The Research Towards The Propagation Characteristics Of Electromagnetic Wave In 25GHz For Tunnel Environment

There will be a generous increase in the number of mobile users,monthly network throughput for each user and the data traffic in the next few years.This is one of the most difficult challenges that the Wireless Communication System(WCS)is facing with.Two other challenges for the WCS are the emerging intelligent services and the constantly improving standard of user requirements.Because of these challenges,a technical innovation is the trend of times.In this case,International Telecommunications Union started the prospective studies of the 5th generation mobile communication in 2012,which was formally named “IMT-2020”.Until now,as for China,the researches and test works of the 5th-Generation are focusing on frequency band below 6 GHz,while the research findings of the 5G with high frequency especially in 25 GHz are relatively few.In this thesis,the transmitting properties of electromagnetic wave with 25 GHz under tunnel environment are studied systematically.The main results of the article includes three aspects as below:1)The relationships between the Propagation Path Loss of 25 GHz electromagnetic wave and the tunnel environmental parameters are tested and analyzed.The results show that the main factors for the Propagation Path Loss include the relative dielectric constant of the tunnel wall,roughness of the tunnel wall,tunnel's sectional area and tunnel's sinuosity.When the relative dielectric constant increases from 1 to 7,both the mean received power and the attenuation constant of far-field change significantly.However,the mean received power and the attenuation constant of far-field tends towards stability after the relative dielectric constant increases over 7.The more coarse the tunnel surface is,the lower the mean received power will be and the bigger the attenuation constant of far-field is.The bigger the tunnel's sectional area is,the higher the mean received power will be.The more curved the tunnel is,the lower the mean received power and attenuation constant will be.But the relationship between the tunnel's sectional area and attenuation constant of far-field is still undetermined.2)The small-scale fading,time delay properties and the Doppler Effects of the 25 GHz electromagnetic wave under tunnel environment are studied.The results show that the further the distance is,the smaller the wave range of received power will be,the smaller the maximum deviation and the fading depth of received power will be.With the increase of distance between transmitting antenna and receiving antenna,the Ricean K factor decreases first and increases later.In straight tunnel with constant sectional area,the time delay increases firstly and then decreases with the increase of the distance.For the same distance between transmitting antenna and receiving antenna,the time delay increases with the sectional area larger.The reflection number increases with the increasing degree of the tunnel's bending,which leads to the bigger distance differences and more complicated relationships between the time delay effect and the distance.Under the tunnel environment,the relative frequency shift only depends on the object's moving speed,while the Doppler shifts depends on both the object's moving speed and the wave's frequency.3)The prediction functions of two different kinds of artificial neural network are tested and analyzed.The artificial neural network is inspired by the biological neural network and has the ability of approximating and predicting the relation between several variables.In this project,the distances and tunnel's bend rates are used as the input neuron,and the received power as the output neuron.The first artificial neural network was built with Back Propagation algorithm and Radial Basis Function model.The second RBF artificial neural network was built with the distance and tunnel's sectional area are used as the input neuron,and the received power as the output neuron.According to the test result,the prediction accuracy of the artificial neural network is quite high with few seconds to operate.The trained BP network can only predict the received power change tendency along with the distance.While RBF network can predict not only the relationship between the received power and the distance,but also the local variation tendency of the received power.