Researchers have developed GROW$^2$ (GROunding Which and Where), a novel approach to enable robots to use objects as tools creatively, even for tasks they weren't designed for. This method addresses the challenge of open-world affordance grounding by hierarchically splitting the process into semantic and geometric levels. GROW$^2$ utilizes Vision-Language Models for commonsense reasoning to select tools and relevant parts, and vision foundation models to precisely ground these parts in 3D space. Experiments demonstrate that GROW$^2$ surpasses current state-of-the-art methods in affordance prediction and shows strong generalization capabilities in both simulated and real-world robot tool use scenarios. AI
IMPACT Enhances robot adaptability and creative problem-solving by enabling flexible tool use beyond predefined functions.
RANK_REASON The cluster contains a research paper detailing a new method for robot tool use.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
