LOG-Nav: Efficient Layout-Aware Object-Goal Navigation with Hierarchical Planning

TL;DR

LOG-Nav is a hierarchical, layout-aware navigation method that uses a global map and local scene memory, managed by an LLM agent, to achieve high success rates in complex indoor environments without extensive training.

Contribution

This paper presents LOG-Nav, a novel hierarchical navigation approach leveraging layout information and LLMs, enabling efficient object-goal navigation in multi-room indoor spaces.

Findings

Achieves 85% success rate on MP3D benchmark

Improves success rate by over 40% compared to existing methods

Demonstrates robustness in both virtual and real-world robotic deployments

Abstract

We introduce LOG-Nav, an efficient layout-aware object-goal navigation approach designed for complex multi-room indoor environments. By planning hierarchically leveraging a global topologigal map with layout information and local imperative approach with detailed scene representation memory, LOG-Nav achieves both efficient and effective navigation. The process is managed by an LLM-powered agent, ensuring seamless effective planning and navigation, without the need for human interaction, complex rewards, or costly training. Our experimental results on the MP3D benchmark achieves 85\% object navigation success rate (SR) and 79\% success rate weighted by path length (SPL) (over 40\% point improvement in SR and 60\% improvement in SPL compared to exsisting methods). Furthermore, we validate the robustness of our approach through virtual agent and real-world robotic deployment, showcasing its capability in practical scenarios.