Detection And Positioning Of Rogue Access Points Based On Wi-Fi Signal Sensing

With the widespread deployment of Wi-Fi networks and the rapid development of mobile intelligent terminal technologies,Wi-Fi networks have become an important way to connect to the Internet,with advantages of low cost,high bandwidth and ease of deployment.The access points(APs)is an important infrastructure in the Wi-Fi networks.However,due to the openness of the network,attackers can deploy a rogue AP(RAP)that mimics some legal AP and induces users' connection,while eavesdroping the traffic or redirecting users to a phishing website,and thus the RAP attacks can cause user privacy leakage and even property damage.Therefore,it is of important significance to defending against the RAP attacks to ensure Wi-Fi network security.To this end,this dissertation comprehensively utilize two different kinds of signal data with different granularity,i.e.,received signal strength(RSS)and channel state information(CSI),and study the RAP detection and positioning technology based on Wi-Fi signal sensing.The contributions of the dissertation are as follows.First,the technology of RSS-based RAP detectionis studied.We use the cluster analysis of RSS vectors to detect the existence of RAP.Due to the fact that distributed acquisition of RSS data in the pracitcal network environment will cause missing values in some components of the aggregated RSS vectors,it will affect the RAP detection results and make the false alarm rate higher.To solve this problem,we propose a practical RAP detection approach that can eliminate the influence caused by missing valuesin RSS data.In the data preprocessing stage,the RSS data is dealed with data filling,filtering and averaging so as to prevent the excessive amplification of the distances between RSS vectors due to the missing RSS values.In the stage of cluster analysis,part of all components are dynamically involved in the calculation of RSS vector distance,so that the distance offset caused by the missing RSS values is further reduced.The experiments are conducted in real environment.The results show that the detection approach can reduce the false alarm rate significantly on the condition that it can ensure a high detection rate.Then the technology of CSI-based RAP detection is investigated.In order to localize the RAP with unfixed transmit power,we propose an RAP positioning approach based on CSI fingerprints that contains CSI data of nodes under multiple transmit power levels.In the off-line phase,for each reference point,CSI data of RAP working on multiple transmit power levels is collected,and multi-power CSI fingerprints will be constructed and saved in database.In the online phase,CSI data of the RAP is collected and the positioning algorithm based on Bayesian probability is used to estimate the position of the RAP.The experiments are conducted in real environment.The results show that the proposed positioning approach can decrease the RAP positioning error without the knowledge of the current RAP transmission power and the approach can achieve a desirable accurancy in RAP positioning.Finally,this dissertation designs and implements a prototype system on the basis of the proposed approaches of RSS-based RAP detection and CSI-based RAP positioning,and this prototype system consists of a central controller,multiple RSS monitoring nodes and CSI monitoring nodes.It has several functional modules such as AP information acquisition,CSI data acquisition,RAP detection and RAP positioning,and it can detect and locate RAP effectively.In summary,we investigate the RAP detection and positioning technology based on Wi-Fi signal sensing in the disseration.This technology can assist network administrators in discovering and removing RAPs in order to ensure access security of terminal and users' privacy in Wi-Fi networks.