Social behavior for autonomous vehicles

Patent metadata

Field	Value
Publication number	US-11884302-B2
Application number	US-202017099035-A
Country	US
Kind code	B2
Filing date	Nov 16, 2020
Priority date	Nov 15, 2019
Publication date	Jan 30, 2024
Grant date	Jan 30, 2024

Abstract

Understanding the intent of human drivers and adapting to their driving styles is used to increased efficiency and safety of autonomous vehicles (AVs) by enabling them to behave in safe and predictable ways without requiring explicit inter-vehicle communication. A Social Value Orientation (SVO), which quantifies the degree of an agent's selfishness or altruism, is estimated by the AV for other vehicles to better predict how they will interact and cooperate with others. Interactions between agents are modeled as a best response game wherein each agent negotiates to maximize their own utility. A dynamic game solution uses the Nash equilibrium, yielding an online method of predicting multi-agent interactions given their SVOs. This approach allows autonomous vehicles to observe human drivers, estimate their SVOs, and generate an autonomous control policy in real time.

First claim

What is claimed is:

1. A method for operating an autonomous vehicle in the presence of one or more mobile entities, the method comprising:
monitoring interactions between the one or more mobile entities and/or between the autonomous vehicle and the one or more mobile entities, including for each mobile entity of the one or more mobile entities, estimating a corresponding social behavior score for said entity based at least in part on the monitored interactions;
planning an operation of the autonomous vehicle using the estimated social behavior score for each of the one or more mobile entities; and
causing the autonomous vehicle to operate according to the planned operation,
wherein planning the operation of the autonomous vehicle includes performing a computer-implemented multi-agent game-theoretically based decision process
wherein estimating the social behavior score comprises estimating a Social Value Orientation (SVO).

2. The method of claim 1 , wherein monitoring the interactions of the one or more mobile entities comprises monitoring motion of said autonomous vehicle and/or said one or more mobile entities.

3. The method of claim 1 , wherein the autonomous vehicle comprises an autonomous automotive vehicle operating on a roadway.

4. The method of claim 1 , wherein the one or more mobile entities comprise an autonomous automotive vehicle.

5. The method of claim 1 , wherein the one or more mobile entities comprise a human-operated vehicle.

6. The method of claim 1 , wherein the one or more mobile entities comprises a pedestrian.

7. A method for operating an autonomous vehicle in the presence of one or more mobile entities, the method comprising:
monitoring operation of the one or more mobile entities, including for each mobile entity of the one or more mobile entities, estimating a corresponding social behavior score for said entity based at least in part on the monitored operation, wherein estimating the social behavior score comprises determining a quantity representing a degree of one or more of cooperative and selfish behavior by the mobile entity;
planning operation of the autonomous vehicle using the estimated social behavior score for each of the one or more mobile entities; and
causing the autonomous vehicle to operate according to the operation,
wherein estimating the social behavior score comprises estimating a Social Value Orientation (SVO).

8. The method of claim 7 , wherein estimating the social behavior score comprises determining a quantity representing a degree of one or more of altruistic, prosocial, egotistic, competitive, and sadistic, behavior by the mobile entity.

9. The method of claim 1 , wherein the social behavior score comprises a tuple of multiple component scores, each associated with a different behavior characteristic.

10. The method of claim 1 , wherein planning the operation of the autonomous vehicle includes, for at least one mobile entity, predicting future operation of the mobile entity based on the social behavior score for said mobile entity.

11. The method of claim 10 , wherein planning the operation of the autonomous vehicle includes, for each of the mobile entities, predicting future operation of the mobile entity based on the social behavior score for said mobile entity.

12. The method of claim 10 , wherein predicting the future operation of the at least one mobile entity includes predicting said future operation according to a combination of a predicted utility of said operation and a utility of planned operation of the autonomous vehicle.

13. The method of claim 12 , wherein the combination depends on the corresponding social behavior for said at least one mobile entity.

14. The method of claim 12 , wherein planning the operation of the autonomous vehicle includes evaluating a utility function for operation of a vehicle according to weights for respective operation characteristics.

15. The method of claim 14 , wherein the operation characteristics include one or more of road progress, comfort, desired velocities, tailgating penalty, collision avoidance, lane centering position, and road departure.

16. The method of claim 14 , further comprising forming the utility function based on prior monitoring of vehicles and/or other mobile entities.

17. The method of claim 16 , wherein forming the utility function comprises estimating the weights for respective operation characteristics.

18. The method of claim 12 , wherein planning the operation of the autonomous vehicle includes using a same utility function for each of the one or more mobile entities.

19. The method of claim 18 , wherein planning the operation of the autonomous vehicle includes using a utility function for the autonomous vehicle.

20. The method of claim 12 , wherein the utility function depends on a location of the operation of the autonomous vehicle on a roadway system.

21. The method of claim 1 , further comprising:
selecting a social behavior score for the autonomous vehicle; and
wherein planning the operation of the autonomous vehicle comprises planning said operation based on the selected social behavior score for the autonomous vehicle.

22. The method of claim 1 , wherein the operation of the autonomous vehicles comprises at least an operation selected from a group of operations consisting of a lane change operation, a merge operation, and a turn across oncoming traffic.

23. An autonomous vehicle comprising a machine-implemented controller configured to:
monitor interactions between one or more mobile entities and/or between the autonomous vehicle and the one or more mobile entities, including for each mobile entity of the one or more mobile entities, estimating a corresponding social behavior score for said entity based at least in part on the monitored interactions;
plan operation of an autonomous vehicle using the estimated social behavior score for each of the one or more mobile entities; and
cause the autonomous vehicle to operate according to the planned operation,
wherein planning the operation of the autonomous vehicle includes performing a computer-implemented multi-agent game-theoretically based decision process wherein estimating the social behavior score comprises estimating a Social Value Orientation (SVO).

24. A non-transitory machine-readable medium comprising instructions stored thereon, execution of the instructions by a data processing system causing:
monitoring interactions between one or more mobile entities and/or between the autonomous vehicle and the one or more mobile entities, including for each mobile entity of the one or more mobile entities, estimating a corresponding social behavior score for said entity based at least in part on the monitored interactions;
planning operation of an autonomous vehicle using the estimated social behavior score for each of the one or more mobile entities; and
causing the autonomous vehicle to operate according to the planned operation,
wherein planning the operation of the autonomous vehicle includes performing a computer- implemented multi-agent game-theoretically based decision process
wherein estimating the social behavior score comprises estimating a Social Value Orientation (SVO).

25. The method of claim 7 , wherein planning the operation of the autonomous vehicle includes performing a computer-implemented multi-agent game-theoretically based decision process.

26. An autonomous vehicle comprising a machi

Assignees

• Massachusetts Inst Technology

Inventors

• Rus Daniela
• Karaman Sertac
• Mora Javier Alonso
• Pierson Alyssa
• Schwarting Wilko

Classifications

• B60W60/00272 Primary

  relying on extrapolation of current movement · CPC title

• B60W30/18163

  Lane change; Overtaking manoeuvres · CPC title

• B60W60/0013

  specially adapted for occupant comfort · CPC title

• B60W60/0015

  specially adapted for safety · CPC title

• B60W2554/4029

  Pedestrians · CPC title