LLM as BT-Planner: Leveraging LLMs for Behavior Tree Generation in Robot Task Planning

2024-09-16

arXiv.org

AI Summary

Long-horizon robotic assembly tasks suffer from lengthy temporal dependencies and complex part interrelationships, rendering behavior tree (BT) planning heavily reliant on manual design—resulting in low efficiency and poor scalability. This paper proposes LLM-as-BT-Planner, the first systematic framework leveraging large language models (LLMs) to autonomously generate semantically valid, modular BTs. We introduce four novel in-context learning strategies and empirically validate the efficacy of supervised fine-tuning for small-scale LLMs on BT generation. To ensure executability and interpretability, our method jointly enforces BT syntactic constraints and instruction semantic parsing. Extensive experiments in simulation and on real robotic platforms demonstrate significant improvements over state-of-the-art LLM-based planners and handcrafted BT baselines: +28.6% task success rate, +34.1% BT structural accuracy, and enhanced execution robustness.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Planning with Language Models
• Intelligent Robots: Motion and Path Planning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Robotic assembly tasks remain an open challenge due to their long horizon nature and complex part relations. Behavior trees (BTs) are increasingly used in robot task planning for their modularity and flexibility, but creating them manually can be effort-intensive. Large language models (LLMs) have recently been applied to robotic task planning for generating action sequences, yet their ability to generate BTs has not been fully investigated. To this end, we propose LLM-as-BT-Planner, a novel framework that leverages LLMs for BT generation in robotic assembly task planning. Four in-context learning methods are introduced to utilize the natural language processing and inference capabilities of LLMs for producing task plans in BT format, reducing manual effort while ensuring robustness and comprehensibility. Additionally, we evaluate the performance of fine-tuned smaller LLMs on the same tasks. Experiments in both simulated and real-world settings demonstrate that our framework enhances LLMs' ability to generate BTs, improving success rate through in-context learning and supervised fine-tuning.

Problem

• Automates behavior tree generation for robotic assembly tasks.
• Reduces manual effort in creating behavior trees using LLMs.
• Enhances LLMs' ability to generate robust and comprehensible task plans.

Innovation

• LLMs generate Behavior Trees for robot tasks.
• In-context learning enhances BT generation efficiency.
• Fine-tuned smaller LLMs improve task success rates.