BikeActions: An Open Platform and Benchmark for Cyclist-Centric VRU Action Recognition

TL;DR

This paper introduces BikeActions, a new multi-modal dataset and benchmark for cyclist-centric VRU action recognition, supported by an open perception platform to advance autonomous driving safety.

Contribution

It presents the first open perception platform and a comprehensive dataset for cyclist action recognition, along with baseline evaluations of state-of-the-art models.

Findings

Established performance baselines for VRU action recognition.

Provided a new multi-modal dataset with 852 annotated samples.

Released open hardware and software tools for future research.

Abstract

Anticipating the intentions of Vulnerable Road Users (VRUs) is a critical challenge for safe autonomous driving (AD) and mobile robotics. While current research predominantly focuses on pedestrian crossing behaviors from a vehicle's perspective, interactions within dense shared spaces remain underexplored. To bridge this gap, we introduce FUSE-Bike, the first fully open perception platform of its kind. Equipped with two LiDARs, a camera, and GNSS, it facilitates high-fidelity, close-range data capture directly from a cyclist's viewpoint. Leveraging this platform, we present BikeActions, a novel multi-modal dataset comprising 852 annotated samples across 5 distinct action classes, specifically tailored to improve VRU behavior modeling. We establish a rigorous benchmark by evaluating state-of-the-art graph convolution and transformer-based models on our publicly released data splits, establishing the first performance baselines for this challenging task. We release the full dataset together with data curation tools, the open hardware design, and the benchmark code to foster future research in VRU action understanding under https://iv.ee.hm.edu/bikeactions/.