From Street to Simulator: Assessing Cyclist Movement Consistency Between Virtual and Real-World Environments

Abstract

Virtual reality (VR) is increasingly used for cycling research, yet the fidelity of these simulations remains under-validated. We introduce a publicly available dataset of synchronized, full-body inertial measurement unit (IMU) data and egocentric video from 11 participants cycling matched 1.4 km routes in Vienna and a VR simulator. Initial analyses indicate that core cycling biomechanics are well-replicated in VR. However, significant differences emerge in upper-body motion and emotional responses. Participants exhibited greater torso rotation in VR and reported lower enjoyment compared to outdoor cycling. The dataset provides a benchmark resource for validating simulator designs and investigating embodied locomotion in VR. To facilitate research and benchmarking, the dataset and analysis tools are shared at: street2simulator.de.