Workshop on Data-Driven Approaches to Transportation: Bridging Research and Practice

10:15 - 10:30 Opening Remarks and Welcome

10:30 - 11:30 Invited Talk

Warren Powell (Princeton University & Optimal Dynamics) - From Reinforcement Learning to Sequential Decision Analytics, with Applications in Transportation and Logistics

11:30 - 12:00 Invited Talk

Ruixiao Sun (USC) - Online Transportation Network Cyber-Attack Detection Based on Stationary Sensor Data

12:00 - 13:00 Lunch

13:00 - 13:30 Invited Talk

Qi Luo (Clemson) - The Path to Multimodal Transportation Systems: A Look at the “System of Systems” Approach

13:30 - 14:15 Panel Discussion

Daniel B. Halsted (SCDOT), Qi Luo (Clemson), Yi Sun (USC), Biplav Srivastava (USC)

14:15 - 14:45 Invited Talk

Gurcan Comert (Benedict College) & Judith Mwakalonge (SC State) 

- The Determination of Pothole Sizes and Locations Using Artificial Intelligence and Vehicle Built-in Technologies

14:45 - 15:15 Invited Talk

Yi Sun (USC) - Kinetic Monte Carlo Simulations of Traffic Flows

15:15 - 15:30 Closing Remarks

Recently we have seen impressive advances and successes in applying artificial intelligence (AI) and machine learning (ML) for data-driven approaches to transportation analysis and management, example application areas including adaptive traffic signal control, advanced driver assistance systems and autonomous driving, vehicle-to-everything (V2X) communication, predictive analytics of on-demand mobility and ride sharing, transportation equity and sustainability, etc. Success in these applications has the potential to dramatically increase the safety and efficiency of transportation systems.

These recent advances of data-driven approaches to transportation systems have been primarily pioneered in research labs, while transferring them into the practice has been at a relatively slower pace. This gap is especially discernible when compared with other domains, such as vision and language processing and understanding, games, robotics, sciences, etc., where AI/ML-based data-driven approaches have already impacted real-world applications therein profoundly and at scale.

The factors causing this gap are compound, many of which are unique to transportation systems. For example, while high-fidelity and easy-to-use simulators, which are crucial for the recent AI/ML advances therein, are readily available or standardized in domains like games and biology, such simulators are still far from being practical for transportation analysis and control, since transportation systems are geographically and functionally distributed, often composing subsystems that are highly interdependent that operate on various spatio-temporal scales.

The aim of this workshop is to better understand and mitigate factors that cause the research-to-practice gap for data-driven approaches to transportation applications. We will gather people from the intelligent transportation community, including both researchers of multiple disciplines and practitioners from transportation agencies and local communities, to conduct talks, poster sessions, and discussions that will help fulfill the aim.

This workshop is sponsored by a National Science Foundation (NSF) planning grant from the CISE Community Research Infrastructure (CCRI) program. The University of South Carolina received this grant to lead an effort to design and build research infrastructure for data-driven transportation applications that is easy-to-use for CISE (Computer and Information Science and Engineering) researchers.