance for a vehicle using a multi-dimensional envelope

Patent metadata

Field	Value
Publication number	US-11827241-B2
Application number	US-201916661228-A
Country	US
Kind code	B2
Filing date	Oct 23, 2019
Priority date	Oct 29, 2018
Publication date	Nov 28, 2023
Grant date	Nov 28, 2023

Abstract

Among other things, we describe techniques for adjusting lateral clearance for a vehicle using a multi-dimensional envelope. A trajectory is generated for the vehicle. A multi-dimensional envelope is generated indicating a drivable region for the vehicle and containing the trajectory. One or more objects are identified located along or adjacent to the trajectory. At least one dimension of the generated multi-dimensional envelope is adjusted to adjust a lateral clearance between the vehicle and the identified one or more objects. A control module of the vehicle navigates the vehicle along the multi-dimensional envelope.

First claim

What is claimed is:

1. A method comprising:
generating, using one or more processors, a trajectory for a vehicle, the trajectory comprising a plurality of spatiotemporal locations;
generating, using the one or more processors, a multi-dimensional envelope indicating first lateral constraints on a drivable region for the vehicle and containing the trajectory;
identifying, using the one or more processors, one or more objects located within a threshold distance to the trajectory;
adjusting, using the one or more processors, at least one dimension of the generated multi-dimensional envelope by a machine learning model trained to modify a lateral clearance between the vehicle and the one or more objects and increase passenger comfort, wherein features extracted from the one or more objects and the threshold distance to the trajectory are input to the machine learning model and the machine learning model outputs the lateral clearance, said lateral clearance positively impacting measured passenger comfort data captured by sensors of the vehicle; and
navigating, using the one or more processors, the vehicle based on the adjusted, generated multi-dimensional envelope.

2. The method of claim 1 , wherein the identifying of the one or more objects comprises determining that the vehicle has a likelihood of collision greater than zero with the one or more objects.

3. The method of claim 1 , wherein the adjusting of the at least one dimension of the generated multi-dimensional envelope is based on at least one of a speed of the vehicle, an acceleration, or a magnitude of jerk.

4. The method of claim 1 , wherein the generated multi-dimensional envelope corresponds to a shape having a geometric volume.

5. The method of claim 4 , wherein the geometric volume is tubular, cubic, or conical.

6. The method of claim 1 , wherein the generated multi-dimensional envelope is generated based on information contained within a map of an environment in which the vehicle is navigating.

7. The method of claim 6 , wherein the information contained within the map of the environment represents at least one of a road, a parking lot, a bridge, a construction zone, a curb of a road, a boundary of a lane, an intersection, or a building.

8. The method of claim 1 , wherein the one or more objects are identified from data captured by one or more sensors comprising at least one of a monocular video camera, a stereo video camera, a visible light camera, an infrared camera, a thermal imager, a LiDAR, a radar, an ultrasonic sensor, or a time-of-flight (TOF) depth sensor.

9. The method of claim 1 , wherein the identified one or more objects comprise at least one of other vehicles, pedestrians, or cyclists.

10. The method of claim 1 , wherein the identified one or more objects comprise construction zones or curbs.

11. The method of claim 1 , wherein the adjusting of the at least one dimension of the generated multi-dimensional envelope comprises:
determining a particular lateral clearance between the vehicle and the one or more objects based on features extracted from the one or more objects and the threshold distance to the trajectory; and
modifying the at least one dimension of the generated multi-dimensional envelope, such that the lateral clearance between the vehicle and the identified one or more objects is greater than the particular lateral clearance.

12. The method of claim 1 , further comprising training the machine learning model to determine a particular lateral clearance between the generated multi-dimensional envelope and the one or more objects based on features extracted from the one or more objects and the threshold distance to the trajectory.

13. The method of claim 1 , wherein the machine learning model is based on vehicular collision data or traffic data.

14. The method of claim 1 , wherein the generating of the multi-dimensional envelope comprises:
sampling, using the one or more processors, the trajectory to identify the plurality of spatiotemporal locations;
for each spatiotemporal location of the plurality of spatiotemporal locations, determining, using the one or more processors, a lateral error tolerance with respect to the identified one or more objects, the generated multi-dimensional envelope based on the determined lateral error tolerance.

15. The method of claim 1 , wherein the generating of the multi-dimensional envelope comprises determining, using the one or more processors, a width of the generated multi-dimensional envelope using a lateral error tolerance with respect to the identified one or more objects for at least one side of the multi-dimensional envelope.

16. The method of claim 1 , further comprising determining, using the one or more processors, a likelihood of collision for the vehicle with the identified one or more objects based on a speed constraint for the vehicle, when the vehicle is traveling along the trajectory between a present spatiotemporal location of the vehicle and a spatiotemporal location on the trajectory where the trajectory intersects a boundary of a lane, wherein the vehicle is traveling within the lane and the likelihood of collision is determined for positions between the present spatiotemporal location and the boundary of the lane.

17. The method of claim 1 , further comprising determining, using the one or more processors, a likelihood of collision for the vehicle with the identified one or more objects relative to a location on a longitudinal axis of the vehicle.

18. The method of claim 1 , wherein the at least one dimension of the generated multi-dimensional envelope is dynamically adjusted responsive to changing traffic conditions and operating parameters.

19. A vehicle comprising:
one or more computer processors; and
one or more non-transitory computer-readable storage media storing instructions which, when executed by the one or more computer processors, cause the one or more computer processors to:
generate a trajectory for the vehicle, the trajectory comprising a plurality of spatiotemporal locations;
generate a multi-dimensional envelope indicating a drivable region for the vehicle and containing the trajectory;
identify, using one or more sensors of the vehicle, one or more objects located within a threshold distance to the trajectory;
adjust at least one dimension of the generated multi-dimensional envelope by a machine learning model trained to modify a lateral clearance between the vehicle and the one or more objects and increase passenger comfort, wherein features extracted from the one or more objects and the threshold distance to the trajectory are input to the machine learning model and the machine learning model outputs the lateral clearance, said lateral clearance positively impacting measured passenger comfort data captured by sensors of the vehicle; and
navigate the vehicle based on the adjusted, generated multi-dimensional envelope.

20. One or more non-transitory computer-readable storage media storing instructions which, when executed by one or more computing devices, cause the one or more computing devices to:
generate a trajectory for a vehicle, the trajectory comprising a plurality of spatiotemporal locations;
generate a multi-dimensional envelope indicating first lateral constraints on a drivable region for the vehicle and containing the trajectory;
identify one or more objects located within a threshold distance to the trajectory;
adjust at least one dimension of the generated multi-dimensional envelope by a machine learning model trained to modify a later

Assignees

• Motional Ad Llc

Inventors

• Seccamonte Francesco
• Wolff Eric
• Frazzoli Emilio
• Kabzan Juraj

Classifications

• B60W60/0013 Primary

  specially adapted for occupant comfort · CPC title

• B60W30/0956

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

• B60W60/0016

  of the vehicle or its occupants · CPC title

• B60W60/0027

  using trajectory prediction for other traffic participants · CPC title

• B60W2540/221

  Physiology, e.g. weight, heartbeat, health or special needs · CPC title