Research On Broadband Power Line Channel Modeling Method

Power line communication technology as an effective solution to solve the "last mile" of Internet of things terminal network,compared with other means of communication,the biggest advantage of power line communication is that it can penetrate into every corner of the power cable,and promote the construction of Internet of things better and faster.As the basis of power line communication,power line channel modeling technology is very important.Improving modeling algorithm and accuracy of modeling can improve the quality of power line communication.Firstly,the research background of power line channel modeling is introduced in this thesis,then the research status of the top-down and bottom-up modeling methods and the characteristics of the power line channel are described,and the processes of the two modeling methods are described in detail respectively.The fading,noise,radiation and impedance characteristics of low-voltage power lines are analyzed.In this thesis,the modeling process of multipath model is briefly introduced,and then the bottom-up modeling method is described based on the introduction of distribution parameters of power lines and two-port network.Secondly,the influence of different network topology parameters on power line channel attenuation is analyzed through measurement and simulation,including trunk length,branch length,branch location,number of branches and load impedance,and their influence modes and degree are also analyzed.Based on the radiation effect of power line,the calculation method of power line radiation is analyzed.By improving the calculation method of propagation constant,the radiation effect of power line is introduced into the bottom-up modeling method,which improves the accuracy of the bottom-up modeling method,and its validity is verified by actual measurement and simulation.Finally,on the basis of introducing the basic principle of generative adversarial networks,a statistical channel model of power line channel based on generating antagonistic network is presented.Based on the fact that the generative adversarial network can learn the potential distribution of real data,a large amount of random channel data is used for training,and the root mean square delay extension is used to verify the characteristics of the generated data.The results show that generative adversarial networks can generate statistical channels for power lines and can be used as random channel generators for power lines.