In this Innovation Lab guest post, our perception partner Discernity AI will showcase their cutting-edge application: radar-based blink monitoring. By leveraging the millimeter accuracy of Acconeer’s Pulsed Coherent Radar (PCR) technology, Discernity AI has developed a non-intrusive, privacy-preserving solution to monitor eyelid dynamics in real time, enabling new potentials for human condition sensing for next generation automotive driver sensing systems.
The use-case
Blink dynamics are critical metrics and essential for judging a person’s drowsiness and cognitive load. Such information can also be used to enable new accessible and advanced human-computer interaction, for users with physical limitations or in GDPR-sensitive environments.
Traditional camera-based systems often struggle with visual obstructions, dynamic lighting, or privacy concerns. Discernity AI solves these challenges by utilizing Acconeer’s latest radar platform, offering continuous, robust, and privacy-safe monitoring regardless of ambient lighting or eyewear.
To implement this solution, Discernity AI utilizes Acconeer’s PCR sensor with small footprint, low power consumption and penetration capability, ideal for integration into larger product without draining battery life, requiring bulky hardware, or demanding complex product design.
Algorithm and how it works
Detecting an eye blink requires tracking incredibly fast and subtle physical variations. Discernity AI’s proprietary smart signal-processing algorithm and Edge AI model extract the eyelid dynamics from surrounding environmental noise and natural head movements and further understand the person’s conditions. Because the system relies entirely on radar signals rather than optical lenses, it ensures total user privacy.
Get in touch
The projects featured in the Innovation Lab are designed to inspire and showcase the endless possibilities of Acconeer’s radar technology.
If you are interested in implementing this blink detection application into your own products, or want to learn more about the algorithm and system capabilities, please reach out directly to the team at Discernity AI or visit https://www.discernity.ai/.