This article presents the construction of WiSense, a human activity recognition dataset based on WiFi signals. With the application of machine learning in the field of WiFi sensing, the availability of datasets becomes a vexing problem because collecting WiFi data is a time-consuming and costly process. WiSense consists of data from ten volunteers performing six different activities in two rooms. We conducted the data collection using publicly available CSI collection tools and simple devices, with illustrated collection steps and data processing methods. Our goal is to provide researchers with a publicly available WiFi dataset for use in activity recognition studies, and to enable non-specialists to replicate the collection process. We also perform experiments on the WiSense dataset using six popular activity recognition methods to demonstrate the effectiveness of the dataset. The dataset can be found in GitHub at https://github.com/lzl19971105/WiSense.

中文翻译：

---

Wisense：基于 WiFi 的人类活动识别数据集


本文介绍了 WiSense 的构建，这是一个基于 WiFi 信号的人类活动识别数据集。随着机器学习在 WiFi 传感领域的应用，数据集的可用性成为一个令人烦恼的问题，因为收集 WiFi 数据是一个耗时且昂贵的过程。 WiSense 由 10 名志愿者在两个房间中进行六种不同活动的数据组成。我们使用公开的CSI收集工具和简单的设备进行数据收集，并附有图解的收集步骤和数据处理方法。我们的目标是为研究人员提供一个公开可用的 WiFi 数据集，用于活动识别研究，并使非专业人士能够复制收集过程。我们还使用六种流行的活动识别方法在 WiSense 数据集上进行实验，以证明数据集的有效性。该数据集可以在 GitHub 中找到：https://github.com/lzl19971105/WiSense。