Open-Architecture End-to-End System for Real-World Autonomous Robot Navigation

TL;DR

This paper introduces a lightweight, open-architecture system enabling real-world autonomous robot navigation using onboard sensors, semantic mapping, and LLM-based planning, demonstrated on a quadruped robot with high success rates.

Contribution

The work presents a novel end-to-end system integrating semantic mapping and LLM planning for real-world robot navigation, moving beyond simulation limitations.

Findings

Achieved over 88% success rate in real-world indoor navigation tasks.

Demonstrated zero-shot navigation capabilities in complex environments.

Integrated semantic scene graphs with LLM planning for adaptive navigation.

Abstract

Enabling robots to autonomously navigate unknown, complex, and dynamic real-world environments presents several challenges, including imperfect perception, partial observability, localization uncertainty, and safety constraints. Current approaches are typically limited to simulations, where such challenges are not present. In this work, we present a lightweight, open-architecture, end-to-end system for real-world robot autonomous navigation. Specifically, we deploy a real-time navigation system on a quadruped robot by integrating multiple onboard components that communicate via ROS2. Given navigation tasks specified in natural language, the system fuses onboard sensory data for localization and mapping with open-vocabulary semantics to build hierarchical scene graphs from a continuously updated semantic object map. An LLM-based planner leverages these graphs to generate and adapt multi-step plans in real time as the scene evolves. Through experiments across multiple indoor environments using a Unitree Go2 quadruped, we demonstrate zero-shot real-world autonomous navigation, achieving over 88% task success, and provide analysis of system behavior during deployment.