AI-Powered 'Smart Surfaces' Could Banish Wi-Fi Dead Zones for Good
Imagine never losing your signal at home again — this makes it possible.
Researchers demonstrated a machine learning-based system that lets Reconfigurable Intelligent Surfaces (RISs) generate new phase masks in real time, instead of relying on pre-stored codebooks. Current RIS setups need computationally heavy offline algorithms to create phase masks, then store them on a microcontroller—but limited storage restricts how well the surface can adapt. Their approach puts a compact ML model directly on the microcontroller, approximating the phase mask generation algorithm and responding to new inputs while taking up less space than a codebook.
- Smart surfaces currently rely on a fixed set of pre-programmed signal patterns, which limits how well they adapt to real-world conditions.
- The new AI runs directly on the surface's tiny chip, generating fresh reflection patterns in real time — no big storage needed.
- This could make 5G and future wireless networks faster and more consistent, especially indoors and in crowded spaces.
Why It Matters
Faster, more reliable wireless in homes and cities means fewer dropped calls, quicker downloads, and smoother video.