Systems and methods for generating simulation scenario definitions for an autonomous vehicle system

Patent metadata

Field	Value
Publication number	US-11643105-B2
Application number	US-202016797103-A
Country	US
Kind code	B2
Filing date	Feb 21, 2020
Priority date	Feb 21, 2020
Publication date	May 9, 2023
Grant date	May 9, 2023

Abstract

A system of creating a simulation scenario definition to simulate behavior of an autonomous vehicle includes a computing device and a computer-readable storage medium having one or more programming instructions. The system identifies an event on which a simulation scenario definition is to be based, identifies one or more log files associated with the event, parses the one or more log files in a time-sequential order and populates the simulation scenario definition with information from the identified log files until an event trigger is detected, identifies an actor from one or more of the log files, infer a shape of the actor, generates one or more simulated tracks that includes the inferred shape for the actor, and adds the simulated tracks to the simulation scenario definition.

First claim

The invention claimed is:

1. A method of creating a simulation scenario definition to simulate behavior of an autonomous vehicle, the method comprising:
by a computing device:
identifying an event on which a simulation scenario definition is to be based;
identifying one or more log files associated with the event, wherein the log files comprise data that is collected by one or more sensors of an autonomous vehicle while experiencing the event and logged by the autonomous vehicle to one or more data stores;
parsing the one or more log files in a time-sequential order and populating the simulation scenario definition with information from the identified log files until an event trigger is detected;
identifying an actor from one or more of the log files;
inferring a shape of the actor by
generating a shape estimate for the actor based on track information associated with the actor,
transforming the shape estimate into a box frame having at least one of a position and orientation of the actor for each track of the one or more log files in which the actor appears,
averaging the box frames to obtain an average box frame,
classifying the actor into an object class using the average box frame, and
setting the shape of the actor to a shape associated with the object class;

generating one or more simulated tracks that includes the inferred shape for the actor; and
adding the simulated tracks to the simulation scenario definition.

2. The method of claim 1 , further comprising using a simulation system to generate a data set from a simulated environment by operating a virtual vehicle in the simulated environment wherein a trajectory of the virtual vehicle is defined by the simulation scenario definition.

3. The method of claim 1 , wherein populating the simulation scenario definition with information from the identified log files comprises populating the simulation scenario definition within information pertaining to a pose of the autonomous vehicle.

4. The method of claim 1 , wherein populating the simulation scenario definition with information from the identified log files comprises populating the simulation scenario definition within information pertaining to a velocity of the autonomous vehicle.

5. The method of claim 1 , wherein identifying an actor from one or more of the log files comprises receiving an indication of the actor from a user.

6. The method of claim 1 , wherein
the average box frame represents a first polygon.

7. The method of claim 6 , wherein the classifying further comprises identifying one or more templates, wherein each identified template comprises shape information pertaining to a category of object, wherein the shape information represents a second polygon.

8. The method of claim 7 , wherein each identified template comprises shape information pertaining to a vehicle.

9. The method of claim 7 , wherein the classifying further comprises:
comparing the average box frame to the shape information from one or more of the identified templates to generate a score that is (i) associated with the average box frame and the template and (ii) indicative of an area covered by a ratio of an intersection of the first polygon and the second polygon to an area covered by a union of an area of the first polygon and the second polygon, wherein the box frame and the template are aligned.

10. The method of claim 9 , wherein the classifying further comprises selecting the identified template associated with the highest score.

11. The method of claim 10 , wherein the classifying further comprises:
determining whether the score associated with the selected template is below a threshold value, and
in response to determining that the score associated with the selected template is below a threshold value, generating and providing a warning to a user.

12. The method of claim 10 , wherein the classifying further comprises scaling the second polygon represented by the selected template to fit the average box frame.

13. A system of creating a simulation scenario definition to simulate behavior of an autonomous vehicle, the system comprising:
a computing device; and
a computer-readable storage medium comprising one or more programming instructions that, when executed, cause the computing device to:
identify an event on which a simulation scenario definition is to be based;
identify one or more log files associated with the event, wherein the log files comprise data that is collected by one or more sensors of an autonomous vehicle while experiencing the event and logged by the autonomous vehicle to one or more data stores;
parse the one or more log files in a time-sequential order and populating the simulation scenario definition with information from the identified log files until an event trigger is detected;
identify an actor from one or more of the log files;
infer a shape of the actor by
generating a shape estimate for the actor based on track information associated with the actor,
transforming the shape estimate into a box frame having at least one of a position and orientation of the actor for each track of the one or more log files in which the actor appears,
averaging the box frame to obtain an average box frame,
classifying the actor into an object class using the average box frame, and
setting the shape of the actor to a shape associated with the object class;

generate one or more simulated tracks that includes the inferred shape for the actor; and
add the simulated tracks to the simulation scenario definition.

14. The system of claim 13 , wherein the computer-readable storage medium further comprises one or more programming instruction that, when executed, causes the computing device to use a simulation system to generate a data set from a simulated environment by operating a virtual vehicle in the simulated environment wherein a trajectory of the virtual vehicle is defined by the simulation scenario definition.

15. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the computing device to populate the simulation scenario definition with information from the identified log files comprise one or more programming instructions that, when executed, cause the computing device to populate the simulation scenario definition within information pertaining to a pose of the autonomous vehicle.

16. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the computing device to populate the simulation scenario definition with information from the identified log files comprise one or more programming instructions that, when executed, cause the computing device to populate the simulation scenario definition within information pertaining to a velocity of the autonomous vehicle.

17. The system of claim 13 , wherein the one or more programming instructions that, when executed, cause the computing device to identify an actor from one or more of the log files comprise one or more programming instructions that, when executed, cause the computing device to receive an indication of the actor from a user.

18. The system of claim 13 , wherein
the average box frame represents a first polygon.

19. The system of claim 18 , wherein the computer-readable storage medium further comprises one or more programming instructions that, when executed, cause the computing device to identify one or more templates, wherein each identified template comprises shape information pertaining to a category of object, wherein the shape information rep

Assignees

• Argo Ai Llc

Inventors

• Ackenhausen Thomas Carl
• Carmody Patrick Michael
• Costa Albert
• Nolley Alexander Allen

Classifications

• G06F16/1734

  Details of monitoring file system events, e.g. by the use of hooks, filter drivers, logs · CPC title

• 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

• G06F11/3684 Primary

  for test design, e.g. generating new test cases · CPC title

• G06F30/20

  Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30) · CPC title

• B60W60/0011 Primary

  involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles · CPC title