REU Mentor: Dr. Jason O. Hallstrom
This project will develop a reality-coupled, GIS-grounded 3D digital twin of downtown West Palm Beach, Florida. The photorealistic twin will mirror the physical world based on real-time sensor information and support prediction of future conditions. The twin will represent buildings, environmental features, pedestrians, vehicles, scooters, litter, and other elements of the urban landscape, using privacy-preserving representations that avoid revealing individual identities.
A small team of undergraduate researchers will develop the end-to-end system, leveraging an existing network of Wi-Fi-based pedestrian sensors and intelligent cameras. Students will gain experience with generative AI, 3D modeling, large-scale software development, and physics-based, probabilistic, and AI-driven prediction techniques. Some of the key research challenges include determining how faithfully the digital twin can represent real-world conditions, how accurately it can predict future conditions, and how uncertainty in observations and predictions can be quantified and communicated.
Students are expected to deploy a working end-to-end prototype in West Palm Beach, advancing the mission of the NSF Center for Smart Streetscapes – supporting urban planning, public safety, mobility, and community engagement. Students should have strong programming skills and some familiarity with generative AI technologies.