☕️Assistant Research Scientist
Engineering Systems Group, University of Michigan Transportation Research Institute (UMTRI)
Efe Yarbasi is an Assistant Research Scientist in the Engineering Systems Group at the University of Michigan Transportation Research Institute (UMTRI). His research covers the safety of automated and driver-assistance systems and the electrification of transportation. He serves as principal investigator and co-principal investigator on projects sponsored by NHTSA, Toyota Motor North America, and the National Cooperative Highway Research Program. He co-leads the Low Altitude Airspace Working Group at M-Air.
He holds a PhD in Aerospace Engineering from Georgia Tech. There he worked in the Aerospace Systems Design Laboratory under Prof. Dimitri Mavris and developed methods to find and reduce the uncertainties that matter most in complex, multi-disciplinary designs. He is a member of AIAA and IEEE.
Students, please get in touch. I am looking for curious, motivated students to join my research. Projects range from crash analysis and AI for safety to EV charging and low-altitude airspace, and I mentor undergraduate and graduate students. If any of this sounds interesting, send me an email with a short note about your background and what you want to work on.
My work sits at the intersection of vehicle safety, electrified transportation, and systems engineering.
Estimating and assuring the real-world safety of automatic emergency braking, driver assistance, and automated driving, using crash data, naturalistic driving data, and simulation.
Using large language models and multi-agent AI to reconstruct and reason about pre-crash events from crash records.
Multi-modal electrification of transportation systems and their integration into broader mobility networks, from EV charging to advanced and regional air mobility.
Methods for finding the uncertainties that matter most in complex, multi-disciplinary designs and for designing computational and physical experiments to reduce them, from my PhD at Georgia Tech.
A two-phase multi-agent LLM framework that reconstructs pre-crash scenarios from crash reports, scene diagrams, and event data recorder records (277 CISS rear-end crashes).
A methodology for designing computational or physical experiments that mitigate system-level uncertainty, demonstrated on an early-stage Blended-Wing-Body aircraft concept.
A systematic methodology to identify and mitigate epistemic uncertainty in complex, multi-disciplinary aircraft design.
Identifying the critical system-level uncertainties in a coupled aerostructural analysis to decide which validation experiments to run.