| Summary: | 碩士 === 國立交通大學 === 電子研究所 === 107 === In this thesis, we introduce the topic of channel identification under millimeter wave (mmWave)
frequency and is used to speed up the establishment of the initial network entries between the base station and the mobile user. The contents include background, introduction, problem formulation,algorithm derivation and simulation results. We focus on the estimation of millimeter wave channel angular propagation information, exploiting the sparsity of the high-frequency channels, we view the channel estimation as a sparse signal reconstruction problem. Through the mathematical derivation,we sample the continuous angle in space by uniform grids then exploit the characteristic that mmWave
channel only has a few of dominant paths. We adopt the orthogonal matching pursuit (OMP) algorithm, make the use of quantized angle as a dictionary. The main channel propagation paths can be gradually found out in a recursive way. At the base station side, we use the uniform planar array antenna to transmit/receive and consider the three-dimensional channel model. The discrete Fourier transform (DFT) weight is applied as a beamforming gain to enhance the directivity of signal transmission to iii ensure the millimeter wave channel communication quality.
In the simulation, the normalized mean square error between the perfect channel and the estimated channel of the different channel direction is presented. We compare the convergence of the OMP method with different parameters (number of antennas, number of sampling points, signal to noise ratio) and under different multipath channels situation. In addition, the three-dimensional DFT beam pattern is also demonstrated.
|
|---|
