Infrastructure Systems: Digital Twins for Smart Streetscapes

Led by Jinwoo Jang, Ph.D.

This REU project aims to create AI-enabled hyperlocal digital twins, advancing the understanding of how multiple people interact with urban infrastructure environments and surrounding objects on the streets. The project will leverage various sensor data to better understand street-level agent behavior, micro-mobility, and their interactions with contextual surroundings (e.g., street objects, city infrastructure, traffic, air quality, and temperature). The objectives of this project include developing 1) situation awareness through the data fusion of heterogeneous datasets collected from the City of West Palm Beach, 2) multimodal multi-human behavior analytics, 3) highly realistic streetscape simulation environments, and 3) agent behavior modeling and prediction.

This project can ultimately improve our streetscape's public safety, design, and sustainability via digital twin technologies. The scientific contribution includes (1) the data fusion of heterogeneous data (e.g., point clouds, video, mobile sensing data) to advance the hyper localization of micro-mobility; (2) dynamic modeling of multimodal human behaviors with respect to contextual information (e.g., wayfinding based on nearby traffic and street objects); and (3) data-driven digital twin agents to advance the real-world representations in simulation environments. The impacts are significant, as a data-driven understanding of spatiotemporal variability in agent/vehicle mobility patterns and human-infrastructure interactions will contribute to advancing human-centric street morphology and design. Human-centric street environments will enhance urban livability, foster economic prosperity, and promote environmental sustainability.