MetaBreath: Multitarget Respiration Detection Based on Space-Time-Coding Digital Metasurface

Breathing is an important health metric for tracking diseases, and non-contact respiration detection under multitarget scenarios has attracted wide attention in recent years. However, conventional solutions are difficult to achieve breathing separation when people sit closely or sleep together sharing the same bed, otherwise complex hardware system is required. In this context, this article proposes MetaBreath: a multitarget and high-resolution respiration detection method based on space-time-coding (STC) digital metasurface. To design MetaBreath, we fully explore the harmonics generated by STC digital metasurface, and these harmonics are equivalent to the multiple virtual sensor nodes evenly distributed in different directions, making the observations more comprehensive. Then we model respiration detection as the problem of blind source separation (BSS) when people are close to each other, and solve it by the technique of independent component analysis (ICA). Experiments show that MetaBreath enables respiration detection and separation whether people sit apart or close together, which is reflected in the recovered waveform keeping consistent with ground truth. Besides, for the first time, we realize up to four-person detection in this single-input single-output (SISO) system, which is outperforming the state-of-the-art methods. With both low complexity and high practicality, MetaBreath is expected to apply to sleeping monitoring, life detection, and baby care.