Link-Specific Fuzzy Classification of Transfer Time Based on Passenger Behavior in Urban Rail Systems

Accurate estimation of transfer time is a critical component of passenger-oriented planning and operational efficiency in urban rail systems. Conventional approaches typically rely on fixed or average transfer times and fail to capture variability arising from passenger behavior and station-specific conditions. This study proposes a link-specific fuzzy classification approach for transfer time analysis based on empirical observations. Transfer time distributions between different rail line connections are analysed and found to follow approximately normal distributions with distinct statistical characteristics. Based on these observations, fuzzy membership functions are constructed separately for each transfer pair using mean and standard deviation values. Passengers are then classified as fast, medium, or slow relative to the transfer context. The results demonstrate that passenger speed classification varies significantly across different transfer environments, highlighting the limitations of global and deterministic assumptions. The proposed approach provides a data-driven and flexible framework for representing behavioral variability in transfer processes and offers a practical basis for developing advanced passenger-oriented transfer time models.