LEDs Power Indoor IoT Devices Wirelessly

In the world of distributed sensors and other IoT devices, one of the biggest pain points is keeping these little gadgets powered up. A battery can only last so long, and manually recharging dozens, hundreds, or thousands of devices that are geographically dispersed is impractical. So for this technology to continue to advance, new and better methods of delivering power will be needed.

The dream, of course, is wireless power delivery. But while some rudimentary wireless power delivery systems do exist, they are of little use beyond very short distances, and they also tend to be highly inefficient. However, when it comes to low-power devices — like those powered by energy-efficient microcontrollers that spend most of their time in a sleep state — existing solar panels can supply enough juice.

Or at least they can when they are outdoors and the sun is shining. For indoor applications, small solar cells are not likely to produce enough power under typical ambient lighting conditions. But that may change in the near future thanks to the work of a pair of researchers at the Institute of Science Tokyo. They have developed an approach that uses LEDs to actively power remote devices wirelessly, both day and night.

Their system is an adaptive dual-mode LED-based optical wireless power transmission (OWPT) platform. It was designed to overcome the major limitations of traditional far-field wireless power methods. Unlike laser-based OWPT — which can deliver significant power but poses safety risks in human-occupied spaces — LEDs offer a safe, low-cost alternative suitable for everyday indoor environments. The challenge, however, has always been that LED beams tend to spread out, making long-distance power transfer inefficient, especially when ambient lighting changes.

The team tackled these issues with a clever combination of adaptive optics and AI-driven beam control. At the heart of their transmitter is a double-layer lens system that includes a tunable liquid lens. By adjusting the focal length on the fly, the system can maintain a tight, energy-dense spot of light at distances of up to five meters. This is a significant improvement over past LED-based designs.

A dual-axis motorized reflector then steers this beam with high precision. Using an RGB-IR depth camera, the system identifies both the photovoltaic receivers and the exact location where the beam is landing. An onboard neural network refines that detection, allowing the system to lock onto receivers of different sizes, even as lighting conditions change.

One particularly clever element is the use of retroreflective material surrounding each receiver. This ensures accurate tracking in the dark, enabling normal operation whether the room lights are on or off.

In tests, the system was able to hop between multiple receivers quickly and maintain stable power delivery without interruption. If scaled and commercialized, this approach could form the backbone of future indoor IoT infrastructure — one where sensors, appliances, and other tiny devices remain powered indefinitely without cables or battery swaps.