基于Wi-Fi信道状态信息的人员身份合法性认证

LEGITIMACY IDENTITY VERIFICATION BASED ON WI-FI CHANNEL STATE INFORMATION

针对传统人员身份合法性认证系统隐私性及舒适度问题,提出基于Wi-Fi信道状态信息(CSI)的人员身份合法认证系统,该系统将人员身份合法认证问题转化为相似度比较问题。对CSI信息进行降维和去噪;采用短时能量进行人员检测并获取行走的有效片段;提取离散小波变换(DWT)的近似系数作为特征;利用K-means算法找出每类合法人员数据的聚类中心点和距离半径,通过比较距离半径与未知人员数据到各个聚类中心点的距离进行人员身份合法性的判定。在真实的室内环境对该系统进行了评估,系统中合法人员为1到5人时,TPR和TNR可达到80%以上。实验表明,该系统能够有效地验证人员身份合法性。

Aiming at the privacy and comfort of the traditional legitimacy identity verification system,we propose legitimacy identity verification system based on Wi-Fi channel state information,which transfers the legitimacy identity verification problem to the similarity comparison problem.CSI information were reduced dimensionally and denoised.We used short-time energy to detect personnel and obtain effective segments of walking.Approximation coefficients of discrete wavelet transform(DWT)were extracted as features.We adopted K-means algorithm to find out the clustering center and distance radius of each class of legitimate personnel data.The identity legitimacy was determined by comparing the distance radius with the distance from unknown personnel data to each cluster center point.We evaluated the system in a real indoor environment.The result shows that TPR and TNR can reach more than 80%when there are 1-5 legal personnel in the system.The system can effectively verify the identity.