ERI: Systems of Systems Approach for Safe and Energy Efficient Control of Humans, Vehicles, and Transportation Systems

About This Grant

This Engineering Research Initiation (ERI) award supports research that will focus on a system of systems approach for human-autonomy collaboration using integration of real-world driving with sensing infrastructure to understand traffic dynamics. It is planned that findings will be used for accurate modeling of intelligent traffic and vehicle control strategies. Transportation is a key driver of smart cities, enabling social development and the movement of goods and people across origins and destinations. However, the random nature of people and vehicle movements makes traffic flow systems stochastic and unpredictable. This research project brings together multidisciplinary expertise in the built environment and in systems modeling to explore creation of a new design paradigm for a future connected built environment that learns from human autonomy collaboration using real autonomous driving scenes. The planned research advances the NSF mission of promoting societal benefits by improving energy efficiency and safety. This project would provide opportunities for undergraduate and graduate students to participate in research efforts and foster a new generation of emerging mobility professionals. The project efforts would develop a gaming concept “Collaborative Racing with Humans and Autonomous Vehicles (CRAV)” to help students and public learn about the concept of human autonomy collaboration and improve broader adoption of emerging mobility technologies. The emergence of advanced driving features and diffusion of autonomous vehicles (AVs) across smart and connected communities is promising in terms of improving traffic flow. However, reducing crashes resulting from human drivers and perception reaction failure, termed human-autonomy collaboration, and, more specifically, interaction of autonomous vehicles with human drivers is challenging. While focusing on centralized control offers potential benefits in improving traffic flow, it can suffer from the limitation of a single point of failure. This project will use a multi-faceted approach to conduct fundamental research in three areas: 1) developing a system of systems approach based on reinforcement learning agents for decentralized cooperative control strategies for human autonomy collaboration that learns from naturalistic and autonomous driving, 2) developing optimization and deep learning-based graph neural networks for recreating safety critical events from real autonomous driving scenes, and 3) developing simulated, connected and virtual reality-based advanced human-in-the loop driving features to test whether autonomous features render human driver behavior passive or aggressive. This project leverages rigorous science-based systems engineering in an attempt to move the state-of-the-art well beyond the current focus on centralized control. The mathematical techniques and empirical contributions of this project will enable future research on a broader class of problems related to human autonomy collaboration and intelligent transportation systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.