Research On Parametric Channel Modeling And Simulation Technologies For 5G Communications

In addition to satisfying traditional continuous wide-area coverage,the fifth-Generation(5G)mobile communication system will also realize inter-connections of everything in future Internet of Things(IoT).Its user experience rate,traffic density,latency,and energy efficiency are the key performance indicators in different application scenarios.High efficiency transmission such as massive antennas,and ultra-dense networking technologies are indispensable for a deep understanding of 5G channel propagation characteristics and models.Compared with 4G,5G channel models are with higher requirement in frequency consistency and space-time continuity at millimeter waves(mmWave).To meet with the requirements of 5G channel model standardization,new transmission technology assessment and system design,in this dissertation,channel measurements,modeling and simulation based on 5G candidate frequency bands and typical scenarios are carried out.The main research contents and innovative contributions are embodied in:(1)Channel modeling and simulations in 5G key mmWavee frequencies for outdoor microcellsBased on Keysight's advanced time domain channel sounder,channel measurements for outdoor microcells are conducted at 26,32 and 39 GHz,respectively.The channel parameter extraction and clustering algorithms and modeling methodologies for path loss and Ricean factor as well as the mm Wave channel characteristics are investigated based on large amout of measurement data.Space-Alternative Generalized Expectation-maximization(SAGE)algorithm is applied to extract channel parameters,and the parametric and non-parametric modeling results are compared.The influence of environments on channel statistical properties are analyzed and investigated in these new mmWave frequency bands.Finally,the statistical channel models are built.Moreover,based on the established channel models and parameters and Quasi-Deterministic Radio Channel Generator(QuaDRiGa)recommended by 3GPP and ITU,research on the statistical and time-evolution features of channel parameters are carried out to validate the accuracy of QuaDRiGa channel simulator in mm Waves.(2)Time-varying channel modeling and simulation for Urban microcellsTraditional GBSM has its inherent shortcoming with poor space-time continuity.In this regard,a novel dynamic time-varying model based on multipath birth and death process is established,in which the advantage of GBSM statistical characteristics are still kept,meanwile,better space-time continuities are studied by using the MIMO channel measurement data at 2.55 GHz in urban microcells to bulid a time varying channel model considering multipath birth and death procedures.The model parameters and simulation steps are described in detail,and it's verified with comparisons of the time-varying power delay profiles and statistical channel parameters between simulation and mesureemnt results.(3)Time-varying channel modeling by neural networksBased on machine learning theory,novel time-varying channel models by neural networks are investigated and established.In order to playback the large-and small-scale fading channels with respect to time or distance,the RBF neural network is applied to model the large-scale path loss including shadow fading and joint small-scale channel parameters by using 5G mmWave channel measurement data.The models built have excellent space-time consistencies which can accurately playback the measured large-and small-scale channel models and parameters.(4)D2D radio channel characteristics and modeling.Using the D2D MIMO channel measurement data performed in indoor corridors and rooms,D2D radio channel characteristics are studied when both of the transceiver ends are moving and one end is fixed(F2M:fixed-to-mobile).,double-Rayleigh and multiple-Rayleigh are applied to analyze the distributions of received powers.SAGE is used to extract channel parameters,the angular distribution of azimuth and elevation AoAs and AoDs as well as correlation are investigated.The analytical expression of 3D Doppler power spectral density for the D2D channel is derived,and 2D and 3D Doppler spectral are modeled and verified by measurements.The statistics of channel parameters between D2D with both terminals moving and traditional F2M are studied as well.The received power of D2D channel follows Rayleigh or Rayleigh and double-Rayleigh joint distributions in the line-of-sight and non-line-of-sight scenarios,respectively.(5)MmWave massive antenna Channel characteristics.MmWave channel measurements using a real array and virtual massive antennas in the waiting hall at Qingdao north high-speed railway station are carried out at 28 GHz for the first time in China,the statistics of channel parameters are investigated and the modeling results are given.The results show that with higher number of antennas,the spatial resolution of multipath increases significantly while the number of clusters remains almost the same,the rms delay and angular spread tends to decrease and increase,respectively.