Smart Optimization of Dynamic Wireless Charging for Electric Vehicles Using GPS Intelligence and Machine Learning

This study presents a robust, GPS-driven optimization framework for Dynamic Wireless Charging (DWC) in Electric Vehicles (EVs), targeting energy efficiency and intelligent route planning in urban settings. Using a simulated 50point drive between Halifax and Dartmouth, a Random Forest Regressor predicts energy consumption based on variables such as speed, elevation, and distance. To support timely charging decisions, a hybrid cost function selects the optimal station by minimizing a score based on geographic proximity and inverse charging capacity. Compared to a nearest-station baseline, the framework reduced charging time by $42 \%$, detour distance by $46 \%$, and improved remaining battery level by $133 \%$. Although trained on a synthetic dataset, the model captures realistic EV energy patterns. Due to the dataset’s limited size, extensive statistical validation (e.g., cross-validation or real-world generalization) was not performed. However, results aligned well with theoretical expectations. This work demonstrates how lightweight, machine-learning-based approaches can enable practical EV route optimization and sets the stage for future integration with live GPS telemetry, dynamic pricing, and realtime decision-making.