Rohan Sinha

Stanford University, Vaughn College of Aeronautics and Technology

Papers

Total Citations

101

H-Index

About

Rohan Sinha is a researcher at the forefront of trustworthy autonomy, specializing in the intersection of machine learning, robotics, and formal safety assurance. His work addresses a critical challenge: how to ensure that robots remain safe when their perception systems encounter unfamiliar, out-of-distribution (OOD) data. Sinha’s most-cited paper, “Semantic Anomaly Detection with Large Language Models” (2023, 85 citations), pioneers the use of LLMs to detect semantic anomalies in real-time, offering a scalable solution for identifying when a robot’s environment deviates from expected norms. His earlier work, “A System-Level View on Out-of-Distribution Data in Robotics” (2022, 7 citations), provides a foundational framework for understanding OOD challenges across the autonomy stack. In “Closing the Loop on Runtime Monitors with Fallback-Safe MPC” (2023, 9 citations), Sinha introduces a novel method that integrates runtime monitoring with model predictive control, enabling robots to safely fall back to conservative behaviors when perception uncertainty is high. Collectively, his research has garnered over 100 citations, establishing him as a rising authority on closing the safety loop for learning-enabled robots.