Evaluating Multimodal LLMs as Generalist Vision-Language-Action Agents for Drone Control: Commanding, Approaching, Track
Dataset Analysis
Introduces DroneCATS-Agent and DroneCATS benchmark to evaluate MLLMs as vision-language-action agents for drone control, finding small models navigate but fail at declaring arrival.
Provenance
Collected from papers.
Derived from paper: Evaluating Multimodal LLMs as Generalist Vision-Language-Action Agents for Drone Control: Commanding, Approaching, Tracking and Searching