GROKE: Vision-Free Navigation Instruction Evaluation via Graph Reasoning on OpenStreetMap

TL;DR

GROKE introduces a vision-free, graph reasoning-based framework using OpenStreetMap data to evaluate navigation instructions, addressing limitations of visual-based metrics and improving evaluation accuracy and scalability.

Contribution

This paper presents GROKE, a novel hierarchical LLM-based framework that evaluates navigation instructions using OSM data without visual input, outperforming existing methods in accuracy and interpretability.

Findings

Reduces navigation error by 68.5% compared to baselines

Structured spatial representations outperform grid-based and visual graphs

Provides scalable, interpretable evaluation without visual dependencies

Abstract

The evaluation of navigation instructions remains a persistent challenge in Vision-and-Language Navigation (VLN) research. Traditional reference-based metrics such as BLEU and ROUGE fail to capture the functional utility of spatial directives, specifically whether an instruction successfully guides a navigator to the intended destination. Although existing VLN agents could serve as evaluators, their reliance on high-fidelity visual simulators introduces licensing constraints and computational costs, and perception errors further confound linguistic quality assessment. This paper introduces GROKE(Graph-based Reasoning over OSM Knowledge for instruction Evaluation), a vision-free training-free hierarchical LLM-based framework for evaluating navigation instructions using OpenStreetMap data. Through systematic ablation studies, we demonstrate that structured JSON and textual formats for spatial information substantially outperform grid-based and visual graph representations. Our hierarchical architecture combines sub-instruction planning with topological graph navigation, reducing navigation error by 68.5% compared to heuristic and sampling baselines on the Map2Seq dataset. The agent's execution success, trajectory fidelity, and decision patterns serve as proxy metrics for functional navigability given OSM-visible landmarks and topology, establishing a scalable and interpretable evaluation paradigm without visual dependencies. Code and data are available at https://anonymous.4open.science/r/groke.