Veronica Chatrath

University of Toronto, Technical University of Munich

Papers

Total Citations

H-Index

About

Veronica Chatrath is a roboticist advancing the frontier of long-term autonomous navigation in dynamic, semi-static environments. Her research lies at the critical intersection of probabilistic mapping, object-level change detection, and semantically safe control. In her highly cited 2022 work, "POCD: Probabilistic Object-Level Change Detection and Volumetric Mapping in Semi-Static Scenes," Chatrath introduced a novel framework that enables robots to maintain up-to-date maps by probabilistically detecting and integrating scene changes during repeated traversals. This work, garnering 14 citations, directly addresses a fundamental challenge: undetected environmental changes degrade map quality and cause localization failures. Building on this foundation, her 2024 paper, "Closing the Perception-Action Loop for Semantically Safe Navigation in Semi-Static Environments," tackles the gap between theoretical safety guarantees and real-world deployment. Chatrath demonstrates that modern control paradigms often assume known, static extrinsic constraints—an assumption that breaks down in changing environments. By closing the perception-action loop with semantic understanding, she enables robots to adapt navigation strategies on the fly. Her contributions are pivotal for applications ranging from warehouse logistics to domestic service robots, ensuring that autonomous systems can operate safely and reliably over extended periods without human intervention.