A Force/Torque Taxonomy for Classifying States During Physical Co-Manipulation.

Abstract

Achieving seamless human-robot collaboration requires a deeper understanding of how agents manage and communicate forces during shared tasks. Force interactions during collaborative manipulation are inherently complex, especially when considering how they evolve over time. To address this complexity, we propose a taxonomy of decomposed force and torque components, providing a structured framework for examining haptic communication and informing the development of robots capable of performing meaningful collaborative manipulation tasks with human partners. We propose a standardized terminology for force decomposition and classification, bridging the varied language in previous literature in the field, and conduct a review of physical human-human interaction and haptic communication. The proposed taxonomy allows for a more effective and nuanced discussion of important force combinations that we expect to occur during collaborative manipulation (between human-human or human-robot teams). We also include example scenarios to illustrate the value of the proposed taxonomy in describing interactions between agents.