RADHOME:Radar based Human Activity Dataset in Home Environments

With the rapid advancements in smart home technologies, health monitoring, and safety management systems, there is a growing demand for improving quality of life and ensuring security. In this study, we present a dataset of daily human activities constructed using millimeter-wave (mmWave) radar technology, aimed at providing a robust foundation for non-contact health monitoring, safety management, and smart home systems. The dataset involves 10 participants of diverse genders and body types, performing six activity transitions in a specified sequence: sitting to walking, sitting to stretching, walking to picking up, sitting to drinking water, walking to falling, and falling to waving for help. Each activity was repeated 20 times, resulting in 1,200 activity samples. The radar system was positioned at a height of 0.8–1 meters, with a detection range of 2–5 meters, ensuring precise capture of participants' indoor movements. This dataset is made publicly available with comprehensive documentation to facilitate reproducibility and usability. It offers a valuable resource for developing advanced human activity recognition algorithms and contributes to research on behavioral patterns, with applications in smart home systems, health monitoring, and safety management.

随着智能家居技术、健康监测与安全管理系统的快速演进，社会对提升生活品质与保障安全的需求愈发迫切。本研究构建了一套基于毫米波（millimeter-wave, mmWave）雷达技术的日常人类活动数据集，旨在为非接触式健康监测、安全管理及智能家居系统提供坚实的支撑基础。该数据集涵盖10名性别与体型各异的受试者，按照指定序列完成6类活动转换动作：坐姿切换至行走、坐姿切换至拉伸、行走切换至拾取物品、坐姿切换至饮水、行走切换至跌倒，以及跌倒后挥手求助。每项动作重复执行20次，总计生成1200组活动样本。雷达系统部署高度为0.8至1米，探测范围为2至5米，可精准采集受试者的室内运动数据。本数据集已公开并附带完整文档说明，以保障研究的可复现性与易用性。该数据集为先进人类活动识别算法的开发提供了宝贵资源，同时助力行为模式相关研究，可应用于智能家居系统、健康监测及安全管理领域。