Answers are grounded in this paper and cite their sources.
The vision for future wireless networks extends far beyond mere connectivity, fostering spaces that autonomously perceive and adapt to the radio environment. Reconfigurable metasurfaces (RMSs) are designed to tailor wireless propagation through programmable electromagnetic responses. However, conventional RMSs, while assisting communication, lack integrated sensing and self-adaptive control capabilities and suffer from severe path loss along the transmitter-RMS-receiver link. Here we report a self-adaptive metasurface (SA-MS) that integrates sensing and active wave control within a unified platform. Each meta-atom integrates dual-biased PIN diodes and an embedded port, enabling 1-bit programmable reflection and in-band sensing. The sensed waveform is fed to a lightweight convolutional neural network that infers the direction of arrival with robustness against array mismatches. The acquired spatial information is then mapped through a physics-informed surrogate model to an amplitude-phase configuration that works in conjunction with the amplifier to enhance the wireless link. Real-world wireless communication experiments in dynamic environments validate the sensing, decision, and execution loop of SA-MS, with an average signal enhancement of 14.6 dB. This work paves the way for intelligent wireless spaces, where sensing and communication co-evolve at the physical layer.
No takes yet. Share an insight, caveat, or question.
Zhou et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: