| 引用本文: | 席瑞,张佳,孙一淼,何源.基于毫米波的人体感知研究进展.软件学报,2025,36(11):5241-5273 |
| |
|
| |
|
|
| 本文已被：浏览 1431次 下载 1785次 |
码上扫一扫！ |
|
|
| 基于毫米波的人体感知研究进展 |
|
席瑞1, 张佳2,3, 孙一淼2,3, 何源2,3
|
|
1.电子科技大学 计算机科学与工程学院, 四川 成都 611731;2.清华大学 软件学院, 北京 100084;3.北京信息科学与技术国家研究中心, 北京 100084
|
|
| 摘要: |
| 随着嵌入式技术、移动计算技术、物联网等的快速发展和广泛应用, 越来越多的感知设备渗透到人们的日常生活中, 例如智能手机、摄像头、智能手环、智能路由器、耳机等, 这些设备上的传感器可以采集与人们的活动、健康、社交等息息相关的个人信息, 催生了一类新的感知应用-以人为中心的感知(human-centric sensing). 与传统的感知方法, 如可穿戴设备感知技术、计算机视觉感知技术、无线信号感知技术等相比, 基于毫米波信号的感知技术具有高感知精度、非视距、被动感知(无须携带传感器)、高时空分辨率、易部署、良好环境鲁棒性等一系列优势. 基于毫米波的感知技术成为近年来学术界和工业界研究热点, 能够实现对人体活动、体征等信息的非接触式细粒度感知. 因此, 在重新梳理相关研究的基础上, 分析基于毫米波的人体感知技术兴起的背景和研究意义, 总结现有基于毫米波人体感知应用工作和技术, 包括人体跟踪和定位、运动识别、生物测量和人体成像, 并介绍目前常用的公开数据集. 最后, 讨论潜在的研究挑战以及展望了未来的方向, 实现精准、泛在、稳定的人体感知. |
| 关键词: 毫米波 人体感知 非接触式 物联网 |
| DOI：10.13328/j.cnki.jos.007378 |
| 分类号:TP393 |
| 基金项目:国家自然科学基金(U21B2007) |
|
| Review on mmWave-based Human Perception |
|
XI Rui1, ZHANG Jia2,3, SUN Yi-Miao2,3, HE Yuan2,3
|
|
1.School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China;2.School of Software, Tsinghua University, Beijing 100084, China;3.Beijing National Research Center for Information Science and Technology, Beijing 100084, China
|
| Abstract: |
| With the rapid development of embedded technology, mobile computing, and the Internet of Things (IoT), an increasing number of sensing devices have been integrated into people’s daily lives, including smartphones, cameras, smart bracelets, smart routers, and headsets. The sensors embedded in these devices facilitate the collection of personal information such as location, activities, vital signs, and social interactions, thus fostering a new class of applications known as human-centric sensing. Compared with traditional sensing methods, including wearable-based, vision-based, and wireless signal-based sensing, millimeter wave (mmWave) signals offer numerous advantages, such as high accuracy, non-line-of-sight capability, passive sensing (without requiring users to carry sensors), high spatiotemporal resolution, easy deployment, and robust environmental adaptability. The advantages of mmWave-based sensing have made it a research focus in both academia and industry in recent years, enabling non-contact, fine-grained perception of human activities and physical signs. Based on an overview of recent studies, the background and research significance of mmWave-based human sensing are examined. The existing methods are categorized into four main areas: tracking and positioning, motion recognition, biometric measurement, and human imaging. Commonly used publicly available datasets are also introduced. Finally, potential research challenges and future directions are discussed, highlighting promising developments toward achieving accurate, ubiquitous, and stable human perception. |
| Key words: millimeter wave (mmWave) human perception non-contact Internet of Things (IoT) |
|
|
|
|
