Naturalistic micromobility data: opportunities and threats

Other conference contribution, 2024

Micromobility data are increasingly available and promise to support not only researchers but also policymakers and authorities in best integrating micromobility into the transport system. When micromobility data are collected in the wild by road users attending to their daily routines, these data bring unprecedented insights into the behavior of micromobility road users and their interaction with other road users and infrastructure. Although GPS data is the most widely available micromobility data, electrified vehicles (e.g., e-bikes and e-scooters) often include sophisticated sensors such as inertial measurement units and cameras. New advances in technology make it possible to analyze video data with artificial eyes and leverage artificial intelligence to model and analyze vehicle dynamics and user behavior, complementing GPS with information crucial for understanding micromobility safety, efficiency, and acceptance.

Within the MicroVision and e-SAFER projects, naturalistic data from e-scooter rental services have been used to investigate crash causation and model rider behavior to improve advanced driving assistance systems, support automated driving functions, and inform Euro NCAP protocols. These projects unveiled the peculiarities and prevalence of leisure riding on e-scooters and created some of the basis for sharing micromobility data and open behavioral models. If shared, the data and models from e-SAFER may serve new analyses and promote new countermeasures based on education, policymaking, and infrastructure design. However, data sharing is a double-edged sword where issues such as ethics, privacy, and security need to find a compromise with commercial interests, while the integrity of the data and analysis results must be warranted.