Curbside EV charging infrastructure planning: fusing choice behavior and demand prediction

As the production of electric vehicles (EVs) continues to increase, urban charging infrastructure has not kept pace with the rising demand, leading to a significant gap in the availability of charging stations. Curbside charging offers a promising solution but suffers from inefficient resource allocation. A comprehensive framework for curbside charging planning is proposed. Spatiotemporal analysis of parking transaction data revealed that EVs exhibit higher short-term parking demand and stronger spatial clustering than internal combustion engine vehicles (ICEVs). To understand the underlying mechanisms, a binary logit model was used to investigate EV users' curbside charging choices, identifying weather, state of charge, subsequent travel distance, and pricing as significant determinants. Simulated vehicle arrivals in different functional zones using the Monte Carlo method showed an average error of 12.03% compared with actual observations, indicating the method’s effectiveness in forecasting the spatial distribution of charging demand. The proposed framework achieved reliable demand estimations and provides practical insights for optimizing charging infrastructure deployment, supporting integrated urban transport and energy system management.