Rapid Conversational Authoring of Interactive Human-Robot Activities
Proceedings of the AAAI Symposium Series · 06.11.2026 · 10 (1) · 871-880
DOI: 10.1609/aaaiss.v10i1.43501 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Symposium Series, online erschienen am 06.11.2026. Verfasst von Shiye Cao und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Social robots are being envisioned for a growing range of applications. Yet their behavior is still predominantly authored by robot experts and developers through hand-crafted interaction content or, more recently, large language model (LLM) prompts. This places control over robot behavior with those who implement the interaction rather than the domain experts and end users who better understand how the robot should serve users’ evolving needs over time. Existing end-user authoring tools, such as visual programming environments, lower this barrier but still require specialized authoring skills, including constructing interaction logic through blocks and flows. We present the Robot Interaction Development Environment (RIDE), an open-source, end-to-end, mobile-compatible system that automates much of the human-robot interaction authoring process. From a natural-language description of an activity, RIDE decomposes the interaction into a sequence of states, generates the interaction flow and state-level prompts that drive the robot, and deploys the resulting program. We evaluated RIDE through three case studies in which a speech language therapist designed social-skills practice for a patient, a high school French teacher designed conversation practice for a student, and a nurse designed a stress-relief activity for herself. All three participants created a deployable activity within seven conversational turns without writing prompts or program logic. While participants readily delegated interaction structure to the system, they retained control over domain-specific details such as target phrases and content to avoid. We discuss how low-effort, conversational authoring can enable domain experts and end users to rapidly adapt general-purpose social robots to their goals and evolving needs.
Zitieren
APA
Cao, S., Tchamdja, A., & Huang, C. (2026). Rapid Conversational Authoring of Interactive Human-Robot Activities. Proceedings of the AAAI Symposium Series, 10(1), 871-880. https://doi.org/10.1609/aaaiss.v10i1.43501
Vancouver
Cao S, Tchamdja A, Huang C. Rapid Conversational Authoring of Interactive Human-Robot Activities. Proceedings of the AAAI Symposium Series. 2026;10(1):871-880. doi:10.1609/aaaiss.v10i1.43501
BibTeX
@article{Cao2026,
author = {Cao, Shiye and Tchamdja, Aureliane and Huang, Chien-Ming},
title = {Rapid Conversational Authoring of Interactive Human-Robot Activities},
journal = {Proceedings of the AAAI Symposium Series},
year = {2026},
volume = {10},
number = {1},
pages = {871--880},
doi = {10.1609/aaaiss.v10i1.43501}
}
EUVIMED