Abstract:
Promoting the accuracy and coverage of the alighting of passengers in public transport is essential to support route planning and policy decisions aiming to sustainable mobility. Although previous studies place several principles for alighting estimation from incomplete smart card data, most remain dispersed and address one single mode. These gaps hinder a comprehensive comparison of the success rates of existing alighting algorithms. To address the above challenges, this work assesses side-by-side state-of-the-art principles for alight stop inference using smart card data from multimodal transport networks. To our best knowledge, this research is the first incrementally measuring the impact of each principle present in the literature. It further discusses uncertainty factors and proposes a confidence metric on the estimated alighted stops.