Abstract:

The daily behavior detection of indoor human based on CSI is developing rapidly in the field of WSN.At present,most of the research is still in the environment of 2.4 GHz,so the detection rate,robustness and overall performance still need to be improved.In order to solve this problem,a passive indoor human behavior detection method HDFi (Human Detection with Wi-Fi) based on CSI signal was proposed.The method was used to detect the indoor human daily behavior in a 5 GHz band environment,which was divided into three steps:data acquisition,data processing,feature extraction,online detection.Firstly,the experiment collected typical daily behavioral data in complex laboratory and relatively empty meeting room.Secondly,the amplitude and phase data with more obvious features were extracted and processed by low-pass filtering to obtain a set of stable and noise-free data,and then the fingerprint database was established effectively.Finally,in the real-time detection stage,the collected data features were classified by SVM algorithm to extract more stable eigenvalues,and a classification model of indoor human daily behavior detection was established,and then matched the data in the fingerprint database.The experimental results show that the proposed method has the characteristics of high efficiency,high precision and good robustness,and the method does not need any testing personnel to carry any electronic equipment,so it has high practicability.

Key words: channel state information, behavior detection, low pass filtering, support vector machine