Study Of Channel Estimation Techniques In Millimeter-Wave Large-scale Antenna Arrays

Millimeter wave technology can provide rich frequency band resources for communication systems and has become one of the key technologies for 5G in recent years.The high frequency and short wavelength of millimeter wave enable the system hardware design to support the integration of large-scale antenna arrays in a small size range.As another key technology in 5G system,large-scale antenna array technology can effectively increase the system capacity and also increase the antenna gain during signal transmission.The combination of millimeter wave technology and large-scale antenna array technology significantly increases the data transmission rate to meet the application requirements of 5G systems.In millimeter-wave large-scale antenna array systems,channel estimation has been the key problem that has been widely studied.Accurate channel state information plays a crucial role in the scheme design of precoding and beam assignment in the whole communication system.Currently,most of the available channel estimation algorithms are serial in computational structure.Due to the significant increase in the number of antennas and the sparse nature of the channel in millimeter-wave largescale antenna array systems,the complexity of the channel estimation algorithms increases,posing a higher challenge for fast and accurate implementation of channel estimation.To address this problem,the main research contents and results of this paper are as follows:(1)To address the problem of high complexity of existing channel estimation algorithms,this paper introduces the accelerated projection consistency algorithm for solving large-scale linear equation systems into channel estimation based on linear system theory,and proposes accelerated projection-based consensus channel estimation algorithm.The algorithm adopts a parallel computing structure to distribute the channel estimation task to several parallel sub-operating units for solving,and then the global solution is obtained by the core-operating unit,and the error between the solved value and the true value is reduced through circular iterations.An upper bound on the error of the estimation performance of the algorithm is derived after theoretical analysis.The simulation validation finds that the channel estimation algorithm based on accelerated projection consistency achieves better convergence by fewer iterations compared with the reference algorithm,which can effectively reduce the computational complexity while ensuring the estimation performance,and provides a new idea for the implementation of parallel processing of channel estimation in millimeter-wave large-scale antenna array systems.(2)For the sparse characteristics of the channel in the millimeter wave large-scale antenna array system,this paper proposes a sparse accelerated projection-based consensus algorithm based on the compressed sensing theory,which further reduces the complexity of the algorithm based on the accelerated projection-based consensus channel estimation algorithm.For simulation verification,the algorithm is compared with three reference algorithms in terms of error parametrization and accurate recovery rate.The experimental results show that the proposed algorithm effectively reduces the complexity of the algorithm while ensuring the estimation accuracy.