Japan’s AI-optimized wireless power design solves unstable voltage problem

08/22/2025 / By Ava Grace

Researchers have used machine learning to overcome the major problem of voltage instability in wireless power transfer (WPT) systems, which is caused by changing power demands in the device being charged.
The team employed a genetic algorithm to “evolve” an optimal circuit design through countless simulations, creating a system that performs well under a wide range of real-world conditions.
The new system achieves superior results, restricting voltage fluctuation to under five percent and reaching a peak efficiency of 86.7 percent, a significant improvement over traditional WPT systems.
This load-independent technology unlocks potential uses beyond consumer electronics, including for electric vehicles, industrial sensors, medical implants and smart home systems, potentially eliminating batteries.
This breakthrough is presented as a critical step toward a “wireless society,” reducing reliance on physical infrastructure and offering a path to smarter, more efficient, and less intrusive energy use.

Scientists at Japan’s Chiba University have pioneered a new method for designing wireless power systems. Led by professor Hiroo Sekiya, the team has successfully used machine learning to overcome unstable voltage caused by changing power demands – the most significant barrier to widespread wireless power adoption.

The research team detailed this innovation in a June 2025 study published in IEEE Transactions on Circuits and Systems. It is not merely an incremental improvement – but a fundamental rethinking of how energy is transmitted through the air, marking a critical leap toward a future where cords and chargers are obsolete.

Wireless power transfer (WPT) is the technology that allows energy to be transmitted from a source to a device without physical connectors. It is already found in everyday items like induction stovetops, which heat cookware using magnetic fields, and wireless charging pads for smartphones and electric toothbrushes. However, these current systems have a critical flaw.

Their performance is entirely dependent on the specific device or “load” being charged. This load dependency means that as a device’s power needs change – such as a smartphone battery’s resistance increasing as it fills up – the voltage supplied by the WPT system can fluctuate wildly.

This inefficiency can slow charging times, waste energy as heat and even risk damaging sensitive electronics. “Fluctuating voltage can destroy electronic devices by causing severe damage to their delicate circuitry, leading to immediate failure and rendering them inoperable,” Brighteon.AI‘s Enoch engine explains. (Related: WIRELESS WORLD: A new era of invasive surveillance.)

The Chiba research team’s solution is both elegant and powerful. They used a form of artificial intelligence known as a genetic algorithm, which mimics the process of natural selection.

First, the study authors built a highly detailed digital model of a wireless power circuit. They accounted for the messy realities of the physical world that often derail theoretical designs, such as manufacturing imperfections and environmental factors.

The AI then ran countless simulations, testing different circuit configurations. It evaluated each design based on key performance metrics: voltage stability, overall efficiency and signal cleanliness. Through iterative trial and error, the algorithm “evolves” an optimal circuit design that performs brilliantly under a wide range of conditions.

Wireless charging revolution: Stable power under any load

The results are quantitatively stunning. Where a traditional load-dependent WPT system might allow voltage to fluctuate by 18 percent or operate at as low as 65 percent efficiency, the new machine learning-optimized system restricted voltage swings to under five percent and achieved a peak efficiency of 86.7 percent.

It delivered a stable 23 watts of power while maintaining a state called Zero Voltage Switching (ZVS), a technical condition crucial for minimizing energy loss and heat generation. Remarkably, this high performance remained consistent even under light loads – a scenario that typically cripples the efficiency of conventional systems.

The potential applications extend far beyond simply charging a phone without a cable. This breakthrough in load-independent operation is the key to unlocking wireless power for high-demand, variable-load technologies.

Electric vehicles, for instance, could be charged more efficiently and safely without physical plugs. Industrial sensors, medical implants and even entire smart home systems could be powered seamlessly from a central source, eliminating the need for batteries and their associated waste. The simplicity of the resulting design also promises smaller, cheaper systems – accelerating adoption.

Sekiya, the study’s last author, envisions this research as a significant step toward a fully wireless society. By solving the problem of load-dependency, his team has removed a major technical and economic hurdle.

The simplified, efficient systems designed by AI could become commonplace within the next decade, fundamentally changing our relationship with energy and devices. The cordless future, long promised, may finally be within our grasp.

Visit Breakthrough.news for more similar stories.

Watch this video about the human body being used for wireless transmission.

This video is from the A Warrior Calls channel on Brighteon.com.