VernaCopter: Disambiguated Natural-Language-Driven Robot via Formal Specifications

2024-09-14

arXiv.org

AI Summary

Unreliable motion planning for unmanned aerial vehicles (UAVs) arises from inherent ambiguity in natural language (NL) instructions and the stochasticity of large language models (LLMs). Method: This paper proposes an NL-driven, high-reliability motion planning framework. Its core innovation is the first integration of Signal Temporal Logic (STL) as a semantic bridge between NL commands and robotic task specifications—enabling semantically interpretable and formally verifiable instruction parsing. We further design an LLM–STL co-architectural pipeline that unifies STL-based formal specification, STL-guided trajectory generation, and low-level motion control. Contribution/Results: Evaluated on two highly challenging experimental scenarios, the framework significantly improves planning stability and task success rate. Compared to conventional NL-prompting approaches, it achieves superior robustness against linguistic ambiguity and provides formal guarantees through STL-based verification.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Safety and Robustness
• 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
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

It has been an ambition of many to control a robot for a complex task using natural language (NL). The rise of large language models (LLMs) makes it closer to coming true. However, an LLM-powered system still suffers from the ambiguity inherent in an NL and the uncertainty brought up by LLMs. This paper proposes a novel LLM-based robot motion planner, named extit{VernaCopter}, with signal temporal logic (STL) specifications serving as a bridge between NL commands and specific task objectives. The rigorous and abstract nature of formal specifications allows the planner to generate high-quality and highly consistent paths to guide the motion control of a robot. Compared to a conventional NL-prompting-based planner, the proposed VernaCopter planner is more stable and reliable due to less ambiguous uncertainty. Its efficacy and advantage have been validated by two small but challenging experimental scenarios, implying its potential in designing NL-driven robots.

Problem

• Resolves ambiguity in natural language commands for robots.
• Reduces uncertainty in LLM-powered robot motion planning.
• Enhances reliability of NL-driven robots using formal specifications.

Innovation

• LLM-based robot motion planner
• Signal temporal logic specifications
• Reduces ambiguity in natural language commands