Researchers have introduced a clever new method to boost wireless power transfer (WPT) systems using machine learning. By merging differential equations with genetic algorithms, this approach maintains a steady voltage output and ensures zero-voltage switching (ZVS) even as loads change. If you’ve ever battled with fluctuating voltages, you’ll find this technique especially reassuring.
WPT technology isn’t new—its roots go back to Nikola Tesla’s experiments in the 1890s—but modern devices, from smartphones to IoT sensors, still rely on these principles. Traditionally, achieving load-independent (LI) operation depended on precise circuit components calculated through complex equations that often struggle to capture real-world dynamics. Now, a team led by Prof. Hiroo Sekiya at Chiba University, alongside colleagues from Tokyo University of Science and Sojo University, is changing that with a machine learning-based design.
The process involves modelling the WPT circuitry with differential equations that accurately reflect actual component behaviours. These equations are then solved numerically, while genetic algorithms iteratively fine-tune system parameters to deliver a stable voltage, optimise efficiency, and minimise harmonic distortion. In a practical test on a class-EF WPT system—which pairs a class-EF inverter with a class-D rectifier—the design kept voltage fluctuations to under 5%, a significant improvement over the 18% variations typically seen in conventional systems.
The researchers even factored in aspects like diode parasitic capacitance, enhancing performance at lighter loads. Power-loss analysis revealed consistent energy dissipation and output currents, with the system achieving an 86.7% power-delivery efficiency at 6.78 MHz and delivering over 23 watts. This careful balancing act not only helps in maintaining stability but also paves the way for more compact, cost-effective designs.
Prof. Sekiya remarked, “We established a novel design procedure for a LI-WPT system that achieves a constant output voltage without control against load variations.” He believes these findings could be a key step toward making wireless power truly ubiquitous—potentially within a decade. For anyone interested in how smart engineering can simplify everyday challenges, this work offers valuable insights into the future of automated circuit design.
