RT info:eu-repo/semantics/article
T1 Behavior tree generation and adaptation for a social robot control with LLMs
A1 Merino Fidalgo, Sergio
A1 Sánchez Girón, Celia
A1 Zalama Casanova, Eduardo
A1 Gómez García-Bermejo, Jaime
A1 Duque Domingo, Jaime
K1 Planning and execution
K1 Networks of robots and intelligent sensors
K1 Mobile robots
K1 Cognitive aspects of automation systems and humans
K1 Large language models
K1 33 Ciencias Tecnológicas
AB Large Language Models have recently emerged as a powerful tool for generating flexible and context-awarerobotic behavior. However, adapting to unforeseen events and ensuring robust task completion remainsignificant challenges. This paper presents a novel system that leverages LLMs and Behavior Trees to enablerobots to generate, execute, and adapt task plans based on natural language commands. The system employsChatGPT to process user instructions, generating initial Behavior Trees that encapsulate the required tasksteps. A modular architecture, combining the BT planner and a Failure Interpreter module, allows the systemto dynamically adjust Behavior Trees when execution challenges or environmental changes arise.Unlike conventional methods that rely on static Behavior Trees or predefined state machines, our approachensures adaptability by integrating a Failure Interpreter capable of identifying execution issues and proposingalternative plans or user clarifications in real time. This adaptability makes the system robust to disturbancesand allows for seamless human–robot interaction. We validate the proposed methodology using experimentson a social robot across various scenarios in our workplace, demonstrating its effectiveness in generatingexecutable Behavior Trees and responding to execution failures. The approach achieves an 89.6% success ratein a realistic home environment, highlighting the effectiveness of LLM-powered Behavior Trees in enablingrobust and flexible robot behavior from natural language input
PB Elsevier
SN 0921-8890
YR 2025
FD 2025
LK https://uvadoc.uva.es/handle/10324/78652
UL https://uvadoc.uva.es/handle/10324/78652
LA eng
NO Robotics and Autonomous Systems, 2025, vol. 194, p. 105165
NO Producción Científica
DS UVaDOC
RD 18-oct-2025