Method and device for determining a collision probability of a vehicle with an object in a three-dimensional space

Patent metadata

Field	Value
Publication number	US-12494126-B2
Application number	US-202118015016-A
Country	US
Kind code	B2
Filing date	Jun 11, 2021
Priority date	Jul 14, 2020
Publication date	Dec 9, 2025
Grant date	Dec 9, 2025

Abstract

Determining a collision probability of a vehicle with an object. A three-dimensional position and orientation of the vehicle and the object is detected. The vehicle is approximated by at least one first geometrical element, that includes a part of the vehicle. The same is done for the object. A first error and a second error is determined in relation to the geometrical element for the vehicle or object. A Minkowski difference is formed for each combination of the at least one first geometrical element with the at least one second geometrical element. A third error is standardized via a transformation. The transformation serves to calculate calculation elements. The collision probability of the vehicle with the object is determined on the basis a spatial union of the calculation elements.

First claim

The invention claimed is:

1 . A method for determining a collision probability of a vehicle with an object in a three-dimensional airspace and enabling collision avoidance, comprising:
receiving a three-dimensional position and orientation of the vehicle and of the object, the position comprising x, y, and z coordinates; approximating the vehicle by at least one first geometric body, wherein the at least one first geometric body includes a portion of the vehicle; approximating the object by at least one second geometric body, wherein the at least one second geometric body includes a portion of the object; determining a first indeterminacy of the at least one first geometric body; determining a second indeterminacy of the at least one second geometric body; taking a Minkowski difference for each combination of the at least one first geometric body with the at least one second geometric body; determining a respective third indeterminacy for each Minkowski difference as a function of the respective first indeterminacy of the at least one first geometric body and the respective second indeterminacy of the at least one second geometric body, the third indeterminacy based on a covariance matrix; standardizing the respective third indeterminacy using a respective transformation; applying the respective transformation to the respective taken Minkowski difference to determine respective computing bodies from the respective Minkowski difference; determining the collision probability of the vehicle with the object based on a set union of the computing bodies; and using the determined collision probability to control the vehicle in real-time to avoid a collision with the object by at least one of maneuvering the vehicle along a trajectory or outputting an information signal to an operator of the vehicle.

2 . The method according to claim 1 , wherein a volume of the at least one first and second geometric bodies are determined as a function of a digital resource.

3 . The method according to claim 2 , further comprising detecting an object type of the object using object recognition, and determining the second indeterminacy for the at least one second body as a function of the object type.

4 . The method according to claim 2 , wherein the respective computing bodies for approximating the vehicle and the object are convex.

5 . The method according to claim 1 , wherein
the at least one first geometric body completely contains the vehicle, and the at least one second geometric body completely contains the object, or the at least one first geometric body underapproximates the vehicle, and the at least one second geometric body underapproximates the object.

6 . The method according to claim 1 , wherein the first indeterminacy and/or the second indeterminacy are determined based on a Gaussian normal distribution and/or error propagation.

7 . The method according to claim 1 , wherein
the three-dimensional position and/or orientation of the object is detected multiple times at different points in time for determining a three-dimensional trajectory of the object, and one of multiple trajectories for the vehicle, which has the lowest collision probability, is selected as a function of the trajectory of the object.

8 . The method according to claim 1 , wherein the collision probability is determined for multiple points of the trajectories for the vehicle.

9 . The method according to claim 1 , wherein the vehicle and the object are each approximated by at least one polyhedron, and the set union of the computing bodies is approximated via a prism, the prism completely containing the set union, and used for determining the collision probability.

10 . The method according to claim 1 , further comprising approximating the computing bodies by multiple prisms that contact one another, the prisms being symmetrical with respect to one another.

11 . The method according to claim 1 , further comprising approximating each of the vehicle and the object by at least one sphere, and approximating the set union of the computing bodies by a cuboid, the cuboid completely containing the set union, and used for determining the collision probability.

12 . The method according to claim 1 , wherein using the determined collision probability to control the vehicle comprises at least one of generating a control signal to maneuver the vehicle along a trajectory to avoid the object or outputting an information signal to an operator of the vehicle via a display or audio device to prompt manual avoidance of the object.

13 . An apparatus for determining a collision probability of a vehicle with an object in a three-dimensional airspace and enabling collision avoidance, comprising:
at least one sensor unit for detecting a three-dimensional position and orientation of the vehicle and of the object via an interface-( 15 ), the position comprising x, y, and z coordinates; a control unit, operatively coupled to the at least one sensor unit, wherein the control unit is configured to
approximate the vehicle by at least one first geometric body, wherein the at least one first geometry body includes a portion of the vehicle;
approximate the object by at least one second geometric body, wherein the at least one second geometric body includes a portion of the object;
determine a first indeterminacy of the at least one first geometric body;
determine a second indeterminacy of the at least one second geometric body;
taking a Minkowski difference for each combination of the at least one first geometric body with the at least one second geometric body;
determine a respective third indeterminacy for each Minkowski difference as a function of the respective first indeterminacy of the at least one first geometric body and the respective second indeterminacy of the at least one second geometric body, the third indeterminacy based on a covariance matrix;
standardize the respective third indeterminacy using a respective transformation;
apply the respective transformation to the respective taken Minkowski difference to determine respective computing bodies from the respective Minkowski difference;
determine the collision probability of the vehicle with the object based on a set union of the computing bodies; and
use the determined collision probability to control the vehicle to avoid a collision with the object by at least one of maneuvering the vehicle along a trajectory or outputting an information signal to an operator of the vehicle.

14 . The apparatus according to claim 13 , wherein a volume of the at least one first and second geometric bodies are determined as a function of a digital resource.

15 . The apparatus according to claim 13 , wherein
the at least one first geometric body completely contains the vehicle, and the at least one second geometric body completely contains the object, or the at least one first geometric body underapproximates the vehicle, and the at least one second geometric body underapproximates the object.

16 . The apparatus according to claim 13 , wherein the first indeterminacy and/or the second indeterminacy are determined based on a Gaussian normal distribution and/or error propagation.

17 . The apparatus according to claim 13 , wherein
the three-dimensional position and/or orientation of the object is detected multiple times at different points in time for determining a three-dimensional trajectory of the object, and one of multiple trajectories for the vehicle, which has the lowest collision probability, is selected as a function of the trajectory

Assignees

• Volkswagen Ag
• Ktm Ag

Inventors

• Hruschka Clemens Markus

Classifications

• G08G5/80

  Anti-collision systems · CPC title

• G08G7/02

  Anti-collision systems · CPC title

• G08G1/165

  for passive traffic, e.g. including static obstacles, trees · CPC title

• G08G1/166 Primary

  for active traffic, e.g. moving vehicles, pedestrians, bikes · CPC title