Accelerating Electric Vehicle (EV) adoption: A remote sensing data driven and deep learning-based approach for planning public car charging infrastructure

Abstract

Car fleet electrification is critical for achieving ambitious climate action goals. Access to charging stations is a major barrier for widespread adoption of EV, especially impacting members of lower socio-economic groups who cannot easily install home chargers in their residences. This research aims to examine the demand for public EV charging stations in residential areas and their geographical distribution. By utilizing advanced deep learning models and high-resolution remote sensing imagery, the study aims to identify specific clusters of households that require access to the public infrastructure. The study uses high-resolution aerial images and property parcels as input to a deep learning model YOLOv8 to recognize properties that may require access to public charging stations. This study presents an innovative approach that addresses challenges pertaining to EV adoption using remote sensing data, machine learning, and geospatial analysis. The results of the study demonstrate spatial analysis using sociodemographic data and household parking data, generated through the innovative method developed in this work, to aid Irish towns in planning public EV charging facilities among residential neighbourhoods. The study's findings are expected to aid in the implementation of expansion strategies for the public EV charging network, which is vital for meeting ambitious EV fleet targets.