Erdal Kayacan

D-SafeMPC achieves safer and more efficient robotic planning by seamlessly integrating diffusion models with model predictive control, overcoming key limitations of both approaches.

You can now achieve robust thermal-only SLAM with a lightweight network trained on inexpensive non-radiometric thermal data, rivaling the performance of methods relying on costly radiometric cameras.

A hybrid event-based SLAM system, Edged USLAM, achieves consistently accurate localization in challenging illumination and structured environments where event-only and learning-based methods falter.

GIANT outperforms existing multi-robot navigation methods by fusing global path awareness with attentive local adjustments, achieving superior collision avoidance and efficiency in complex, dynamic environments.