Method and system for determining optimal coefficients of controllers for autonomous driving vehicles

Patent metadata

Field	Value
Publication number	US-10613489-B2
Application number	US-201715627970-A
Country	US
Kind code	B2
Filing date	Jun 20, 2017
Priority date	Jun 20, 2017
Publication date	Apr 7, 2020
Grant date	Apr 7, 2020

Abstract

Driving parameters (e.g., speed, heading direction) that an autonomous driving vehicle (ADV) likely utilize as target driving parameters are grouped into a number of ranges and one of the driving parameters in each range is selected as a driving parameter representative or a target driving parameter representing the respective range or segment. For each of the target driving parameters representing the ranges, a particle swarm optimization method is utilized to derive a set of most optimized coefficients for a controller (e.g., speed controller, steering controller) for controlling an ADV. A driving parameter to coefficient (parameter/coefficient) mapping table is generated to map a particular driving parameter representing a range of driving parameter to a set of one or more coefficients of a particular controller. The parameter/coefficient mapping table is utilized at real-time to configure a controller in response to a particular target driving parameter using the corresponding coefficients.

First claim

What is claimed is:

1. A computer-implemented method for determining coefficients of a controller for controlling an autonomous driving vehicle (ADV), the method comprising:
selecting a plurality of target driving parameters in a plurality of predetermined ranges, a first target driving parameter of the plurality of target driving parameters representing driving parameters in a first predetermined range of the plurality of predetermined ranges:
selecting a controller coefficient candidate from a set of controller coefficient candidates within the first predetermined range for controlling the ADV to drive based on the first target driving parameter;
calculating a cost using a cost function based on the selected controller coefficient candidate and the first target driving parameter;
determining a local best controller coefficient for the first predetermined range, such that the cost is minimized amongst the controller coefficient candidates of the set of the first predetermined range; and
determining a global best controller coefficient based on the local best controller coefficient in view of other local best controller coefficients of other ranges of controller coefficient candidates, wherein the global best controller coefficient is utilized by the controller to control the ADV with the first target driving parameter.

2. The method of claim 1 , wherein for each of the predetermined ranges, the method further comprises performing calculating a cost using the cost function and determining a local best controller coefficient for the respective predetermined range.

3. The method of claim 2 , wherein the first target driving parameter represents one of a plurality of driving parameters, and wherein for each of the driving parameters, the method further comprises performing calculating a cost using the cost function, determining a local best controller coefficient for each of the plurality of predetermined ranges, and determining a global best controller coefficient for the respective driving parameter.

4. The method of claim 3 , wherein each of the driving parameters represents a range of driving parameter values.

5. The method of claim 1 , wherein determining a local best controller coefficient for the predetermined range comprises:
comparing the cost associated with the selected controller coefficient candidate with a current local best cost associated with a current local best controller coefficient of the first predetermined range; and
in response to determining that the cost associated with the selected controller coefficient candidate is lower than the local best cost associated with the current local best controller coefficient,
replacing the current local best controller coefficient with the selected controller coefficient candidate, and
replacing the current local best cost with the cost associated with the selected controller coefficient candidate.

6. The method of claim 1 , wherein determining a global best controller coefficient based on the local best controller coefficient comprises:
comparing the cost associated with the selected controller coefficient candidate with a global best cost associated with a current global best controller coefficient; and
in response to determining that the cost associated with the selected controller coefficient candidate is lower than the global best cost associated with the current global best controller coefficient,
replacing the current global best controller coefficient with the selected controller coefficient candidate, and
replacing the global best cost with the cost associated with the selected controller coefficient candidate.

7. The method of claim 1 , wherein each of the controller coefficient candidates represents at least one of Kp, Ki, or Kd of a proportional-integral-derivative controller (PID) controller.

8. The method of claim 1 , wherein the first target driving parameter is one of a target speed or a heading direction of the ADV.

9. The method of claim 1 , further comprising generating a driving parameter to controller coefficient mapping (parameter/coefficient) table to be used to configure a controller for driving the ADV with a similar target driving parameter, wherein the parameter/coefficient table includes a plurality of mapping entries, each mapping entry mapping a particular driving parameter to a particular controller coefficient.

10. A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations of determining optimal controller coefficients, the operations comprising:
selecting a plurality of target driving parameters in a plurality of predetermined ranges, a first target driving parameter of the plurality of target driving parameters representing driving parameters in a first predetermined range of the plurality of predetermined ranges;
selecting a controller coefficient candidate from a set of controller coefficient candidates within the first predetermined range for controlling the ADV to drive based on the first target driving parameter;
calculating a cost using a cost function based on the selected controller coefficient candidate and the first target driving parameter;
determining a local best controller coefficient for the first predetermined range, such that the cost is minimized amongst the controller coefficient candidates of the set of the first predetermined range; and
determining a global best controller coefficient based on the local best controller coefficient in view of other local best controller coefficients of other ranges of controller coefficient candidates, wherein the global best controller coefficient is utilized by the controller to control the ADV with the first target driving parameter.

11. The machine-readable medium of claim 10 , wherein for each of the predetermined ranges, the method further comprises performing calculating a cost using the cost function and determining a local best controller coefficient for the respective predetermined range.

12. The machine-readable medium of claim 11 , wherein the first target driving parameter represents one of a plurality of driving parameters, and wherein for each of the driving parameters, the method further comprises performing calculating a cost using the cost function, determining a local best controller coefficient for each of the plurality of predetermined ranges, and determining a global best controller coefficient for the respective driving parameter.

13. The machine-readable medium of claim 12 , wherein each of the driving parameters represents a range of driving parameter values.

14. The machine-readable medium of claim 10 , wherein determining a local best controller coefficient for the predetermined range comprises:
comparing the cost associated with the selected controller coefficient candidate with a current local best cost associated with a current local best controller coefficient of the first predetermined range; and
in response to determining that the cost associated with the selected controller coefficient candidate is lower than the local best cost associated with the current local best controller coefficient,
replacing the current local best controller coefficient with the selected controller coefficient candidate, and
replacing the current local best cost with the cost associated with the selected controller coefficient candidate.

15. The machine-readable medium of claim 10 , wherein determining a global best controller coefficient based on the local best controller coefficient comprises:
comparing the cost associated with the selected contro

Assignees

• Baidu Usa Llc

Inventors

• Luo Qi
• Kong Qi
• Zhu Fan
• Yu Xiang
• Hu Sen
• Zhuang Li
• Zhu Weicheng
• Yang Guang
• Wang Jingao

Classifications

• B60W60/001

  Planning or execution of driving tasks · CPC title

• G05B13/044 Primary

  not using a perturbation signal · CPC title

• B60W2050/0043

  Signal treatments, identification of variables or parameters, parameter estimation or state estimation · CPC title

• B60W50/00 Primary

  Details of control systems for road vehicle drive control not related to the control of a particular sub-unit {, e.g. process diagnostic or vehicle driver interfaces} · CPC title

• G05D2201/0212

  Physics · mapped topic