Wireless charging may soon become even more efficient and reliable, thanks to a pioneering machine learning approach developed by researchers at Chiba University. Their innovation tackles one of the biggest challenges in wireless power transfer (WPT) maintaining stable performance despite load changes.
From smartphones and electric toothbrushes to IoT sensors, WPT systems have quietly entered our daily lives. These systems transmit electrical energy without physical connectors, relying on electromagnetic fields, a concept that traces back to Nikola Tesla’s early experiments. However, maintaining stable performance has always required precise tuning of inductors and capacitors. In reality, environmental factors, manufacturing tolerances, and parasitic capacitance often cause voltage instability and disrupt zero voltage switching (ZVS), a key factor for efficiency.
To overcome this, Professor Hiroo Sekiya and his team introduced a machine-learning-based WPT design method that factors in real-world imperfections. Their approach models the WPT circuit through differential equations that track voltage and current behavior, solving them numerically until steady-state operation is reached. Performance is then evaluated based on voltage stability, efficiency, and harmonic distortion.
A genetic algorithm, an optimization technique inspired by natural selection fine-tunes the circuit parameters. This cycle repeats until LI operation requirements are met. The method was tested on a class-EF WPT system combining a class-EF inverter and class-D rectifier. Traditional setups can only maintain ZVS at a single operating point, but the new design restricted voltage fluctuation to under 5%, compared to 18% in conventional systems.
Performance gains were significant: the LI system consistently maintained ZVS and delivered 23 watts at 86.7% efficiency at 6.78 MHz, even under varying loads. Loss analysis showed transmission coil efficiency remained stable, indicating steady output current.
Looking ahead, Prof. Sekiya sees wide-reaching implications: “We are confident that the results of this research are a significant step toward a fully wireless society.”
He also highlighted the potential for simpler, cheaper, and smaller WPT systems: “Our goal is to make WPT commonplace within the next 5 to 10 years.”
The study was published in IEEE Transactions on Circuits and Systems I: Regular Papers.
 maintaining stable performance despite load changes. From smartphones and electric toothbrushes to IoT sensors, WPT systems have quietly entered […]&media=https://wonderfulengineering.com/wp-content/uploads/2025/08/image-140.png){kind=link}