Technology Category
Intelligent Robots: ManipulationHumans and AI: Human-Aware Planning and Behavior PredictionMachine Learning: Transfer, Domain Adaptation, Multi-Task Learning
Task-Oriented Robot-Human Handovers on Legged Manipulators
🤖 AI Summary
This work addresses the limited generalizability of existing robot handover methods to novel object–task combinations and their neglect of human post-handover usage intent. The authors propose AFT-Handover, a framework that uniquely integrates affordance reasoning from large language models with texture-based point cloud feature transfer to enable zero-shot affordance transfer across objects and tasks. By modeling part-level correspondences, the method significantly enhances the efficiency of immediate human use after handover without requiring task-specific training. Experiments demonstrate higher handover success rates across diverse object–task pairs, and user studies confirm its substantial superiority over current approaches. The framework has been successfully deployed on a legged manipulation robot platform.
Technology Category
Application Category
📝 Abstract
Task-oriented handovers (TOH) are fundamental to effective human-robot collaboration, requiring robots to present objects in a way that supports the human's intended post-handover use. Existing approaches are typically based on object- or task-specific affordances, but their ability to generalize to novel scenarios is limited. To address this gap, we present AFT-Handover, a framework that integrates large language model (LLM)-driven affordance reasoning with efficient texture-based affordance transfer to achieve zero-shot, generalizable TOH. Given a novel object-task pair, the method retrieves a proxy exemplar from a database, establishes part-level correspondences via LLM reasoning, and texturizes affordances for feature-based point cloud transfer. We evaluate AFT-Handover across diverse task-object pairs, showing improved handover success rates and stronger generalization compared to baselines. In a comparative user study, our framework is significantly preferred over the current state-of-the-art, effectively reducing human regrasping before tool use. Finally, we demonstrate TOH on legged manipulators, highlighting the potential of our framework for real-world robot-human handovers.
Problem
Research questions and friction points this paper is trying to address.
Task-Oriented Handover
Human-Robot Collaboration
Affordance Generalization
Legged Manipulators
Zero-shot Transfer
Innovation
Methods, ideas, or system contributions that make the work stand out.
Task-Oriented Handover
Affordance Transfer
Large Language Model (LLM)
Zero-Shot Generalization
Legged Manipulator
🔎 Similar Papers
No similar papers found.
