[CoRL 2020 Best System Paper] SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving

CoRL 2020

[CoRL 2020 Best System Paper] SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving

Dec 16, 2020
|
96 views
|
|
Code
Details
"**SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving** Ming Zhou (Shanghai Jiao Tong University); Jun Luo (Huawei Technologies Canada Co. Ltd.)*; Julian Villella (Independant); Yaodong Yang (Huawei); David Rusu (Huawei); Jiayu Miao (Shanghai Jiao Tong University); Weinan Zhang (Shanghai Jiao Tong University); Montgomery Alban (Huawei); IMAN FADAKAR (HUAWEI TECHNOLOGIES CANADA); Zheng Chen (Huawei); Chongxi Huang (Huawei Technologies); Ying Wen (Huawei); Kimia Hassanzadeh (Huawei); Daniel Graves (Huawei); Zhengbang Zhu (Huawei Noah’s Ark Lab); Yihan Ni (Huawei Technologies); Nhat Nguyen (Huawei); Mohamed Elsayed (Huawei Technologies Canada Co., Ltd); Haitham Ammar (Huawei); Alexander Cowen-Rivers (Huawei R&D UK); Sanjeevan Ahilan (Independant); Zheng Tian (Huawei); Daniel Palenicek (Huawei); Kasra Rezaee (University of Toronto); Peyman Yadmellat (Huawei Technologies Canada); Kun Shao (Huawei Noah's Ark Lab); dong chen (Huawei Technologies Co., Ltd.); Baokuan Zhang (Huawei Technologies Co., Ltd.); Hongbo Zhang (Huawei Noah's Ark Lab); Jianye Hao (Tianjin University); Wulong Liu (Huawei Noah's Ark Lab); Jun Wang (UCL) Publication: https://corlconf.github.io/paper_53/ **Abstract** Interaction is fundamental in autonomous driving (AD). Despite more than a decade of intensive R&D in AD, how to dynamically interact with diverse road users in various contexts still remains unsolved. Multi-agent learning has recently seen big breakthroughs and has much to offer towards solving realistic interaction in AD. However, to realize this potential we need multi-agent AD simulation of realistic interaction. To break this apparent chicken-and-egg circularity, we built an AD simulation platform called SMARTS (Scalable Multi-Agent Rl Training School), which is designed to accumulate behavior models of road users towards increasingly realistic and diverse interaction that in turn enables deeper and broader multi-agent research on interaction. In this paper, we describe the design goals of SMARTS, explain its key architectural ideas, illustrate its use for multi-agent research through experiments on concrete interaction scenarios, and introduce a set of benchmarks and metrics. As an open-source, industrial-strength platform, the future of SMARTS lies in its growth along with the multi-agent research it enables in the years to come.

Comments
loading...