Research On Channel Estimation In Millimeter Wave Massive MIMO Hybrid Beamforming System

Millimeter wave technology can be perfectly combined with Massive MIMO technology to improve the data throughput of next-generation wireless communication systems and ease the pressure on spectrum resources.At present,due to the large channel dimensions of Massive MIMO systems,traditional channel estimation methods cannot be used to estimate mmWave channel.Thus,exploring reliable and effective channel estimation algorithms for mmWave massive MIMO systems is a major challenge in academia.To develop efficient mmWave Massive MIMO channel estimation algorithms,this thesis mainly uses the sparsity of mmWave channels and treats the channel estimation problem as a sparse reconstruction problem.Firstly,a multi-resolution hierarchical codebook structure is proposed in this thesis,and an adaptive channel tracking algorithm based on such codebook is designed.The AoA/AoD of the channel and path gain are estimated by searching codebook to find beamforming vectors which maximize the received signal.Simulation results show that the proposed adaptive channel estimation algorithm based on hierarchical codebook can achieve the spectral efficiency equivalent to exhaustive search algorithm within a certain SNR range.The number of iteration is less and the complexity of this algorithm is lower.Additionally,in order to further explore mmWave channel estimation problems,an algorithm is proposed by extracting out-of-band spatial information,with lower complexity and less training overhead.The algorithm mainly uses the spatial consistency of high and low frequency channels to establish the frequency-dependent Sub-6GHz channel and mmWave channel,and extracts Sub-6GHz channel information to obtain angle estimations of the mmWave channel.MmWave codebook is then designed in the direction of the larger mmWave channel gains,and hybrid beamforming can be finished by beam training.Simulation results show that the angle information of mmWave channel can be obtained by leveraging the out-of-band information of Sub-6GHz.Also,hybrid beamforming can achieve higher spectral efficiency.Moreover,with low SNR of systems,the out-of-band information aided mmWave channel estimation algorithm can achieve better performance and avoid the disadvantage that long-distance users may not be accessed.The algorithm also effectively reduces the search range of mm Wave beam,and the complexity is lower,which can effectively save time resources and reduce training overhead.