Review maps AI-powered metasurfaces for communication and cloaking

A new review in Opto-Electronic Technology examines how AI is turning programmable electromagnetic metasurfaces into adaptive systems for wireless communication and invisibility applications. The paper, published July 2, 2026, highlights design, control and deployment challenges as the field moves toward real-time, intelligent EM platforms.

Why it matters: - AI-enabled programmable metasurfaces could reshape wireless networks by actively steering signals, improving coverage and supporting integrated sensing and communication. - The same technology could make electromagnetic cloaking more adaptive by changing scattering patterns and Doppler signatures in real time. - The review says these systems are moving from simple wave-control surfaces toward intelligent electromagnetic platforms that can respond to changing environments.

What happened: - Opto-Electronic Technology published an invited review on July 2, 2026, with DOI 10.29026/oet.2026.260014. - The article is titled “AI-enabled electromagnetic metasurfaces for wireless communication and invisibility cloak.” - The review was contributed by a team led by Prof. Hongsheng Chen, Prof. Bin Zheng and Researcher Jiwei Zhao at Zhejiang University. - The paper focuses on programmable electromagnetic metasurfaces, AI-assisted design, closed-loop control and applications in wireless communication and electromagnetic invisibility.

The details: - Programmable metasurfaces are ultrathin electromagnetic interfaces built from subwavelength elements that act like controllable EM pixels. - These surfaces can reshape wave direction, energy distribution, polarization and frequency content. - Spatial coding assigns different electromagnetic responses across a surface to redirect or reshape wavefronts. - Temporal modulation changes a surface response over time and can move energy into new frequency components. - Space-time coding combines spatial and temporal control to coordinate beam direction and frequency channels on one platform. - AI-assisted inverse design can speed up the search for meta-atom structures that produce target electromagnetic responses. - Data-driven methods can help synthesize large metasurface arrays and manage interactions among neighboring elements. - Closed-loop systems can combine electromagnetic, visual or deformation sensing to update control strategies in real time. - In wireless communication, intelligent metasurfaces can reshape propagation environments, extend signal coverage, track moving users and directly modulate information onto electromagnetic waves. - In electromagnetic invisibility, programmable metasurfaces can reconstruct scattering fields for adaptive cloaking under changing conditions. - Temporal modulation can also regulate radar-detected Doppler signatures, extending cloaking beyond spatial scattering control. - The review says progress depends on physically reliable algorithms, efficient hardware, experimental validation and robust operation in realistic environments.

Between the lines: - The review frames AI as the missing layer that lets metasurfaces do more than execute preset commands. - That shift matters because real environments change constantly, especially in wireless and radar settings where static designs are limited. - The biggest gap is no longer the concept, but the ability to run low-power, large-scale, real-time systems outside the lab.

What's next: - The field still needs low-power hardware, real-time control, large-scale integration and reliable validation in complex environments. - Further progress will likely come from tighter links between programmable devices, intelligent algorithms and sensing-feedback loops. - The review expects future metasurfaces to evolve into smart EM platforms that can perceive, decide and adapt on their own. - The Zhejiang University team says its broader research spans artificial EM materials, metasurfaces, EM invisibility, intelligent EM control and intelligent wireless communication. - The team reports more than 100 papers in journals including Nature Photonics, Nature Communications and Science Advances, along with more than 40 invention patents. - More information is available in the journal’s official website, editorial board page, online archive and submission portal.