T echnical and Legal Challenges for Urban Autonomous Driving - PowerPoint PPT Presentation

T echnical and Legal Challenges for Urban Autonomous Driving Seung-Woo Seo, Prof. Vehicle Intelligence Lab. Seoul National University sseo@snu.ac.kr

I. Main Challenges for Urban Autonomous Driving I. Dilemma in Autonomous Driving II. Approach to Human‐like Driving I. Intention‐Aware Decision Making II. Imitation Learning III. Autonomous Driving Research in SNU I. Demonstration of SNUver IV. Conclusion 2

Challenges for Urban Autonomous Driving

Considerations for Urban Autonomous Driving  Moving & static objects • Pedestrians • Other vehicles • Traffic light & signs • Unforeseen events  Crossing intersection  Turning  Lane changes  Parking  Entering and exiting drop off stations  Etc.

First Self-driving in City Road in Korea(2017. 6. 22)

Yeouido Area in Seoul

Demonstration at Yeouido Area in Seoul Driving course on Yeuido 2 3 1 4 7 5 6 7

Dilemma in Autonomous Driving In urban environments, dilemma situations frequently occur Crossing a double-yellow line to pass by an illegally parked car Lane-change in heavy traffic Decisions at a yellow traffic light 8

Dilemma in Autonomous Driving 3 Different Aspects I. Legal aspect II. Interactivity aspect III.Technology aspect 9

Legal Aspect Crossing a double-yellow line to pass by an illegally parked car Crossing a double-yellow line Waiting until an illegally parked car leaves VS. illegal & socially compliant decision legal & impractical decision

“AV violating the traffic law”

Interactivity Aspect  Interactive driving (ex. Lane cut‐in) ‐12‐

Dilemma in Autonomous Driving 3 Aspects I. Legal aspect EX) Crossing a double‐yellow line to pass an illegally parked car II. Interactivity aspect EX) Lane‐change in heavy traffic Human-Like Driving unsignalized intersection III.Technology aspect 13

Approach to Human-Like Driving

TASK 1. LANE‐CHANGE IN HEAVY TRAFFIC TASK 2. INTERSECTION TASK N. HIGHWAY Policy Optimization Policy Optimization Policy Optimization Single‐Task Policy Single‐Task Single‐Task 1 Policy 2 Policy N 15

 Model for Decision Making  The state space “S” is a joint space � � � � �  � : Ego-vehicle’s state space ��� � � � ��� , � � ��� , � � ��� � � A A  � : Other vehicles’ state space � � � � � , � � � , � � � � � X   � : Other vehicles’ driving intention X 1 t t � � Θ � �� ���� , � ��� � R R O  O  The action space “A” : A = ���. , ���. , �����. 1 t t  The reward model  Very high penalty when vehicle is predicted Y  Y 1 t t to collide.  Very high reward when vehicle arrives at its goal.  Low penalty when vehicle moves at each step    1 t t Passing through intersection as fast as possible without any collision 17

 Experimental Environment SNU Campus road Total length : ~4km 기숙사삼거리 국제대학 원 Start 행정대학 원 자동 대운동장 화 시스 템 연구 Goal 소 18 18

Imitation Learning  Learning from Expert Drivers • Expert drivers understand human interactions on the road and comply with mutually accepted rules, which are learned from countless experience Behavior Cloning Inverse Reinforcement Learning � � � �,� � � � ���� � � � �,� � ���, �� � � ���� Learning Learning Policy Technique Technique Derivation Mapping from states to actions Reconstruct reward function (Supervised Learning) Brenna D. Argall, at el. “A survey of robot learning from demonstration”, Robotics and Autonomous Systems 57 (2009): 469‐483 19

Imitation Learning  Driving dilemma in single lane road • Crossing a double-yellow line to pass by an illegally parked car Demonstration of expert drivers Sang‐Hyun Lee and Seung‐Woo Seo, “A Learning‐Based Framework for Handling Dilemmas in Urban Automated Driving”, IEEE International Conference on Robotics and Automation(ICRA), 2017 20

Imitation Learning  Experimental Environments SNU Campus road Total length : ~4km 21

Autonomous Driving Research in SNU

SNUver SNU Automated Drive [November 19, 2013] [June 22, 2017] Automated Driving in Grand Prize in unmanned self‐driving car contest Urban Environments [November 4, 2015] [November 15, 2016] Driverless taxi on Door‐to‐Door Automated SNU Campus Driving on SNU Campus 23

SNUver 1 (2015)

SNUver 2 (2016)

SNUvi (2017)

 Discussed several key issues related to dilemma in urban autonomous driving  Briefly introduced our learning-based approaches to human-like driving  There still remain many challenges that make the urban autonomous driving very hard  Future Work 27