SmallPlan: Leverage Small Language Models for Sequential Path Planning with Simulation-Powered, LLM-Guided Distillation

TL;DR

SmallPlan introduces a framework where small language models are trained using large models as teachers to perform efficient, real-time path planning in robotics, suitable for edge devices, by combining simulation, supervised fine-tuning, and reinforcement learning.

Contribution

The paper presents a novel method for training lightweight language models for path planning using LLM-guided distillation, enabling efficient deployment in robotics.

Findings

SLMs achieve competitive performance with GPT-4o on path planning tasks.

SLMs are resource-efficient and suitable for edge deployment.

The approach reduces hallucination and overfitting in small models.

Abstract

Efficient path planning in robotics, particularly within large-scale, complex environments, remains a significant hurdle. While Large Language Models (LLMs) offer strong reasoning capabilities, their high computational cost and limited adaptability hinder real-time deployment on edge devices. We present SmallPlan - a novel framework leveraging LLMs as teacher models to train lightweight Small Language Models (SLMs) for high-level path planning tasks. In SmallPlan, the SLMs provide optimal action sequences to navigate across scene graphs that compactly represent full-scaled 3D scenes. The SLMs are trained in a simulation-powered, interleaved manner with LLM-guided supervised fine-tuning (SFT) and reinforcement learning (RL). This strategy not only enables SLMs to successfully complete navigation tasks but also makes them aware of important factors like distance travel, providing more efficient path planning. Through experiments, we demonstrate that the fine-tuned SLMs perform competitively with larger models like GPT-4o on sequential path planning, without suffering from hallucination and overfitting. SmallPlan is resource-efficient, making it well-suited for edge-device deployment and advancing practical autonomous robotics. Our source code is available here: https://github.com/quangpham2006/SmallPlan