SLIDE 1
T echnical and Legal Challenges for Urban Autonomous Driving
Seung-Woo Seo, Prof.
Vehicle Intelligence Lab. Seoul National University sseo@snu.ac.kr
SLIDE 2 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
2
SLIDE 3
Challenges for Urban Autonomous Driving
SLIDE 4 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
SLIDE 5
First Self-driving in City Road in Korea(2017. 6. 22)
SLIDE 6
Yeouido Area in Seoul
SLIDE 7 Demonstration at Yeouido Area in Seoul
7 Driving course on Yeuido 5 4 3 2 1 6 7
SLIDE 8 Lane-change in heavy traffic Crossing a double-yellow line to pass by an illegally parked car
In urban environments, dilemma situations frequently occur
Decisions at a yellow traffic light
8
Dilemma in Autonomous Driving
SLIDE 9 9
Dilemma in Autonomous Driving
- I. Legal aspect
- II. Interactivity aspect
III.Technology aspect
3 Different Aspects
SLIDE 10
Legal Aspect
Crossing a double-yellow line to pass by an illegally parked car
VS.
Crossing a double-yellow line illegal & socially compliant decision Waiting until an illegally parked car leaves legal & impractical decision
SLIDE 11
“AV violating the traffic law”
SLIDE 12
- Interactive driving (ex. Lane cut‐in)
‐12‐
Interactivity Aspect
SLIDE 13 13
Human-Like Driving
Dilemma in Autonomous Driving
EX) Crossing a double‐yellow line to pass an illegally parked car
EX) Lane‐change in heavy traffic unsignalized intersection III.Technology aspect 3 Aspects
SLIDE 14
Approach to Human-Like Driving
SLIDE 15 15
TASK 1. LANE‐CHANGE IN HEAVY TRAFFIC
TASK 2. INTERSECTION TASK N. HIGHWAY
Single‐Task Policy 1 Policy Optimization Single‐Task Policy 2 Policy Optimization Single‐Task Policy N Policy Optimization
SLIDE 16
SLIDE 17
- Model for Decision Making
17
1 t
X
1 t
Y
1 t
A R
1 t
O
t
O
t
Y
t
t
X A R
- The state space “S” is a joint space
- : Ego-vehicle’s state space
- : Other vehicles’ state space
- : Other vehicles’ driving intention
- The action space “A” : A = . , . , .
- The reward model
- Very high penalty when vehicle is predicted
to collide.
- Very high reward when vehicle arrives at its goal.
- Low penalty when vehicle moves at each step
Passing through intersection as fast as possible without any collision
Θ
,
,
SLIDE 18
18 18
SNU Campus road
Total length : ~4km
행정대학 원 국제대학 원 기숙사삼거리 대운동장
자동 화 시스 템 연구 소
Start Goal
SLIDE 19
- Learning from Expert Drivers
- Expert drivers understand human interactions on the road and comply with
mutually accepted rules, which are learned from countless experience
Brenna D. Argall, at el. “A survey of robot learning from demonstration”, Robotics and Autonomous Systems 57 (2009): 469‐483
Behavior Cloning Inverse Reinforcement Learning
Learning Technique Policy Derivation Learning Technique , , ,
Mapping from states to actions (Supervised Learning) Reconstruct reward function
19
Imitation Learning
SLIDE 20
- 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
SLIDE 21
- Experimental Environments
21
SNU Campus road
Total length : ~4km
Imitation Learning
SLIDE 22
Autonomous Driving Research in SNU
SLIDE 23 [November 19, 2013] Grand Prize in unmanned self‐driving car contest [November 4, 2015] Driverless taxi on SNU Campus [November 15, 2016] Door‐to‐Door Automated Driving on SNU Campus [June 22, 2017] Automated Driving in Urban Environments
23
SNUver
SNU Automated Drive
SLIDE 24
SNUver 1 (2015)
SLIDE 25
SNUver 2 (2016)
SLIDE 26
SNUvi (2017)
SLIDE 27
- 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
27
