technical and legal challenges for urban autonomous...
TRANSCRIPT
Technical and Legal Challenges forUrban Autonomous Driving
Seung-Woo Seo, Prof. Vehicle Intelligence Lab.Seoul National [email protected]
I. Main Challenges for Urban Autonomous DrivingI. Dilemma in Autonomous Driving
II. Approach to Human‐like DrivingI. Intention‐Aware Decision MakingII. Imitation Learning
III. Autonomous Driving Research in SNUI. 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
7
Driving course on Yeuido
5
4
3
2
1
6
7
Lane-change in heavy traffic
Crossing a double-yellow line to passby an illegally parked car
In urban environments, dilemma situations frequently occur
Decisions at a yellow traffic light
8
Dilemma in Autonomous Driving
9
Dilemma in Autonomous Driving
I. Legal aspect
II. Interactivity aspect
III.Technology aspect
3 Different Aspects
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
“AV violating the traffic law”
Interactive driving (ex. Lane cut‐in)
‐12‐
Interactivity Aspect
13
Human-Like Driving
Dilemma in Autonomous Driving
I. Legal aspectEX) Crossing a double‐yellow line to pass an
illegally parked car
II. Interactivity aspectEX) Lane‐change in heavy traffic
unsignalized intersection
III.Technology aspect
3 Aspects
Approach to Human-Like Driving
15
TASK 1. LANE‐CHANGE IN HEAVY TRAFFIC TASK 2. INTERSECTION TASK N. HIGHWAY
Single‐Task Policy 1
Policy Optimization
Single‐TaskPolicy 2
Policy Optimization
Single‐TaskPolicy N
Policy Optimization
Model for Decision Making
17
1tX
1tY
1t
A
R1tO tO
tY
t
tX
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 ,
, ,
, ,
Experimental Environment
18
18
SNU Campus roadTotal length : ~4km
행정대학원
국제대학원
기숙사삼거리
대운동장자동화
시스템
연구소
Start
Goal
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
PolicyDerivation
Learning Technique
, , ,
Mapping from states to actions(Supervised Learning) Reconstruct reward function
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
21
SNU Campus roadTotal length : ~4km
Imitation Learning
Autonomous Driving Research in SNU
[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 inUrban Environments
23
SNUverSNU Automated Drive
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