Bio:
Dr. Javier Alonso-Mora is an Associate Professor at the Cognitive Robotics department of the Delft University of Technology , where he leads the Autonomous Multi-robots Laboratory . He is a Principal Investigator at the Amsterdam Institute for Advanced Metropolitan Solutions ( AMS Institute) and co-founder of The Routing Company . He is actively involved in the Delft robotics ecosystem, including the Robotics Institute, the Transportation Institute and Robovalley.

Before joining TU Delft, Dr. Alonso-Mora was a Postdoctoral Associate at the Computer Science and Artificial Intelligence Lab (CSAIL) of the Massachusetts Institute of Technology (MIT). He received his Ph.D. degree in robotics from ETH Zurich, where he worked in the Autonomous Systems Lab, and in partnership with Disney Research Zurich. Dr. Alonso-Mora holds a Diploma in Engineering and a Diploma in Mathematics from the Technical University of Barcelona ( UPC) , where he was part of the Interdisciplinary Higher Education Centre (CFIS) and the Robotics Institute (IRI).

His main research interest is in navigation, motion planning and control of autonomous mobile robots, with a special emphasis on multi-robot systems, on-demand transportation and robots that interact with other robots and humans in dynamic and uncertain environments. He has published over 60 papers in premier venues and his work has appeared in major media outlets. He is the recipient of multiple prizes and grants, including the ICRA Best Paper Award on Multi-robot Systems (2019), an Amazon Research Award (2019) and a talent scheme VENI award from the Netherlands Organisation for Scientific Research (2017).

Title: Social motion planning for autonomous vehicles

Abstract: We move towards an era of smart cities, where autonomous vehicles will provide on-demand transportation and make our streets safer - while coexisting with humans and other road users. The motion plan of the autonomous vehicle must therefore account for the interaction with other traffic participants and consider that they are, as well, decision-making agents. For example, when humans drive a car, they are fully aware of their environment and how other drivers and pedestrians may react to their future actions. In this talk I will give an overview of our work in motion planning among other decision-making vehicles, which combines reinforcement learning and trajectory optimization to be able to provide a certain level of guarantees while leveraging learning to achieve social behaviors. After a brief introduction to receding-horizon trajectory optimization framework, I will discuss three methods to enhance the motion planner with learning. Namely: using inverse reinforcement learning and social value estimation to achieve social behaviors; employing a learned policy to guide the motion planner in dense traffic scenarios; and achieving social trajectories by learning from a dataset of human-driven vehicles.

Bio:
Dr. Alexandre Bayen is the Liao-Cho Professor of Engineering at UC Berkeley. He is a Professor of Electrical Engineering and Computer Science, and Civil and Environmental Engineering. He is currently the Director of the Institute of Transportation Studies (ITS). He is also a Faculty Scientist in Mechanical Engineering, at the Lawrence Berkeley National Laboratory (LBNL). He received the Engineering Degree in applied mathematics from the Ecole Polytechnique, France, in 1998, the M.S. and Ph.D. in aeronautics and astronautics from Stanford University in 1998 and 1999 respectively. He was a Visiting Researcher at NASA Ames Research Center from 2000 to 2003. Between January 2004 and December 2004, he worked as the Research Director of the Autonomous Navigation Laboratory at the Laboratoire de Recherches Balistiques et Aerodynamiques, (Ministere de la Defense, Vernon, France), where he holds the rank of Major. He has been on the faculty at UC Berkeley since 2005. Bayen has authored two books and over 200 articles in peer reviewed journals and conferences. He is the recipient of the Ballhaus Award from Stanford University, 2004, of the CAREER award from the National Science Foundation, 2009 and he is a NASA Top 10 Innovators on Water Sustainability, 2010. His projects Mobile Century and Mobile Millennium received the 2008 Best of ITS Award for ‘Best Innovative Practice’, at the ITS World Congress and a TRANNY Award from the California Transportation Foundation, 2009. Mobile Millennium has been featured more than 200 times in the media, including TV channels and radio stations (CBS, NBC, ABC, CNET, NPR, KGO, the BBC), and in the popular press (Wall Street Journal, Washington Post, LA Times). Bayen is the recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE) award from the White House, 2010. He is also the recipient of the Okawa Research Grant Award, the Ruberti Prize from the IEEE, and the Huber Prize from the ASCE.

Title: Lagrangian control at large and local scales in mixed autonomy traffic flow
Abstract:
This talk investigates Lagrangian (mobile) control of traffic flow at local scale (vehicular level). The question of how self-driving vehicles will change traffic flow patterns is investigated. We describe approaches based on deep reinforcement learning presented in the context of enabling mixed-autonomy mobility. The talk explores the gradual and complex integration of automated vehicles into the existing traffic system. We present the potential impact of a small fraction of automated vehicles on low-level traffic flow dynamics, using novel techniques in model-free deep reinforcement learning, in which the automated vehicles act as mobile (Lagrangian) controllers to traffic flow.
Illustrative examples will be presented in the context of a new open-source computational platform called FLOW, which integrates state of the art microsimulation tools with deep-RL libraries on AWS EC2. Interesting behavior of mixed autonomy traffic will be revealed in the context of emergent behavior of traffic: https://flow-project.github.io

Bio:
Dr. Zhenhui (Jessie) Li is the Chief Scientist at Yunqi Academy of Engineering. Her research has been focusing on innovating spatial-temporal data mining techniques to address real-world problems in cities, environment, and society. She is now dedicated to work on city intelligence.
Dr. Li obtained her PhD in computer science from University of Illinois at Urbana-Champaign and was a tenured associate professor at the Pennsylvania State University.

