The increasing adoption of electric vehicles (EVs) has placed significant pressure on existing charging infrastructure. Public charging stations often struggle with issues such as overcrowding, extended waiting times, and inefficient energy distribution. To address these challenges, this study introduces an AI-powered demand prediction framework capable of forecasting when, where, and how much charging activity will occur. By applying Long Short-Term Memory (LSTM) networks alongside regression models, the system evaluates historical usage data, traffic conditions, and weather influences to generate accurate forecasts. Beyond prediction, the framework categorizes congestion levels into low, medium, and high, offering practical insights for both users and operators. EV drivers receive personalized recommendations such as the most suitable charging station, estimated waiting time, cost projections, and optimal charging periods. Meanwhile, infrastructure managers and urban planners gain valuable support in balancing utilization and planning future expansions.
Sujini et al. (Tue,) studied this question.