Constrained Natural Language Action Planning for Resilient Embodied Systems

2025-10-07

AI Summary

To address the low reliability, poor reproducibility, and opaque decision-making of LLM-driven robotic task planning in real-world environments, this paper proposes a constraint-driven framework that synergistically integrates large language models (LLMs) with symbolic planning. Symbolic planning explicitly encodes hard constraints—such as physical feasibility and action temporal ordering—while the LLM handles high-level semantic understanding and generalizable reasoning. A bidirectional supervision mechanism dynamically couples the two components, ensuring robustness and interpretability without sacrificing the LLM’s strong generalization capability in open-world settings. Experiments demonstrate that our approach achieves 99% success rate on the ALFWorld benchmark and 100% task completion rate on a real quadrupedal robot—significantly outperforming both pure-LLM and pure-symbolic baselines. The framework thus bridges the gap between neural scalability and symbolic rigor in embodied AI planning.

Technology Category

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

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

Replicating human-level intelligence in the execution of embodied tasks remains challenging due to the unconstrained nature of real-world environments. Novel use of large language models (LLMs) for task planning seeks to address the previously intractable state/action space of complex planning tasks, but hallucinations limit their reliability, and thus, viability beyond a research context. Additionally, the prompt engineering required to achieve adequate system performance lacks transparency, and thus, repeatability. In contrast to LLM planning, symbolic planning methods offer strong reliability and repeatability guarantees, but struggle to scale to the complexity and ambiguity of real-world tasks. We introduce a new robotic planning method that augments LLM planners with symbolic planning oversight to improve reliability and repeatability, and provide a transparent approach to defining hard constraints with considerably stronger clarity than traditional prompt engineering. Importantly, these augmentations preserve the reasoning capabilities of LLMs and retain impressive generalization in open-world environments. We demonstrate our approach in simulated and real-world environments. On the ALFWorld planning benchmark, our approach outperforms current state-of-the-art methods, achieving a near-perfect 99% success rate. Deployment of our method to a real-world quadruped robot resulted in 100% task success compared to 50% and 30% for pure LLM and symbolic planners across embodied pick and place tasks. Our approach presents an effective strategy to enhance the reliability, repeatability and transparency of LLM-based robot planners while retaining their key strengths: flexibility and generalizability to complex real-world environments. We hope that this work will contribute to the broad goal of building resilient embodied intelligent systems.

Problem

• Enhancing reliability of LLM planners with symbolic oversight
• Improving repeatability through transparent constraint definition
• Maintaining LLM generalization while ensuring planning success

Innovation

• LLM planners augmented with symbolic planning oversight
• Transparent hard constraints replace traditional prompt engineering
• Preserves LLM reasoning while ensuring reliability and repeatability