A Real-Time Data-Driven Cycling Routing Approach: A Case Study in Singapore

Since cyclists must consider dynamic criteria such as weather and traffic safety when selecting routes, incorporating real-time road conditions into route planning is essential. However, few existing studies have addressed real-time cycling route planning. Therefore, this study proposed a real-time data-driven cycling routing approach. The approach integrated multiple criteria concerning cyclists' health, safety, and comfort, while accounting for their dynamic variations. To respond to these real-time changes in criteria, two Dijkstra-based dynamic routing models were introduced, distinguished by their mechanisms for triggering path recalculation: the Periodic Update Model (PUM) and the Node-Based Update Model (NUM). The proposed method was validated through simulation experiments of cyclist travel behavior in a case study conducted in Singapore. The experimental results, measured by the Optimal Route Provision Rate (ORPR), confirmed that both proposed dynamic routing models consistently outperformed the traditional shortest-path model across all evaluated scenarios. Nuanced differences in the models' performance were also revealed: while NUM proved superior in route quality for short-distance scenarios, PUM demonstrated robust scalability for long-distance scenarios. The conclusions affirm the effectiveness of the proposed method in supporting cyclists' route choice decisions and highlight the potential for advancing more systematic bicycle route planning.