User demand prediction and cloud-based smart mobile interface for electric vehicle charging

With recent years' fast growing penetration of Plug-in Electric Vehicles (PEV), the energy source under demand is gradually transferring from gasoline to electricity. It would be of significant interest to the utility and energy service community to predict charging demand on users' side. In this paper, we find and propose a practical PEV user demand prediction algorithm through user-based charging session data. The analysis is completed with the data collected from the charging stations installed on campus of University of California, Los Angeles (UCLA). We also present a mobile interface for PEV users to interact with the charging system. Initial investigations recommend to use the median value of each individual user's charging history as a personal energy demand value. The SMAPE accuracy of this approach is 0.55 for users with small and large sample sizes.