Researchers from the Beijing National Laboratory for Molecular Sciences (BNLMS), CPU Hydrogen Power Technology (Suzhou), Fudan University, Chinese Academy of Sciences and Zhengzhou University have developed smart glasses that track eye position using arrays of perovskite light sensors instead of cameras or contact lenses. Their system measures light reflected from the eyeball to determine gaze direction with five-degree precision.
The researchers solved a key materials challenge by developing a novel crystal growth method inspired by biological mineralization processes. They added a layer of polyacrylic acid sodium (PAAS) that guides perovskite crystals to form in larger, more organized structures – similar to how sea creatures control shell formation. This resulted in methylammonium lead iodide films with superior light-detecting capabilities.
The improved perovskite sensors demonstrated remarkable sensitivity, responding to light changes 300 times more strongly under typical indoor lighting compared to darkness. They generated 22.09 amperes of current per watt of incident light power. The devices maintained 91% of their initial performance after extended testing in normal humidity.
The smart glasses contain a grid of these sensors positioned to detect light reflected from different regions of the eye. Neural network algorithms process the sensor signals in real-time to determine eye position. Testing showed 99.86% accuracy in distinguishing nine different gaze directions. The system maintained this precision even as the distance between sensors and eye varied from 14 to 56 millimeters.
To demonstrate practical applications, the researchers connected their smart glasses to a remote-controlled car. Users successfully navigated the vehicle through complex paths using only eye movements. The system responded within 130 milliseconds – fast enough for smooth control. Performance remained stable across varying light conditions typical of indoor environments.
The technology addresses several limitations of existing eye tracking approaches. It avoids the computational overhead and privacy concerns of camera-based systems while providing similar accuracy. The non-contact sensors eliminate the discomfort of contact lens devices. Solution-based manufacturing of the perovskite materials suggests potential for cost-effective production.
The system's speed and reliability make it suitable for real-world applications requiring precise eye tracking. Medical professionals could monitor eye movement patterns to diagnose neurological conditions. People with limited mobility could control assistive devices through eye gestures. Augmented reality systems could provide more natural hands-free interfaces.
This research demonstrates that perovskite light sensors can enable accurate, non-invasive eye tracking in a practical form factor. The combination of precise movement detection, fast response time, and stable operation in varying conditions addresses key requirements for widespread adoption of eye-controlled interfaces. The documented performance metrics and successful real-world testing establish a foundation for further development of this approach.
