2018_ubicomp Wistep: device-free step counting with wifi signals

本文研究WiFi信号来估计步频计数的问题

*step counting: selecting subcarriers and limb speed 

*50volunteers

*two indoor environments

*90.2% and 87.59%, respectively

*device-free

FMCW bandwidth:1.79GHz

WiFi bandwidth:20 or 40MHz

*****************************************************************

分为两步：walking detection, step counting

 

butterworth bandpass filter:带通滤波主要获取一定频率范围内的信号,具体来说就是移除高频的噪音和低频的呼吸以及心跳

收发设备的配置：

transmitter with one directional antenna(这个就是限制啊)

receiver with 3 omni-directional antennal

sampling packet:1000Hz(发包率有点高)

##Considering human activities in traditional indoor environment introduce frequencies of no more than 300Hz in CSI amplitudes(CARM), thus WiStep is configured with a sampling rate of 1000Hz that is larger than the Nyquist sampling rate##

上述图展示了躯干部份与腿部分对信号反射产生的频率不同。

参考的文献：

[1]WifiU: gait recognition using wifi signal, 2016

[2]CARM: understanding and modeling of wifi signal based human activity recognition,2015

 

 

 

evaluation metrics:

 

***************************************************************

acceleration and deceleration phases:加减速阶段

step counting=count walking steps

accelerometers

gyroscopes

sensory data:感知数据

light condition, power consuming, and personal privacy

stem from