From Steering to Pedalling: Do Autonomous Driving VLMs Generalize to Cyclist-Assistive Spatial Perception and Planning?

TL;DR

This paper introduces CyclingVQA, a benchmark for evaluating vision-language models on cyclist-centric perception and reasoning, revealing strengths and gaps in current models for cyclist-assistive traffic understanding.

Contribution

The paper presents CyclingVQA, a new diagnostic benchmark for cyclist-centric perception and reasoning, and evaluates 31+ VLMs to identify strengths and limitations in cyclist-assistive applications.

Findings

Current models show promising capabilities but need improvement in cyclist-specific cues.

Several models underperform in cyclist-assistive scenarios compared to general-purpose models.

Error analysis highlights key failure modes to guide future development.

Abstract

Cyclists often encounter safety-critical situations in urban traffic, highlighting the need for assistive systems that support safe and informed decision-making. Recently, vision-language models (VLMs) have demonstrated strong performance on autonomous driving benchmarks, suggesting their potential for general traffic understanding and navigation-related reasoning. However, existing evaluations are predominantly vehicle-centric and fail to assess perception and reasoning from a cyclist-centric viewpoint. To address this gap, we introduce CyclingVQA, a diagnostic benchmark designed to probe perception, spatio-temporal understanding, and traffic-rule-to-lane reasoning from a cyclist's perspective. Evaluating 31+ recent VLMs spanning general-purpose, spatially enhanced, and autonomous-driving-specialized models, we find that current models demonstrate encouraging capabilities, while also revealing clear areas for improvement in cyclist-centric perception and reasoning, particularly in interpreting cyclist-specific traffic cues and associating signs with the correct navigational lanes. Notably, several driving-specialized models underperform strong generalist VLMs, indicating limited transfer from vehicle-centric training to cyclist-assistive scenarios. Finally, through systematic error analysis, we identify recurring failure modes to guide the development of more effective cyclist-assistive intelligent systems.