Scene security level determination method, device and storage medium

Patent metadata

Field	Value
Publication number	US-12397822-B2
Application number	US-202217821456-A
Country	US
Kind code	B2
Filing date	Aug 22, 2022
Priority date	Aug 31, 2021
Publication date	Aug 26, 2025
Grant date	Aug 26, 2025

Abstract

The present application provides a scene security level determination method, relates to the technical field of autonomous driving, and is used for determining the security level of an unknown scene. The method comprises: determining, according to a first parameter set and a first parameter value set of the target scene, whether at least one predefined scene with a similarity between the at least one predefined scene and a target scene being larger than a set similarity threshold exists in a predefined scene library; selecting, in response to the at least one predefined scene existing, one predefined scene from the at least one predefined scene as a control scene; generating, according to the parameter subset of the control scene, a parameter value subset of the target scene; and determining, according to the parameter subset and the parameter value subset, a target security level of the target scene.

First claim

The invention claimed is:

1. A scene security level determination method, comprising:
determining, by a processor in a target vehicle according to a first parameter set and a first parameter value set of a target scene where the target vehicle is located, whether at least one predefined scene exists in a predefined scene library, wherein a similarity between the at least one predefined scene and the target scene being larger than a set similarity threshold; wherein the first parameter set includes parameters whose values have been acquired for the target scene, and the first parameter value set includes parameter values of the parameters in the first parameter set;
selecting, by the processor in response to the at least one predefined scene existing, one predefined scene from the at least one predefined scene as a control scene; wherein there is a parameter subset of the control scene, and the parameter subset of the control scene is a subset of the first parameter set of the target scene;
generating, by the processor according to the parameter subset of the control scene, a parameter value subset of the target scene; wherein the parameter value subset of the target scene includes values of parameters in the parameter subset of the control scene, and the parameter value subset of the target scene is a subset of the first parameter value set;
determining, by the processor according to the parameter subset of the control scene and the parameter value subset of the target scene, a target security level of the target scene; and
causing, by the processor, the target vehicle to perform according to the target security level of the target scene;
wherein the method further comprises:
determining value ranges of each parameter in the parameter subset of the predefined scene corresponding to different security levels, comprising:
in response to the fact that at least two parameters in the parameter subset of the predefined scene are non-unicity parameters,
determining, according to importance degrees of the at least two parameters on driving security of the predefined scene, a reference parameter from the at least two parameters;
determining a value range of the reference parameter at each security level; and

determining, according to the value range of the reference parameter at each security level, value ranges of other non-unicity parameters in the at least two parameters corresponding to each security level respectively;
wherein the non-unicity parameters comprise at least two parameters in which if one of the at least two parameters meets a predetermined condition, the other parameter is classified into a same security category no matter how the other parameter changes.

2. The method according to claim 1 , wherein the determining, according to the parameter subset of the control scene and the parameter value subset of the target scene, the target security level of the target scene comprises:
for each parameter of the parameter subset, determining, according to value ranges of the parameter corresponding to different security levels and the parameter value of the parameter in the parameter value subset, a parameter security level corresponding to the parameter; and
determining, according to the parameter security levels, the target security level of the target scene.

3. The method according to claim 1 , further comprising: constructing a predefined scene library, wherein the predefined scene library comprises the at least one predefined scene, and the constructing the predefined scene library specifically comprises:
for each predefined scene,
determining, according to a value of each parameter acquired for the predefined scene, a second parameter set and a second parameter value set of the predefined scene; wherein the second parameter set includes the parameters whose values have been acquired for the predefined scene, and the second parameter value set includes values of the parameters in the second parameter set;
determining, according to an importance degree of each parameter of the predefined scene, a parameter subset of the predefined scene from the second parameter set; and
determining, according to a parameter value subset of the predefined scene, the value range of each parameter in the parameter subset of the predefined scene corresponding to a security level, wherein the parameter value subset of the predefined scene includes values of the parameters in the parameter subset of the predefined scene.

4. The method according to claim 3 , wherein the importance degree includes a first importance degree or a second importance degree, and the first importance degree is determined according to historical road measured data or historical simulation data, and the second importance degree is set according to an empirical value.

5. The method according to claim 3 , wherein the importance degree includes a first importance degree and a second importance degree, and the determining, according to the importance degree of each parameter of the predefined scene, the parameter subset of the predefined scene from the second parameter set comprises:
determining, according to the first importance degree, a first parameter subset from the second parameter set, wherein the first importance degree is determined according to historical road measured data or historical simulation data; and
determining, according to the second importance degree, the parameter subset of the predefined scene from the first parameter subset, wherein the second importance degree is set according to an empirical value.

6. The method according to claim 3 , wherein the parameter value of the non-unicity parameters has a correlation with a value of at least one parameter in the parameter subset of the predefined scene;
the value range of the reference parameter at each security level are determined according to historical road measured data, historical simulation data or empirical values.

7. The method according to claim 1 , further comprising:
in response to the fact that no predefined scene exists in the predefined scene library with a similarity above the similarity threshold with the target scene, determining a preset security level as the target security level of the target scene.

8. The method according to claim 1 , further comprising:
in response to the fact that no predefined scene exists in the predefined scene library with a similarity above the similarity threshold with the target scene, adding the target scene to the predefined scene library.

9. The method according to claim 1 , wherein the non-unicity parameters comprise inter-vehicle distance, a speed of the target vehicle, and a preceding vehicle speed.

10. A device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the computer program, implements a method comprising:
determining, by a process in a target vehicle according to a first parameter set and a first parameter value set of a target scene where a target vehicle is located, whether at least one predefined scene exists in a predefined scene library, wherein a similarity between the at least one predefined scene and the target scene being larger than a set similarity threshold; wherein the first parameter set includes parameters whose values have been acquired for the target scene, and the first parameter value set includes parameter values of the parameters in the first parameter set;
selecting, by the process in response to the at least one predefined scene existing, one predefined scene from the at least one predefined scene as a control scene; wherein there is a parameter subset of the control scene, and the parameter

Assignees

• Beijing Tusen Weilai Tech Co Ltd
• Beijing Tusen Weilal Tech Co Ltd

Inventors

• Wu Nan

Classifications

• B60W40/02

  related to ambient conditions · CPC title

• B60W60/0015

  specially adapted for safety · CPC title

• B60W2556/10

  Historical data · CPC title

• B60W60/0011 Primary

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

• G06V20/56 Primary

  exterior to a vehicle by using sensors mounted on the vehicle · CPC title