Research On Hardware Design Of Beamforming In High And Low Frequency Fusion Networking

Under the current background of full commercial use of 5G,the millimeter wave frequency band has received extensive attention from researchers due to its large bandwidth and narrow beam as a key frequency band in the 5G system.A series of technologies suitable for millimeter wave communication such as massive MIMO transmission,multiple access,and denser networking connection methods,have emerged to effectively improve the efficiency of wireless resource utilization.Massive MIMO beamforming technology can take full advantage of the strong directivity of millimeter waves to form high gain narrow beams with high spatial resolution,the quality of received signal has been improved and the inter-user interference has been much reduced,thus,higher system capacity and spectrum efficiency was achieved to make up for the high path loss of the millimeter waves in non-ideal propagation environments.The use of a fully digital beamforming system in the low band can achieve optimal system performance.However,due to hardware complexity,energy consumption and cost constraints in high frequency communication systems with massive MIMO arrays,most researchers process signals based on the hybrid beamforming architecture.Existing experimental results have proved that the sub-6GHz band and the millimeter wave band have a high spatial correlation.Some researchers take advantage of this property to use low band information to assist beamforming in high and low frequency fusion networking systems,which greatly reduces millimeter wave training costs.This article first analyzes several commonly used beamforming architectures,technical principles and applicable conditions,as well as their advantages and disadvantages,and models millimeter wave related channels and received signals.Then the software simulation and performance comparison of beam selection algorithms,which exploits out-of-band information assistance and millimeter wave in-band information,are carried out.The results show that the beam selection algorithm based on out-of-band information assistance can reduce training overhead by several times compared with exhaustive search,and the performance is better than in-band beam selection under the same conditions.Next,several SVD beamforming algorithms commonly used in hardware systems are studied and compared.The optimized beamforming algorithm is more suitable for hardware implementation and has advantages in terms of computational complexity and resource consumption.Based on the software simulation and hardware implementation of the beamforming module of 16x8 MIMO communication system,this paper compares the performance between digital beamforming and hybrid beamforming schemes,designs a five-channel parallel beamforming hardware framework,performs functional verification and FPGA board is used to complete the test,the optimized beamforming module takes into account the trade-off between processing performance and resource consumption to meet system performance requirements.Perform the bit error rate simulation for the 4x4 MIMO communication system,develop the serial beamforming hardware module and the corresponding receiver test system based on resource saving,and complete the hardware test.