Title: The Reality Gap: Lessons Learnt Traffic Signal Control

Abstract:
Motivated to address the real-world traffic problem, our research group has produced a series of studies on reinforcement learning for traffic signal control. However, our research results face difficulties to be implemented in the real world because of the reality gap between the simulation environment and the real environment. In this talk, I would like to share what I learnt in attempt to bridge the reality gap.

Bio:
Jun Luo studied computer science at Peking University and has a PhD in Computer Science and Cognitive Science from Indiana University Bloomington. He previously taught Cognitive Science at the University of Toronto and worked for several small and large companies. He joined Huawei Technologies Canada in 2016, where he currently serves as a Distinguished Researcher as part of Huawei Noah’s Ark Lab for Artificial Intelligence.

Title: Bootstrapping Behavior Simulation for Autonomous Driving
Abstract:
Simulations are essential for the development of autonomous driving solutions. In addition to simulation of physical dynamics and static environment structures, simulation of the interactive behaviors of social agents sharing the same traffic environment as the autonomous driving AI is also a huge challenge needing an adequate solution. Without adequate simulation of interactive behaviors of social agents, autonomous driving simulators will never be adequate for testing or training the autonomous driving AI. In this talk, we look at how to bootstrap the simulation of interactive behaviors for autonomous driving research without having to have already solved the challenge of building fully competent autonomous driving, or for that matter autonomous walking and biking, AI.

Bio:
Pradeep Varakantham is an Associate Professor in the School of Information Systems at Singapore Management University.  He also has a joint appointment as visiting researcher at Google Research, India. His research is at the intersection of Artificial Intelligence, Operations Research and Machine Learning with specific focus on solving (multi-agent) sequential planning/learning problems.  Pradeep has published extensively in top tier conferences and journals in Artificial Intelligence and Machine Learning. Apart from publishing his research,  he has led translational projects that have been successfully deployed and in use by multiple government agencies pertaining to transportation, safety and security in Singapore.  He is a senior member of Association for Advancement of Artificial Intelligence (AAAI), gave the "early career spotlight” invited talk at International Joint Conference on Artificial Intelligence, IJCAI 2016, has received google research award for projects on “ AI for Social Good”. His papers have won the best application, best-demo and finalist at ICAPS-19, AAMAS-2018 and INFORMS Innovative Applications in Analytics awards respectively. 

Title: Online Sequential Matching for Improving Efficiency in Urban Environments

Abstract:
Rapid “urbanization” (more than 50% of worlds’ population now resides in cities) coupled with the natural lack of coordination in usage of common resources (ex: bikes, ambulances, taxis, traffic personnel, attractions) has a detrimental effect on a wide variety of response (ex: waiting times, response time for emergency needs) and coverage metrics (ex: predictability of traffic/security patrols) in cities of today. Motivated by the need to improve response and coverage metrics in urban environments, I will talk about our research on building intelligent agent systems that make sequential decisions to continuously match available supply of resources to an uncertain demand for resources. Our broad approach to generating these sequential decision strategies is through a combination of machine learning and sequential optimization methods, while exploiting structure (anonymity, limited influence, abstraction etc.) in problems. I will specifically delve into our recent research on developing online optimization methods (including Deep RL) that can handle operational constraints and show some of our recent results on real world datasets.

Bio:
Prof. Hai Yang is currently a Chair Professor at The Hong Kong University of Science and Technology. He is internationally known as an active scholar in the field of transportation, with more than 260 papers published in SCI/SSCI indexed journals and a SCI H-index citation rate of 60. Most of his publications appeared in leading international journals, such as Transportation Research, Transportation Science and Operations Research. Prof. Yang received a number of national and international awards, including Frank M. Masters Transportation Engineering Award, American Society of Civil Engineers (2020); National Natural Science Award bestowed by the State Council of PR China (2011). He was appointed as Chang Jiang Chair Professor of the Ministry of Education of PR China; Prof. Yang served as the Editor-in-Chief of Transportation Research Part B: Methodological from 2013 to 2018 and is now a distinguished editorial board member of this prestigious journal.

Title: Learning to delay for ride-sourcing services

Abstract:
Ride-sourcing services are now reshaping the way people travel by effectively connecting drivers and passengers through mobile internets. The matching time interval (the time interval over which waiting passengers and idle drivers are accumulated and then subjected to peer-to-peer matching) and matching radius (or maximum allowable pick-up distance, within which waiting passengers and idle drivers can be matched or paired) are two key control variables that a platform can employ to optimize system performance in an online matching system. In this study, we propose an optimization model to dynamically adjust the matching time interval and matching radius in response to real-time supply and demand conditions. We show that the flexible matching time interval and matching radius in the system level have potentials to improve overall system performance. In the individual level, a specific passenger request can benefit from a delayed matching since he/she may be matched with closer idle drivers after waiting for a few seconds. Motivated by the potential benefits of delayed matching, we further establish a two-stage framework which incorporates a combinatorial optimization and multi-agent deep reinforcement learning methods. The multi-agent reinforcement learning methods are used to dynamically determine the delayed time for each passenger request (or the time at which each request enters the matching pool), while the combinatorial optimization conducts an optimal bipartite matching between idle drivers and waiting passengers in the matching pool. The proposed framework is shown to substantially improve system performances, by well balancing the trade-off among pick-up time, matching time and successful matching rate.