Dynamic scenario parameters for an autonomous driving vehicle

Patent metadata

Field	Value
Publication number	US-12195036-B2
Application number	US-202217805000-A
Country	US
Kind code	B2
Filing date	Jun 1, 2022
Priority date	Jun 1, 2022
Publication date	Jan 14, 2025
Grant date	Jan 14, 2025

Abstract

According to some embodiments, systems, methods and media for dynamically generating scenario parameters for an autonomous driving vehicles (ADV) are described. In one embodiment, when an ADV enters a driving scenario, the ADV can invoke a map-based scenario checker to determine the type of scenario, and invokes a corresponding neural network model to generate a set of parameters for the scenario based on real-time environmental conditions (e.g., traffics) and vehicle status information (e.g., speed). The set of scenario parameters can be a set of extra constraints for configuring the ADV to drive in a driving mode corresponding to the scenario.

First claim

What is claimed is:

1. A computer-implemented method of operating an autonomous driving vehicle (ADV), comprising:
determining, by a scenario checker on the ADV, a scenario that the ADV has entered on a road segment based on a position of the ADV and a boundary parameter associated with a traffic signal in the road segment;
invoking, by the ADV, a trained neural network model corresponding to the scenario determined by the scenario checker;
generating, with the trained neural network model corresponding to the scenario determined by the scenario checker, a set of parameters based on environmental data at the scenario and vehicle status information of the ADV, the set of parameters for the scenario replacing a set of fixed scenario parameters of the trained neural network model, wherein the set of fixed scenario parameters are based on training data collected from human-driven vehicles navigating a similar scenario as the scenario determined by the scenario checker; and
operating the ADV in a driving mode defined by the set of parameters generated based on the environmental data at the scenario and the vehicle status information of the ADV to navigate the scenario.

2. The computer-implemented method of claim 1 , wherein the scenario checker determines the scenario that the ADV has entered based on a current position of the ADV, information on a map surrounding the current position, information of a route that the ADV is taking, and one or more of the set of fixed scenario parameters.

3. The computer-implemented method of claim 2 , wherein the set of fixed scenario parameters are generated by the trained neural network model based on historical environmental data and historical vehicle status information.

4. The computer-implemented method of claim 1 , wherein the environmental data at the scenario includes traffics at the scenario, and wherein the vehicle status information of the ADV includes a speed of the ADV.

5. The computer-implemented method of claim 1 , wherein the set of scenario parameters represents a set of additional constraints that are used by the ADV when generating a planned trajectory.

6. The computer-implemented method of claim 1 , wherein the scenario checker on the ADV is a convolutional neural network (CNN).

7. The computer-implemented method of claim 1 , wherein the scenario is one of a plurality of scenarios that are predefined based on map information, wherein the plurality of scenarios include a junction, a yield sign, a stop sign, a one-lane right turn, a one-lane left turn, a multi-lane right turn, and a multi-lane left turn.

8. A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for operating an autonomous driving vehicle (ADV), the operations comprising:
determining a scenario, by a scenario checker, that the ADV has entered on a road segment based on a position of the ADV and a boundary parameter associated with a traffic signal in the road segment;
invoking a trained neural network model corresponding to the scenario;
generating, with the trained neural network model corresponding to the scenario, a set of parameters based on environmental data at the scenario and vehicle status information of the ADV, the set of parameters for the scenario replacing a set of fixed scenario parameters of the trained neural network model, wherein the set of fixed scenario parameters are based on training data collected from human-driven vehicles navigating a similar scenario as the scenario; and
operating the ADV in a driving mode defined by the set of parameters generated based on the environmental data at the scenario and the vehicle status information of the ADV to navigate the scenario.

9. The non-transitory machine-readable medium of claim 8 , wherein the determining that the ADV has entered is based on a current position of the ADV, information on a map surrounding the current position, information of a route that the ADV is taking, and one or more of the set of fixed scenario parameters.

10. The non-transitory machine-readable medium of claim 9 , wherein the set of fixed scenario parameters are generated by the trained neural network model based on historical environmental data and historical vehicle status information.

11. The non-transitory machine-readable medium of claim 8 , wherein the environmental data at the scenario includes traffics at the scenario, and wherein the vehicle status information of the ADV includes a speed of the ADV.

12. The non-transitory machine-readable medium of claim 8 , wherein the set of scenario parameters represent a set of additional constraints that are used by the ADV when generating a planned trajectory.

13. The non-transitory machine-readable medium of claim 8 , wherein the scenario checker on the ADV is a convolutional neural network (CNN).

14. The non-transitory machine-readable medium of claim 8 , wherein the scenario is one of a plurality of scenarios that are predefined based on map information, wherein the plurality of scenarios include a junction, a yield sign, a stop sign, a one-lane right turn, a one-lane left turn, a multi-lane right turn, and a multi-lane left turn.

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 operating an autonomous driving vehicle (ADV), the operations comprising:
determining a scenario, by a scenario checker, that the ADV has entered on a road segment based on a position of the ADV and a boundary parameter associated with a traffic signal in the road segment;
invoking a trained neural network model corresponding to the scenario;
generating, with the trained neural network model corresponding to the determined scenario, a set of parameters based on environmental data at the scenario and vehicle status information of the ADV, the set of parameters for the scenario replacing a set of fixed scenario parameters of the trained neural network model, wherein the set of fixed scenario parameters are based on training data collected from human-driven vehicles navigating a similar scenario as the scenario determined by the scenario checker; and
operating the ADV in a driving mode defined by the set of parameters generated based on the environmental data and the vehicle status information to navigate the scenario.

16. The data processing system of claim 15 , wherein the determining that the ADV has entered is based on a current position of the ADV, information on a map surrounding the current position, information of a route that the ADV is taking, and one or more of the set of fixed scenario parameters.

17. The data processing system of claim 16 , wherein the set of fixed scenario parameters are generated by the trained neural network model based on historical environmental data and historical vehicle status information.

18. The data processing system of claim 15 , wherein the environmental data at the scenario includes traffics at the scenario, and wherein the vehicle status information of the ADV includes a speed of the ADV.

19. The data processing system of claim 15 , wherein the set of scenario parameters represent a set of additional constraints that are used by a planning module when generating a planned trajectory.

20. The data processing system of claim 15 , wherein the scenario checker on the ADV is a convolutional neural network (CNN).

Assignees

• Baidu Usa Llc

Inventors

• Jiang Shu
• Wu Szu Hao
• Cao Yu
• Lin Weiman
• Hu Jiangtao

Classifications

• B60W2420/403

  Image sensing, e.g. optical camera · CPC title

• G06F18/2148

  characterised by the process organisation or structure, e.g. boosting cascade · CPC title

• G06V20/588 Primary

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title

• G06V20/582

  of traffic signs · CPC title

• B60W2554/4041

  Position · CPC title