Recent Results on the Use of Artificial Intelligence Techniques Applied to Wireless Power Transfer Systems

Abstract

This article reviews the application of machine learning (ML) techniques in wireless power transfer (WPT) systems, focusing on their role in optimizing system performance, enhancing safety, and improving efficiency. With the growing demand for wireless charging applications such as electric vehicles (EVs), IoT devices, and medical implants, WPT systems face challenges in terms of coil alignment, foreign object detection, and power efficiency. The use of ML algorithms, particularly neural networks and reinforcement learning has emerged as a promising solution to address these challenges. We explore how ML can optimize the geometric and structural design of WPT coils, predict the optimal parameters for inductive couplers, and enhance coupling efficiency under varying conditions. Additionally, ML is being used for foreign object detection (FOD) to ensure safety by identifying metallic and living objects that may interfere with power transmission. The article discusses various approaches, including supervised learning, regression models, and Q-learning algorithms, highlighting their ability to reduce design time, improving system efficiency, and mitigate risks associated with misalignment and object interference. By reviewing recent advancements and ongoing research, this paper provides a comprehensive overview of the potential and limitations of integrating ML into WPT systems, paving the way for smarter, safer, and more efficient wireless charging technologies.

Keywords

• Electric vehicle (EV) charging
• foreign object detection (FOD)
• inductive coupling
• machine learning (ML)
• magnetic field optimization
• misalignment compensation