Research On Physical Layer Key Generation Mechanism Based On Wireless Channel Characteristics

Wireless communication network has the advantages of good mobility,high scalability,flexible deployment and control,etc.,which makes the application of wireless communication more and more popular.However,the open channel-based communication process makes such wireless applications more vulnerable to various security threats.The current security mechanisms are typically established through asymmetric encryption or symmetric encryption.However,these encryption methods are either computationally complex or have challenges such as secure key distribution.In contrast,the physical layer key generation mechanism ensures wireless nodes can directly generate shared keys by utilizing the reciprocity and spatial decorrelation of wireless channels,and thus achieving efficient and flexible protection for wireless communication.The main work of this thesis is as follows:(1)The commonly used randomness sources in the physical layer key generation technology are the received signal strength and channel state information.Compared with the received signal strength,the channel state information contains more channel states.However,the traditional physical layer key generation methods only use its amplitude or phase to generate the shared key,which significantly impedes the efficiency of key generation.In addition,the existing physical layer key generation algorithms usually use traditional quantization methods which is not flexible.Based on the above analysis,this thesis proposes an improved quantization method using k-means clustering,which can more efficiently convert the channel states into bits.Further the impact of the proposed quantization method on physical layer key generation is analyzed.The experimental and simulation results show that the proposed quantization method is superior to the existing quantization methods and can improve the efficiency of key generation.(2)Aiming at the problem that although the channel states collected by nodes have a high correlation,there are still many differences,and it is proposed to use data normalization processing combined with an improved information coordination method to effectively improve the efficiency of the key agreement process.The improved information reconciliation scheme proposed in this thesis is based on a digital sum operation.Based on the above mechanism combined with the K-means clustering and quantization process,a complete physical layer key generation process is realized.The analysis results show that the improved information reconciliation process can effectively enhance the efficiency of physical layer key generation.