Research On Physical Layer Security Of Power Line Carrier Communication For Cooperative Relay

Security is the cornerstone of communication network.It is an important index to measure the performance of communication system.Compared with the traditional communication network,the power line carrier communication network faces more serious security threats.The traditional encryption and decryption algorithms and key negotiation mechanism have been unable to meet people's demand for the security performance of power line communication.Therefore,the demand for the security performance of the physical layer of power line communication,which is the lowest layer of OSI model,has emerged.The purpose of this paper is to improve the security of the physical layer of power line carrier communication by studying the eavesdropping channel model,the optimal power distribution strategy and the optimal relay link selection strategy.First of all,according to the Wyner physical layer security model,four nodes channel model is established,which contains artificial noise.The signal of the relay node and the legal destination node and eavesdrop node is analysed.Through the instantaneous channel SNR of the destination node and the eavesdrop node,secrecy capacity is calculated.According to the features of lognormal distribution of power line channel,the average secret channel capacity is calculated by properties of moment generating function.The influence of artificial noise and the distance of relay node and eavesdrop node to the average channel secrecy capacity is simulation and the influence of artificial noise effect on the properties of the physical layer security is analysed.Secondly,in the problem of node power distribution in communication system,the traditional power distribution method leads to low efficiency,and optimization of power distribution strategy can improve the instantaneous channel security capacity.The power optimization problem of the power wiretap channel model is a non-convex optimization problem.In this paper,a power optimization strategy based on fully connected neural network is proposed to optimize the power distribution by fitting the nonlinear relationship between the channel coefficient and the power distribution of the source node,the relay node and the artificial noise.The results show that with the change of channel coefficient,the maximum instantaneous channel security capacity can be guaranteed by allocating power according to the predicted value of neural network.Finally,in view of power line channel relay selection problem of the Multi-relay eavesdrop model,this paper proposes a relay selection algorithm based on convolution neural network,with the second order tensor relay link channel coefficients for the input,the relay selection tags for the output,and convolution neural network training,the relay link with maximum instantaneous channel secrecy capacity is chosen.This algorithm transforms the relay selection problem into the classification problem,and directly solves the classification problem through the convolutional neural network,greatly simplifying the computation.The 1*3 convolution kernel based convolutional neural network established in this paper and the accuracy of selecting the optimal relay link is up to 96.7%.The theoretical analysis of this paper is verified by Python language simulation and tensorflow2.1 deep learning framework.