Max Robotics — Global Service Robot Directory + US Certification

About

Devendra Singh Chaplot is a pioneering researcher in embodied AI and autonomous robot navigation, whose work has fundamentally advanced how intelligent agents perceive, explore, and interact with real-world environments. His research centers on semantic navigation, modular learning systems, and embodied active learning — areas where he has consistently bridged the gap between simulation and real-world deployment. Chaplot's most influential contribution, "Object Goal Navigation using Goal-Oriented Semantic Exploration" (2020, 221 citations), introduced a modular framework that overcomes the long-term planning limitations plaguing end-to-end learning approaches. This work established a foundational paradigm for navigating to object categories in unseen environments. His subsequent research extended this vision through universal navigation systems like GOAT, which enables robots to reach targets specified via language, images, or object categories — a critical step toward versatile home robotics. Chaplot has also made substantial infrastructure contributions through the Habitat simulation platforms (Habitat 2.0 and 3.0), enabling the broader research community to train and benchmark embodied agents in increasingly realistic, physics-enabled, and human-populated environments. His self-supervised learning framework SEAL further demonstrates his commitment to reducing reliance on costly labeled data. Collectively, his work reflects a systematic effort to make autonomous navigation practical, generalizable, and deployable in everyday unstructured environments.