SmartChargeNet: PCA-Enhanced Attention-LSTM for Wireless Electric Vehicle Charging Optimization

Wireless charging for electric vehicles (EVs) has gained significant popularity due to its contactless nature, eliminating the need for physical connections between the transmitter and receiver. This advancement in transportation contributes to environmental sustainability by reducing greenhouse gas emissions. However, one of the key challenges in wireless power transfer (WPT) systems is the decrease in transmitted power and efficiency as the distance between the coils increases, resulting in power wastage. This paper focuses on the analysis of a WPT system designed for EV charging, taking into account various electrical and environmental factors to optimize the output power. The study explores how changes in the distance between adjacent coils impact the wireless power transfer process. Furthermore, the research involves data collection involving different shapes of transmitter and receiver coils, varying distances, and alignment scenarios. The proposed lightweight system, SmartChargeNet, is a novel system that uses a PCA-Enhanced Attention-LSTM network to predict received power in wireless EV charging systems. The system achieved an impressive R2 Score of 99.29%, demonstrating its effectiveness and accuracy using several metrics including Mean Squared Error (MSE), R2 Score, Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE).