A Data-Driven Approach for Micromobility Hazards Detection and Safety Enhancement

Abstract—Micromobility users may face various types of risks and hazards such as road accidents, conflicts with other road users and poor road infrastructure. Effectively, micromobility safety in urban environments represents one of the main challenges and concerns which may affect the adoption and sustainability of this travel mode. The aim of this paper is to propose a data-driven method for the practical implementation of micromobility hazards detection. Therefore, a sensor kit has been developed to be used for collecting micro-level data about micromobility users and providing a common situational awareness of infrastructure and traffic conditions. In the proposed approach, the sensor kit represents the main source of data used to implement machine learning (ML) technologies for detecting several types and severities of micromobility hazards. Therefore, real time data collected by the sensor kit may feed the produced ML classifiers to detect several types and severities of micromobility hazards.