基于Wi-Fi信道状态信息的身份识别

Identity Recognition Based on Wi-Fi Channel State Information

人的行为感知技术在人机交互中起着重要作用,其中动作识别和身份识别技术应用广泛。传统的行为感知技术需要人们佩戴传感器,且设备成本高。为此,本文提出了一种基于Wi-Fi信道状态信息(Channel State Information,CSI)的身份识别系统。该系统包括数据采集,数据预处理,行走区间检测,分类识别4个阶段。首先,在实验室环境下采集Wi-Fi网卡中的CSI数据并提取幅值信息;其次,通过Butterworth滤波器消除环境噪声从而得到稳定且无噪声干扰的数据;使用行走区间检测算法(Anomaly Detection Algorithm,ADA),检测出行走区间;最后,提取特征值,通过支持向量机(Support Vector Machine,SVM)算法进行分类识别。实验结果表明,随着人数从2~4人变化,平均识别率为87.5%~95%。

Human behavior perception technologies play an important role in the human-computer interaction.Among them,action recognition and identification technology are widely used.The traditional behavior perception technology requires people to wear sensors,so the cost of implementation is high.In this work,we propose a system of human identification based on Wi-Fi channel state information(CSI)signals.The system includes four stages:data collection,data preprocessing,walking interval detection,and classification and recognition.At first,the CSI data were collected by NIC in the laboratory environment and the CSI amplitude information was extracted.Then,the environmental noises were removed through the Butterworth filter to obtain stable and noise-free data.The anomaly detection algorithm(ADA)was applied to detect the walking interval.Finally,the system extracts the features from the CSI time series and employs the support vector machine(SVM)classifier to classify a detected target.Our experimental results indicate that the identification average accuracy is about 87.5%to 95%from a group of 2 to 6 people,respectively.