Sim-to-Real Transfer of Vision-Language Navigation in Continuous Environments Using an Ackermann-Steered Mobile Robot

2026-10-05

AI Summary

This study addresses the limitations of conventional vision-language navigation, which relies on topological graphs, panoramic views, and precise localization, thereby hindering adaptation to real-world continuous environments. To overcome this, we propose a Sim-to-Real transfer framework that eliminates these prior dependencies. Methodologically, we design a cross-modal attention architecture to align visual and language representations, achieving efficient domain adaptation through photometric adjustments and fine-tuning with limited real-world data. Furthermore, the system integrates LiDAR-based perception and Ackermann steering control to support offline deployment. Experimental results demonstrate that our approach achieves robust performance in terms of SPL and nDTW metrics, successfully enabling natural language-driven autonomous navigation on dedicated hardware platforms.

Abstract

Vision-Language Navigation (VLN) enables robots to navigate through environments using natural language instructions, making human-robot interaction intuitive. Traditional VLN models often rely on navigation graphs, 360-degree views, and perfect localization which pose significant challenges when adapting these models to real-world settings. This work addresses these limitations by performing a simulation-to-real domain shift of a VLN approach that operates in continuous environments without requiring navigation graphs or panoramic views. The proposed system integrates vision-language models that align visual inputs and linguistic instructions within a shared embedding space, facilitating natural language-driven navigation. We employ a Cross-Modal Attention (CMA) based architecture trained on an existing dataset in a simulated environment and fine-tune it using real-world data collected from a custom-built Ackermann-steered robot equipped with a camera and a LiDAR sensor. By utilising linear photometric adjustments and fine-tuning on a limited number of episodes, our model successfully adapts to real-world environments, achieving effective navigation while running offline on dedicated hardware. Experimental results, evaluated using Success weighted by Path Length (SPL) and Normalized Dynamic Time Warping (nDTW) metrics, demonstrate the robustness and adaptability of our approach. Keywords: Vision-Language Navigation, Cross-Modal Attention, Natural Language Instructions, Sim-to-Real Transfer, Autonomous Navigation, Ackermann-steering.

Problem

• Vision-Language Navigation
• Sim-to-Real Transfer
• Continuous Environments
• Ackermann-steered Robot

Innovation

• Vision-Language Navigation
• Sim-to-Real Transfer
• Cross-Modal Attention
• Continuous Environments
• Ackermann-steered Robot