ElderAL-CSI

We present the first Wi-Fi CSI dataset for aging-in-space that jointly labels daily activities, fall-risk events, and location. Collected in a single indoor room, it explicitly covers nine positions (3\u00d73 grid) and six actions, including confusable pairs (e.g., lying down vs. falling). (ZTE_WiFi_Sensing) using a ZTE AX3000 as the RF front-end. Because acquisition depended on instantaneous link quality, the sampling rate and MIMO configuration varied; most recordings used a 2\u00d73 TX\/RX setup with 512 subcarriers. Experiments took place in an unfurnished 172 m\u00b2 room (Fig. 2). Three volunteers (2 males, 1 female; mean age 26) performed six daily activities at nine grid positions spaced 1.2 m apart in both axes; at each position, every user repeated each activity 15 times. The activities were: 0\u2014stationary standing, 1\u2014walking, 2\u2014falling, 3\u2014lying down (sleep), 4\u2014drinking, and 5\u2014sitting down while turning on the TV.