CPS: Small: Reachability Analysis for Nonlinear Cyber-Physical Systems.

NSF Award Search · 01002627DB NSF RESEARCH & RELATED ACTIVIT · $600,000 · view on nsf.gov ↗

Abstract

Modern cyber-physical systems — such as autonomous vehicles, medical devices, and power grid controllers — must operate safely despite uncertainty in their environments, models, and decision-making components. A key challenge in ensuring their safety is predicting all possible future behaviors of these systems to verify that unsafe conditions are avoided. Doing so, in practice, allows us to encounter unforeseen and potentially dangerous behaviors in these systems. This problem, known as reachability analysis, is central to the design and certification of safety-critical systems. However, existing methods struggle to scale to complex, nonlinear systems and are especially challenged when systems incorporate artificial intelligence (AI) components. This project addresses a fundamental gap by developing new mathematical and computational tools for analyzing such systems under uncertainty. The proposed advances will improve the ability to provide strong safety guarantees for emerging technologies, particularly AI-enabled autonomous systems that are increasingly deployed in domains critical to national health, economic prosperity, and public safety. By enabling more reliable verification of these systems, the project contributes to safer transportation, more robust medical technologies, and resilient infrastructure. In addition to its research contributions, the project will support education and workforce development through open-source software, benchmark problems, and a summer s

Key facts

NSF award ID
2534982
Awardee
University of Colorado at Boulder (CO)
SAM.gov UEI
SPVKK1RC2MZ3
PI
Sriram Sankaranarayanan
Primary program
01002627DB NSF RESEARCH & RELATED ACTIVIT
All programs
Artificial Intelligence (AI), CYBER-PHYSICAL SYSTEMS (CPS), SMALL PROJECT
Estimated total
$600,000
Funds obligated
$600,000
Transaction type
Standard Grant
Period
06/01/2026 → 05/31/2029