Research And Design Of Human Motion Detection Technology Based On WLAN

In recent years,due to the rapid development of WLAN technology and the popularity of indoor WLAN devices,with the use of Wi-Fi signal to realize basic communication,utilize the existing WLAN technology for indoor human motion detection has attracted the attention of the numerous scholars and holds potential to ubiquitous smart applications.This technology does not require intruders to bring wireless electronic devices into the monitoring area.It can be implemented on the existing WLAN infrastructure,no additional hardware devices.Thus,this technology is very practical and commercial application protect.Despite of the prevalent Wi-Fi signal feature Received Signal Strength(RSS),more reliable and robust signal feature Channel State Information(CSI)in the IEEE 802.11n standard has been proposed recently years.CSI contains finer description of the state of the wireless channel and more complete description of the environment feature in the multipath transmission of the wireless signal.With the new CSI information,more reliable and robust indoor human detection system can be realized.In this thesis,we explore to present an indoor human motion detection system(CETW)based on CSI phases.The system can achieve accurate detection of intrusions in complex indoor scenes.Based on the original human motion detection method,two improvements are proposed to adjust the actual application scene of the dynamic moving speed of invaders:(1)We improve the correlation distance method to extract the eigenvalues of CSI matrix,and measure the difference the CSI matrix and the standard CSI matrix in static sciences by the eigenvalue.As demonstrated in the experiments,compared with the previous method based on the mean and variance extraction of CSI matrix eigenvalues,this method can avoid abnormal CSI eigenvalue omission on intrusion detection and false alarm.It can obtain the accuracy of 4%to 7%improved.(2)We propose the LCR/SCR parameters to improve the accuracy of intrusion detection in the dynamic speed of intruders.The LCR with larger scrolling window is adopted when the moving speed is faster.When moving slowly,use the SCR with smaller scrolling window.The actual test results show that using the LCR/SCR parameters can improve the accuracy of detection by 8%to 10%where the intruders are moving too slowly(0.3m/s)or too fast(2m/s).The CETW system is evaluated from different indoor environments,numerous antenna pairs and the dynamic speed of invaders.As demonstrated in the experiments,our human motion detection system gains a high detection rate which reaches more than 95%and a low false alarm rate.