Interested in CR2C2 activities and engagement opportunities?
12:00 - 1:00 PM
Zoom Link: https://ncat.zoom.us/j/91626835577?pwd=METHlYu26jbBzFNj5j6TJIEA1hn0Ki.1 (Registration not required)
Part 1: Enhancing Automated Vehicle Operation and Safety on Rural Roads - Challenges and Innovative Solutions
Speaker: Dr. Subhadeep Chakraborty
Abstract: Automated vehicle technologies have made rapid progress in urban settings, yet their safe and reliable deployment on rural roads remains a critical challenge. In this seminar, Dr. Subhadeep Chakraborty will present a comprehensive research effort focused on advancing Connected and Automated Vehicle (CAV) safety in rural environments characterized by poor road markings, limited infrastructure, and unpredictable driving conditions. The talk will introduce a rural-specific CAV architecture, along with two key innovations: a Perception Error Model (PEM) based on a modified FETSGAN framework to realistically capture sensor uncertainties, and GENESIS-RL, a reinforcement learning approach for generating naturalistic safety-critical edge cases. The seminar will highlight results from large-scale simulations and Vehicle-in-the-Loop experiments using instrumented vehicle platforms integrated with digital twins of Tennessee’s rural roadways. The discussion will focus on how AI-driven simulation and real-world testing can be combined to create scalable, data-driven safety validation frameworks that support safer automated mobility in rural America.
Part 2: Rural Road Infrastructure Assessment and Enhancement for Autonomous Vehicle Deployment
Speaker: Dr. Tesfamichael Getahun
Abstract: Rural road networks often lack the physical and digital infrastructure required for the safe and reliable operation of automated vehicle (AV) technologies. Inadequate or poorly maintained lane markings, inconsistent traffic signage, variable road surface conditions, limited GPS reliability, and uneven 5G network coverage pose challenges for automated driving systems. While such infrastructure may be acceptable for human drivers, its readiness for AV operation has not been systematically evaluated, contributing to limited focus to rural-focused AV infrastructure studies. This project developed a framework for evaluating rural road infrastructure readiness for AV deployment. The framework considers multiple physical and digital infrastructure elements, including lane markings, traffic signs, potholes, GPS accuracy, and 5G network coverage. To implement this framework, a software tool was developed to process, analyze, quantify, and visualize offline data collected using cameras, GPS, and 5G communication modules deployed on AV platform vehicles and commercially available ADAS-equipped vehicles in rural areas of Georgia and North Carolina. Through experimental evaluation, the framework and tool enabled the systematic identification and quantification of infrastructure limitations that affect/limit AV deployment. Together, these outcomes provide an objective, data-driven basis for assessing rural AV infrastructure readiness and informing future research, planning, and investment decisions.
Dr. Subhadeep Chakraborty is an Associate Professor in the Department of Mechanical and Aerospace Engineering at the University of Tennessee, Knoxville, where he leads research at the intersection of artificial intelligence, autonomous systems, and transportation safety. His work focuses on developing scalable, simulation-driven validation frameworks for Connected and Automated Vehicles (CAVs), with particular emphasis on perception uncertainty, reinforcement learning, and vehicle-in-the-loop experimentation. Dr. Chakraborty has served as Principal Investigator on numerous federally funded projects supported by the CR2C2, CSCRS, National Science Foundation, Department of Homeland Security, and Oak Ridge National Laboratory, including recent efforts to improve automated vehicle safety on rural and underserved roadways. His research combines digital twins, AI-based perception error modeling, and real-vehicle testing to address safety-critical challenges in real-world deployment. Dr. Chakraborty is an active collaborator with transportation agencies and national laboratories and is committed to translating research outcomes into practical, data-driven tools while training the next generation of engineers in intelligent mobility systems.
Dr. Tesfamichael Getahun is a Postdoctoral Researcher at the Autonomous Cooperative Control of Emergent Systems of Systems (ACCESS) Laboratory at North Carolina A&T State University. His research interests include computer vision, advanced driver-assistance systems (ADAS), and control for autonomous vehicles.