Wi-Fi Device Identification Scheme Based On Channel State Informatio

With the rapid development of wireless networks,Wi-Fi technology has become indispensable in our daily lives.While convenience has been brought by wireless internet,network security issues have also been raised.One of the most serious issues is rogue access point attacks.A rogue access point attack refers to a counterfeit access point being privately deployed by an attacker,identity information of legitimate access points is imitated,unintended users are induced to connect,and sensitive information is stolen.Large amounts of property damage are caused every year,causing widespread attention in the academic community.An effective way to address the above security threats is to establish an effective identity authentication mechanism between access point devices and networked devices.However,the current wireless device identity detection schemes have the following defects and deficiencies: 1.High complexity,complex algorithms,and additional hardware configurations are required by existing wireless device detection schemes,and difficulties are brought to system maintenance and management.2.Inefficiency,plenty of computing resources is required by existing schemes,and efficiency is affected.3.Limited application scenarios,solutions are most often focused on scenarios such as the Internet of Things and distributed networks,but lack research in the field of Wi-Fi devices.Therefore,the identity detection of commercial Wi-Fi devices has been targeted for research,the channel state information of the signal is analyzed,the hardware fingerprint of the device is calculated,convolutional networks are used as device fingerprint recognizers,and accurate identification of Wi-Fi device identity is achieved.The detailed research content and implementation plan are as follows.· A lightweight and efficient Wi-Fi device recognition scheme has been proposed to solve device identity spoofing attacks in wireless networks.The wireless signal of WiFi devices is connected,the channel state information of the signal is extracted,and fingerprints with unique features are calculated.A simple and effective convolutional network model is designed,and the fingerprints of wireless Wi-Fi devices are learned and recognized.Additional software or hardware has been abandoned,only Wi-Fi signals are adopted.In addition,the computational cost required to extract channel state information is tiny,so the scheme consumes little system resources and can be theoretically implemented on mobile devices.After experimental verification,the recognition accuracy of this scheme can reach 91%.· To further enhance the stability,the amplitude matrix fingerprint is proposed,and the amplitude of the signal is calculated to obtain the amplitude matrix fingerprint,which,together with the phase error,is used as the device fingerprint.This scheme is called P-A.In the P-A scheme,wireless devices are connected,and CSI was analyzed to obtain P and A fingerprints,which were input into the P-A convolutional network model and trained and recognized.A more stable and accurate device identification method is implemented by P-A,and through the combination of two fingerprints,more accurate and reliable identification results are provided.After experimental verification,the performance is maintained stable when conditions such as time,location,and environment change.Through extensive experiments,it has been proven that the P-A scheme can reach the recognition speed within 0.5 seconds and improve the accuracy to around 95%.