Weighted planning trajectory profiling method for autonomous vehicle

Patent metadata

Field	Value
Publication number	US-12157495-B2
Application number	US-202117396405-A
Country	US
Kind code	B2
Filing date	Aug 6, 2021
Priority date	Aug 6, 2021
Publication date	Dec 3, 2024
Grant date	Dec 3, 2024

Abstract

In one embodiment, an exemplary method includes the operations of receiving, at a profiling application, a record file recorded by the ADV for a driving scenario in an area, and a high definition map matching the area; extracting planning messages and perception messages from the record file; and aligning the planning message and the perception messages based on their timestamps. The method further includes calculating an individual performance score for each planning cycle of the ADV for the driving scenario based on the planning messages; calculating a weight for each planning cycle based on the perception messages and the high definition map; and then calculating a weighted score for the driving scenario based on individual performance scores and their corresponding weights.

First claim

What is claimed is:

1. A computer-implemented method of evaluating a planning performance of an autonomous driving vehicle (ADV), the method comprising:
generating, for each of a plurality of planning cycles, one or more control commands including a throttle command, a brake command, and a steering command;
driving the ADV in an area using the one or more control commands;
storing driving records of the ADV associated with the driving of the ADV in a driving scenario in the area;
providing, by a profiling application executed by a processor, an interface to a user;
receiving, through the interface, i) a record file that includes the driving records of the ADV associated with the driving scenario in the area, and ii) a map associated with the area;
extracting, by the profiling application executed by a processor, a plurality of planning messages and a plurality of perception messages from the record file;
aligning each of the plurality of planning messages and each of the plurality of perception messages in time to each of the plurality of planning cycles;
calculating, by the profiling application, an individual performance score for each of a plurality of planning cycles of the ADV for the driving scenario based on the plurality of planning messages wherein each of the plurality of planning cycles represent a time interval during which a planned trajectory is generated for a following time period;
calculating, by the profiling application, a weight for each of the plurality of planning cycles based on the perception messages and the map associated with the area wherein each weight for each of the plurality of planning cycles varies based on one or more factors associated with an environment of the ADV;
calculating, by the profiling application, a weighted score for the driving scenario based on the individual performance score for each of the plurality of planning cycles and the corresponding weight, wherein calculating the weighted score comprises multiplying the individual performance score of each of the plurality of planning cycles and the corresponding weight for that planning cycle, and summing up the weighted score for each of the plurality of planning cycles to generate a total weighted score.

2. The method of claim 1 , wherein the profiling application is configured to run either on the ADV or on a cloud server.

3. The method of claim 1 , wherein the one or more factors include a curvature of a planned path, a number of obstacles that have overlapped trajectories with the ADV, or a junction size, wherein each of the one or more factors is derived either from the plurality of perception messages or from the map.

4. The method of claim 1 , wherein calculating the weighted score for the driving scenario further comprises:
dividing the total weighted score by a number of the plurality of planning cycles.

5. The method of claim 1 , wherein the driving scenario includes a plurality of different stages, each stage corresponding to one or more planning cycles and having a different driving environment.

6. The method of claim 1 , wherein the individual performance score for each of the plurality of planning cycles is based on factors measuring controllability and comfort.

7. The method of claim 1 , further comprising providing one or more interfaces that enable a client to upload the record file and the map to a server for determining the planning performance of the ADV.

8. A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations of evaluating a planning performance of an autonomous driving vehicle (ADV), the operations comprising:
generating, for each of a plurality of planning cycles, one or more control commands including a throttle command, a brake command, and a steering command;
driving the ADV in an area using the one or more control commands;
storing driving records of the ADV associated with the driving of the ADV in a driving scenario in the area;
providing, by a profiling application executed by a processor, an interface to a user;
receiving, through the interface, i) a record file that includes the driving records of the ADV associated with the driving scenario in the area, and ii) a map associated with the area;
extracting the plurality of planning messages and a plurality of perception messages from the record file;
aligning each of the plurality of planning messages and each of the plurality of perception messages in time to each of the plurality of planning cycles;
calculating an individual performance score for each of a plurality of planning cycles of the ADV for the driving scenario based on the plurality of planning messages wherein each of the plurality of planning cycles represent a time interval during which a planned trajectory is generated for a following time period;
calculating a weight for each of the plurality of planning cycles based on the perception messages and a map associated with the area wherein each weight for each of the plurality of planning cycles varies based on one or more factors associated with an environment of the ADV;
calculating a weighted score for the driving scenario based on the individual performance score for each of the plurality of planning cycles and the corresponding weight, wherein calculating the weighted score comprises multiplying the individual performance score of each of the plurality of planning cycles and the corresponding weight for that planning cycle, and summing up the weighted score for each of the plurality of planning cycles to generate a total weighted score.

9. The non-transitory machine-readable medium of claim 8 , wherein the operations are performed either on the ADV or on a cloud server.

10. The non-transitory machine-readable medium of claim 8 , wherein the one or more factors include a curvature of a planned path, a number of obstacles that have overlapped trajectories with the ADV, or a junction size, wherein each of the one or more factors is derived either from the plurality of perception messages or from the map.

11. The non-transitory machine-readable medium of claim 8 , wherein calculating the weighted score for the driving scenario further comprises:
dividing the total weighted score by a number of the plurality of planning cycles.

12. The non-transitory machine-readable medium of claim 8 , wherein the driving scenario includes a plurality of different stages, each stage corresponding to one or more planning cycles and having a different driving environment.

13. The non-transitory machine-readable medium of claim 8 , wherein the individual performance score for each of the plurality of planning cycles is based on factors measuring controllability and comfort.

14. The non-transitory machine-readable medium of claim 8 , wherein the operations further comprise providing one or more interfaces that enable a client to upload the record file and the map to a server for determining the planning performance of the ADV.

15. A data processing system, comprising:
a processor; and
a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations of evaluating a planning performance of an autonomous driving vehicle (ADV), the operations comprising:
generating, for each of a plurality of planning cycles, one or more control commands including a throttle command, a brake command, and a steering command;
driving the ADV in an area using the one or more control commands;
storing driving records of the ADV associated with the driving of the ADV in a drivin

Assignees

• Baidu Usa Llc

Inventors

• Jiang Shu
• Luo Qi
• Cao Yu
• Lin Weiman

Classifications

• B60W40/072

  Curvature of the road · CPC title

• B60W40/04

  Traffic conditions · CPC title

• B60W2556/45

  External transmission of data to or from the vehicle · CPC title

• B60W2556/40

  High definition maps · CPC title

• B60W60/0013

  specially adapted for occupant comfort · CPC title