Fluent Robotics Lab, University of Michigan
Date:
The Fluent Robotics Lab, directed by Prof. Christoforos Mavrogiannis, builds algorithms and systems that let robots work fluently with and around people in dynamic, unstructured environments.
Role:
Developed MPPI-based goal-intent expression algorithms in PyTorch for human-aware navigation, and studied how contextual legibility is affected by human distraction/attention in social navigation using the Hello Robot Stretch 2 platform (Rethinking Legibility in Social Robot Navigation, HRI 2026 LBR)
Contributing to a scenario-centric social navigation dataset (ACME) for training foundational navigation models, and to the Bi3 biplatform, bicultural dataset accepted to ICRA 2026
Formalizing information-theoretic models of implicit human-robot communication across multiple modalities
Contributing to open-source infrastructure for human-aware social navigation research in the lab