How to Plan `Good’ Paths for Urban Autonomous Driving?
Over the decade, researchers have been investigating viable technologies to realize autonomous driving in urban environments. While large improvement has been made in technologies, there still remain many challenges that make autonomous driving hard in urban situations. Some of them are accurate detection and tracking of a large number of moving objects and reliable path planning in real time. Especially for path planning, a great amount of research works has been published based on the recent progresses in Reinforcement Learning. In addition, due to the advances of Deep Leaning technologies, large-scale and generalized solutions have been feasible for various applications. In this talk, I discuss a few key issues of urban autonomous driving and introduce some recent technological advances. I mainly focus on the path planning problem and discuss how Deep Reinforcement Learning techniques can be applied to autonomous driving. I introduce several prominent learning-based approaches to find optimal and safe paths, and describe their cons and pros for achieving the computational and sample efficiency while guaranteeing safety, which are critical in urban autonomous driving.
Seung-Woo Seo is a professor in the Department of Electrical Engineering at the Seoul National University, and Director of Intelligent Vehicle IT (IVIT) Research Center in Seoul. He received his Ph.D. from the Pennsylvania State University, and B.S. and M.S. degrees from Seoul National University, all in Electrical Engineering. He was a Faculty of the Department of Computer Science and Engineering at the Pennsylvania State University, and served as a Member of the Research Staff in the Department of Electrical Engineering at Princeton University. He was a visiting professor at Stanford University, Center of Automotive Research in 2014. He has served as a Chair or Committee Member in numerous international conferences and workshops. He was the General Co-chair of 2015 IEEE Intelligent Vehicle Symposium held in Seoul, and has been serving as a Member of the Steering Committee of the IEEE Transactions on Intelligent Vehicles. He has also served as an Organizing Committee Chair of the International Unmanned Solar Vehicle Challenge in 2012, and served for five years as a Director of the Information Security Center at the Seoul National University. He has been very active in the research of autonomous driving and has been making key technological achievements especially in autonomous driving in urban areas. He has developed an autonomous driving car called SNUver in 2015 and has successfully demonstrated autonomous driving in the city roads of Seoul in 2017 for the first time in Korea. His research areas include autonomous driving and information security.
AI and Robotics in LGE
LG전자는 “고객을 위한 가치창조”와 “인간존중의 경영”이라는 경영 이념으로 지금까지 고객의 삶의 질 증진과 새로운 경험을 제공해 왔습니다.
로봇 사업을 미래 성장동력으로 육성하기 위해 LG전자는 상업용에서 가정용에 이르기까지 다양한 로봇을 개발하고 새로운 사업 기회를 지속 모색하고 있습니다.
AI 등 신기술과 로봇의 융합을 통해 고객들이 일상에서 새로운 가치와 경험을 느낄 수 있도록 준비중인 내용을 공유드리고자 합니다.
주요 수상 실적