The Center for Smart Streetscapes (CS3) is pleased to announce a new supplemental award from the National Science Foundation supporting research on AI-based surrogate safety measures and their governance — and welcoming two new faculty members to the Center.
Welcoming Professors Zhengbo Zou and Henry Lam. CS3 is delighted to introduce Zhengbo Zou (Civil Engineering and Engineering Mechanics, Columbia University) and Henry Lam (Industrial Engineering and Operations Research, Columbia University), who join the Center to lead the supplement’s two interlocking research threads. Professor Zou brings deep expertise in computer vision and uncertainty-aware sensing for the built environment; Professor Lam is a leading researcher in uncertainty quantification, robust optimization, and simulation.
From measurement to governance. Surrogate safety measures — indicators like time-to-collision that reveal risk before crashes occur — allow cities to evaluate street designs proactively rather than waiting for collision records to accumulate. Professor Zou’s team will develop camera-only AI pipelines that extract these measures from streetscape video with calibrated confidence, building on CS3’s ongoing collaboration with NYCDOT and validating the work across the DataCity and COSMOS testbeds.
But safety measures are only as reliable as the AI that produces them. Professor Lam’s team will develop the governance layer: runtime-assurance and uncertainty-quantification methods that detect when AI predictions become unreliable and safely qualify or override them, with provable bounds on failure probability. Together, the two threads aim to establish new principles and practices for governing safety-critical AI in smart streetscapes — a framework for deciding not just what AI systems can measure, but when their outputs can be trusted to inform decisions that affect people on the street. The project embodies the convergence at the heart of CS3’s mission: situational-awareness models produce the measures, and governance methods certify them, so public agencies can act on AI outputs with confidence.
The supplement will support two graduate research assistants working at the intersection of urban safety analytics and trustworthy AI, with results integrated into CS3 testbed workflows and shared through peer-reviewed publication and innovation ecosystem partners.
Save the date: CS3 Student Conference. The award also funds a three-day, Center-wide student conference hosted at Florida Atlantic University in Boca Raton, bringing together students and postdocs from Columbia, Rutgers, Lehman College, and FAU around the theme of integrating AI into smart streetscape applications. Details and dates will be announced soon — students across all thrusts and institutions should stay tuned.