Feedback-based control model generation for an autonomous vehicle

• US 10,449,958 B2
• Filed: 02/15/2017
• Issued: 10/22/2019
• Est. Priority Date: 02/15/2017
• Status: Active Grant

• Current Assignee: Ford Global Technologies, LLC (Ford Motor Company)
• Original Assignee: Ford Global Technologies, LLC (Ford Motor Company)
• Inventors: Jiang, Zhen, Fu, Yan, Li, Yuzhou
• Primary Examiner(s): Mott, Genna M
• Application Number:
  US15/433,781
• Publication Number:
  US 20180229723A1
• Time in Patent Office: 979 Days
• CPC Class Codes: B60W 10/18 including control of braking systems
• CPC Class Codes: B60W 10/20 including control of steering systems
• CPC Class Codes: B60W 2050/0031 Mathematical model of the vehicle
• CPC Class Codes: B60W 2050/0075 Automatic parameter input, automatic initialising or calibrating means
• CPC Class Codes: B60W 2050/0088 Adaptive recalibration
• CPC Class Codes: B60W 2540/215 Selection or confirmation of options
• CPC Class Codes: B60W 2556/10 Historical data
• CPC Class Codes: B60W 2556/45 External transmission of data to or from the vehicle
• CPC Class Codes: B60W 30/09 Taking automatic action to avoid collision, e.g. braking and steering
• CPC Class Codes: B60W 30/0953 the prediction being responsive to vehicle dynamic parameters
• CPC Class Codes: B60W 40/10 related to vehicle motion
• CPC Class Codes: B60W 60/0013 specially adapted for occupant comfort
• CPC Class Codes: G05B 13/048 using a predictor
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00
• CPC Class Codes: G05D 1/0231 using optical position detecting means position-fixing by using electromagnetic waves other than radio waves, e.g. optical position detecting means G01S5/16
• CPC Class Codes: G05D 1/0255 using acoustic signals, e.g. ultra-sonic singals sonar systems designed for anti-collision purposes G01S15/93
• CPC Class Codes: G05D 1/0257 using a radar radar systems designed for anti-collision purposes between land vehicles or between land vehicle and fixed obstacles G01S13/931

Abstract

A computer system receives data sets for a driving maneuver at a location, such as data regarding a state of a vehicle, an external environment indicated by external sensors, and user feedback regarding the driving maneuver. Data sets may be collected from users having various demographic and behavioral attributes. Users may be selected using design of experiment (DOE) algorithms to cover a wide range of possible combinations. A stochastic response surface model (SRSM) is generated that relates sensor data (vehicle state and environmental) to user feedback regarding safety and comfort. The SRSM may be generated using Gaussian process regression (GPR) in order to model uncertainty. The SRSM is then used to generate a control model using an optimization under uncertainty (OUU) algorithm.

Claims

1. A method comprising, by a computer system:

receiving feedback from one or more passengers of one or more autonomous vehicles;

receiving sensor data and control data from the one or more autonomous vehicles;

generating a behavioral model according to the feedback, sensor data, and control data;

determining uncertainty for the behavioral model; and

generating a control model according to the behavioral model and the uncertainty for the behavioral model;

wherein generating the behavior model according to the feedback and sensor data comprises generating a stochastic response surface model (SRSM) according to the feedback and sensor data, the SRSM relating the feedback to environmental factors indicated by the sensor data, status of the one or more autonomous vehicles indicated in the sensor data, and control actions taken by the autonomous vehicle indicated in the control data.

2. The method of claim 1, wherein the feedback includes one or more driving maneuver ratings, each rating including a location.

3. The method of claim 2, wherein receiving feedback includes receiving feedback from a plurality of passengers.

4. The method of claim 3, further comprising selecting, by the computer system, the plurality of passengers according to design of experiment criteria.

5. The method of claim 3, further comprising selecting, by the computer system, the plurality of passengers according to a design of experiment algorithm that processes potential passengers for selection according to demographic attributes of the potential passengers and driving preferences of the potential passengers.

6. The method of claim 1, further comprising:

evaluating, by the computer system, the SRSM according to a portion of the feedback, sensor data, and control data;

determining, by the computer system, that the SRSM does not satisfactorily correspond to the portion of the feedback, sensor data, and control data;

in response to determining that the SRSM does not satisfactorily correspond to the portion of the feedback, sensor data, and control data, collecting additional feedback, additional sensor data, and additional control data; and

generating an updated SRSM using the additional feedback, additional sensor data, and additional control data.

7. The method of claim 1, wherein generating the control model according to the behavior model and the uncertainty for the model comprises performing a multi-objective optimization under uncertainty (OUU) algorithm using the behavioral model and the uncertainty for the behavioral model.

8. The method of claim 1, wherein the sensor data includes outputs of at least one of a light detection and ranging (LIDAR) sensor, a radio detection and ranging (RADAR) sensor, and one or more cameras.

9. The method of claim 1, further comprising:

uploading the control model into an autonomous vehicle;

receiving, by a controller of the autonomous vehicle, outputs of one or more sensors; and

autonomously driving, by the controller, the autonomous vehicle using the outputs processed according to the control model.

10. A system comprising one or more processing devices and one or more memory devices operably coupled to the one or more processing devices, the one or more memory devices storing executable code effective to cause the one or more processing devices to:

receive feedback from one or more passengers of one or more autonomous vehicles;

receive sensor data and control data from the one or more autonomous vehicles;

generate a behavioral model according to the feedback, sensor data, and control data;

determine uncertainty for the behavioral model; and

generate a control model according to the behavioral model and the uncertainty for the behavioral model;

wherein the executable code is further effective to cause the one or more processing devices to generate the behavior model according to the feedback and sensor data by generating a stochastic response surface model (SRSM) according to the feedback and sensor data, the SRSM relating the feedback to environmental factors indicated by the sensor data, status of the one or more autonomous vehicles indicated in the sensor data, and control actions taken by the autonomous vehicle indicated in the control data.

11. The system of claim 10, wherein the feedback includes one or more driving maneuver ratings, each rating including a location.

12. The system of claim 11, wherein the one or more passengers include a plurality of passengers.

13. The system of claim 12, wherein the executable code is further effective to cause the one or more processing devices to select the plurality of passengers according to design of experiment criteria.

14. The system of claim 12, wherein the executable code is further effective to cause the one or more processing devices to select the plurality of passengers according to a design of experiment algorithm that processes potential passengers for selection according to demographic attributes of the potential passengers and driving preferences of the potential passengers.

15. The system of claim 10, wherein the executable code is further effective to cause the one or more processing devices to:

evaluate the SRSM according to a portion of the feedback, sensor data, and control data;

if the SRSM does not satisfactorily correspond to the portion of the feedback, sensor data, and control data;

collect additional feedback, additional sensor data, and additional control data; and

generate an updated SRSM using the additional feedback, additional sensor data, and additional control data.

16. The system of claim 10, wherein the executable code is further effective to cause the one or more processing devices to generate the control model according to the behavior model and the uncertainty for the model by performing a multi-objective optimization under uncertainty (OUU) algorithm using the behavioral model and the uncertainty for the behavioral model.

17. The system of claim 10, wherein the sensor data includes outputs of at least one of a light detection and ranging (LIDAR) sensor, a radio detection and ranging (RADAR) sensor, and one or more cameras.

18. The system of claim 10, wherein the executable code is further effective to cause the one or more processing devices to upload the control model into an autonomous vehicle.