Urban Cycling: Intelligent Bicycle Sensors for Road Safety and Sustainability

Urban transportation is rapidly evolving with a growing emphasis on sustainability and smart city development. Cycling, a cornerstone of sustainable urban mobility, requires reliable data for effective integration into city planning. However, there remains a critical gap in the comprehensive collection and analysis of cycling-specific environmental and safety data. Addressing this gap, this work introduces an innovative bicycle sensor system that utilises embedded artificial intelligence (AI) to process data on-device. By reducing data overhead and enhancing privacy, this system offers a novel approach to improving cycling safety and contributing to smarter urban transportation planning. Our sensor system integrates road surface classification and overtaking detection, employing edge computing to minimise energy consumption and data transmission, while still delivering actionable insights for city planners and cyclists alike. A workshop conducted with citizens, followed by a survey, evaluated user experiences, revealing valuable insights into usability, connectivity, and areas for improvement, including connection stability and data continuity. Although our AI models showed promising results, they were constrained by limited training data and sampling rates, highlighting areas for future refinement. This work shows the potential of integrating EmbeddedAI within cycling infrastructure planning to enhance road safety, and suchmobile sensor systems to engage communities in urban mobility efforts.