An Ontology for Human-Human Interactions and Learning Interaction Behavior Policies

Abstract

Robots are expected to possess similar capabilities that humans exhibit during close proximity dyadic interaction. Humans can easily adapt to each other in a multitude of scenarios, ensuring safe and natural interaction. Even though there have been attempts to mimic human motions for robot control, understanding the motion patterns emerging during dyadic interaction has been neglected. In this work, we analyze close-proximity human-human interaction and derive an ontology that describes a broad range of possible interaction scenarios by abstracting tasks and using insights from attention theory. This ontology enables us to group interaction behaviors into separate cases, each of which can be represented by a particular graph. Using imitation learning, we train unique interaction policies with recurrent neural networks for each case. The ontology offers a unified and generic approach to categorically analyze and learn close-proximity interaction behaviors that can both be utilized as base models for future studies and enhance natural human-robot collaboration. CCS Concepts: • Human-centered computing → Collaborative interaction; Interaction design theory, concepts and paradigms; • Computing methodologies → Robotic planning;

ROSGPT: Next-Generation Human-Robot Interaction with ChatGPT and ROS

“…When developing prompts for human-robot interaction, it is crucial to consider the development of appropriate ontologies for context-specific applications. This is necessary to facilitate the accurate mapping between unstructured and structured command data, ultimately enhancing the efficiency and effectiveness of the interaction [17].…”

ROSGPT: Next-Generation Human-Robot Interaction with ChatGPT and ROS

“…When developing prompts for human-robot interaction, it is crucial to consider the development of appropriate ontologies for context-specific applications. This is necessary to facilitate the accurate mapping between unstructured and structured command data, ultimately enhancing the efficiency and effectiveness of the interaction [16].…”

MILD: Multimodal Interactive Latent Dynamics for Learning Human-Robot Interaction

“…3) Interaction Modeling with Recurrent Neural Networks: When large datasets are available, Recurrent latent space models are powerful tools in approximating latent dynamics with some form of a forward propagating distribution [23], [30], [34], [42], [46]. Given their power of modeling temporal sequences, they yield themselves naturally for learning interactive/collaborative tasks in HRI [58], [74], [86]. To simplify the various scenarios of interaction tasks Oguz et al [58] develop an ontology to categorize interaction scenarios and Latent Space HMM Conditioning…”

“…Given their power of modeling temporal sequences, they yield themselves naturally for learning interactive/collaborative tasks in HRI [58], [74], [86]. To simplify the various scenarios of interaction tasks Oguz et al [58] develop an ontology to categorize interaction scenarios and Latent Space HMM Conditioning…”