Research On Beamspace Channel Estimation Algorithm For Mmwave Massive Mimo Systems

Millimeter-wave(mm Wave)massive multi-input multi-output(massive MIMO)technology has the characteristics of huge bandwidth resources and higher multiplexing gain,which can significantly improve the communication performance indicators such as user throughput,spectrum efficiency,energy efficiency and so on.It is extremely important in the next generation and future mobile communication development.However,mm Wave massive MIMO also brings a series of technical challenges.Among them,the multiplexing gain of multiple antennas can be achieved by beamforming strategy,and the mm Wave spatial channel will be converted into beamspace channel.and the high channel dimension brought by a large number of antennas makes the beamspace channel estimation quite difficult.Moreover,because the beamspace channel has different properties from the traditional spatial channel,the traditional channel estimation method cannot continue to be used and needs to be improved.Therefore,this dissertation investigates the channel estimation algorithm under the beamspace of mm Wave massive MIMO system.In this dissertation,we address the problems of high complexity,low accuracy,and high pilot overhead of the channel estimation method under beam space in the currently available millimeter-wave massive MIMO system.From the characteristics of the millimeter-wave channel,the channel estimation problem is constructed as a sparse reconstruction problem by combining the MIMO system structure,and the sparse recovery algorithm is improved to enhance the estimation accuracy of the beamspace channel estimation.First,a block-sparse based support detection(BSD)algorithm is proposed for narrowband beamspace channel,where the different channel component elements of each propagation path are rearranged and combined to obtain a new equivalent channel vector,which has the nature of block sparsity,and the special nature of beamspace channel is used to estimate the support,so that the support of each channel component can be treated as a whole to achieve improvement of channel estimation accuracy and reduction of algorithm complexity.Furthermore,due to the large bandwidth characteristics of millimeter waves,future mm Wave communications are more likely to be broadband systems,and the wideband context will bring new challenges to channel estimation.To this end,this dissertation firstly describes the channel estimation problem as a carrier frequency beam direction estimation problem by exploiting the special frequency-dependent sparse structure of wideband beamspace channel,and then proposes a beam-band based orthogonal matching pursuit(BBOMP)channel estimation algorithm by constructing a beam-band function,using it to obtain the signal power of each path component,and then the support of each path component at different sub-carriers is estimated jointly to minimize the effect of beam squint and to break the assumption of common support of existing schemes so as to improve the estimation accuracy,and finally the obtained support is used to complete the sparse recovery of the OMP algorithm to obtain the estimated channel vector.Simulation results show that the BSD algorithm and the BBOMP algorithm proposed in this dissertation have better estimation accuracy and lower pilot overhead than the conventional algorithms and existing algorithms,respectively,under their respective circumstances.