Research And Implementation Of 5G NR Channel Modeling And Beam Management Mechanism

Combining large-scale antenna technology with millimeter-wave communication is an indispensable means for improving the transmission rate of the fifth generation mobile communication system(5G).With the increase of the working carrier frequency,more antennas need to be deployed in the communication system to obtain beamforming gain to compensate for the high path loss brought by the high frequency band.The beam used for communication between the base station and the user will become thinner in this scenario.The determination of the beam pair used for communication between transceiver will take longer,which will have a significant impact on the normal communication of the user.Therefore,beam management becomes more important in the millimeter wave band.The research on the transmission characteristics of millimeter-wave channels is helpful for more efficient beam management.Therefore,this paper first studies and implements the millimeter-wave channel modeling based on TR 38.901 protocol,and verifies the accuracy of the realized channels.Secondly,the beam management mechanism in the 5G NR protocol is analyzed.Based on the 5G NR link-level simulation platform,the synchronization signal block for the initial access phase and the sounding reference signal for the beam thinning phase are implemented.The reception performance is verified in the single/multi-user scenario.The above simulations provide an important reference for the R&D of the base station products of the project partners,and have certain engineering significance.Finally,combining an optimization algorithm with beam management,a two-stage fast beam search algorithm based on stochastic gradient descent is proposed for ultra-large-scale antenna scenarios in the frequency band above 52.6 GHz in the beyond 5G period.It can be used in the broadband hybrid beamforming system based on partial connection architecture.The number of beam searches and interaction delays are significantly reduced.Compared with exhaustive search,the performance loss of the algorithm is less than 0.95%and 0.8%in non-line of sight and line of sight scenarios respectively.