Presentation | 2020-12-11 An Automated Driving Strategy for Microscopic Road Traffic Using Multi-Agent Deep Reinforcement Learning Ryota Suwa, Toshiharu Sugawara, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This study proposes a method for interaction-aware automated driving in microscopic road traffic simulations using multi-agent deep reinforcement learning. In recent years, so many studies proposed an algorithm for automated driving using reinforcement learning, and some companies conducted experimental self-driving on public roads. However, existing reinforcement learning algorithms do not sufficiently consider interactions with other vehicles, where it is essential for safe driving on public roads. Therefore, we introduced a safety-oriented success rate and proposed a method that considers interactions between multiple agents in microscopic traffic environments. We also conducted an experimental evaluation in a driving simulation environment to see if the proposed method can improve self-driving safety. One experimental results suggest that it can achieve a higher success rate in microscopic road traffic situations, such as at an intersection and a junction, than existing methods that used single-agent reinforcement learning. Finally, we discuss the reasons for the improvement and future issues. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Multi-Agent / Deep Reinforcement Learning / Self-Driving |
Paper # | AI2020-7 |
Date of Issue | 2020-12-03 (AI) |
Conference Information | |
Committee | AI |
---|---|
Conference Date | 2020/12/10(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Online and HAMAMATSU ACT CITY |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Foundations and application technologies for AI systems on the new normal |
Chair | Naoki Fukuta(Shizuoka Univ.) |
Vice Chair | Yuichi Sei(Univ. of Electro-Comm.) / Yuko Sakurai(AIST) |
Secretary | Yuichi Sei(Nagoya Inst. of Tech.) / Yuko Sakurai(Tokyo Univ. of Agriculture and Technology) |
Assistant |
Paper Information | |
Registration To | Technical Committee on Artificial Intelligence and Knowledge-Based Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | An Automated Driving Strategy for Microscopic Road Traffic Using Multi-Agent Deep Reinforcement Learning |
Sub Title (in English) | |
Keyword(1) | Multi-Agent |
Keyword(2) | Deep Reinforcement Learning |
Keyword(3) | Self-Driving |
1st Author's Name | Ryota Suwa |
1st Author's Affiliation | Waseda University(Waseda Univ.) |
2nd Author's Name | Toshiharu Sugawara |
2nd Author's Affiliation | Waseda University(Waseda Univ.) |
Date | 2020-12-11 |
Paper # | AI2020-7 |
Volume (vol) | vol.120 |
Number (no) | AI-281 |
Page | pp.pp.34-38(AI), |
#Pages | 5 |
Date of Issue | 2020-12-03 (AI) |