RoboDawn Transfers VLM Intelligence to Robotic Control

RoboDawn utilizes a compact interface of discrete translation, rotation, and gripper commands to enable vision-language models to control robots in closed-loop environments. The system allows VLMs to reason about visual states and adapt actions based on resulting physical feedback.