Federal Project Grant Award Summary

The National Science Foundation (NSF) Directorate for Engineering awarded $225,000 to The Ohio State University (Columbus, OH) under CFDA 47.041—Engineering—to support collaborative research on "Reinforcement Learning with High-Probability Safety Constraints: Theory, and Applications." The three-year project (July 1, 2026–June 30, 2029) will develop foundational theory and algorithms for safe reinforcement learning (RL) systems capable of operating in safety-critical cyber-physical applications such as unmanned aerial vehicles, mobile robots, and networked infrastructure. The research will explicitly integrate probabilistic safety guarantees into learning methods, enabling these systems to respect stringent safety constraints while adapting to uncertain conditions, limited data, and changing operational environments—addressing a critical gap in current RL methods that typically optimize only average performance rather than high-confidence safety outcomes.

The project will produce intellectual contributions including novel algorithms and theoretical frameworks for safe decision-making under model uncertainty and environmental shifts, alongside broader-impact deliverables such as open-source software, benchmark scenarios, and educational curriculum modules to advance the field of reliable autonomous systems. These outputs are designed to support national priorities in reliable autonomy relevant to health, prosperity, welfare, and defense while building workforce capacity in STEM through direct student training and research experiences.