| Multiple Input Multiple Output (MIMO) and Orthogonal Frequency Division (OFDM) are adopted as the core technique in the physical layer of LTE systems. MIMO-OFDM can combat the frequency-selective fading, improve the system throughput and spectral efficiency. The channel estimation technique for LTE is not only the prerequisite of equalization and detection, but also the basis of link adaption, and plays an important role for the system performance. In addition, the complexity and variability of wireless channels may result in the degradation of open-loop system throughput. Therefore, it is necessary to research the closed-loop transmission scheme for the LTE downlink system. In this thesis, channel estimation and adaptive transmission for LTE downlink systems are investigated.Firstly, channel estimation algorithms for LTE downlink are investigated. As the frequency division multiplexing (FDM) technique is used to arrange the pilots, there exists no interference between antennas. MIMO channel estimation model can be simplified into single input single output (SISO) channel estimation model. Based on the SISO channel estimation model, least square (LS) channel estimation, linear minimum square error (LMMSE) channel esimation, discrete cosine transform (DCT) based channel estimation, partial symmetric extension discrete fourier transform(PSE-DFT) based channel estimation are studied for LTE downlink. Meanwhile, linear interpolation and cubic spline interpolation are also investigated. The interpolation algorithm makes use of the estimated channel value in pilot subcarriers to get the whole channel estimates in the frequency domain.Subsequently, compressed channel estimation algorithm using the time domain sparsity of the wireless channels is investigated. Firstly, three basic contents of compressive sensing are reviewed, including signal sparse representation, design of measurement matrix, and reconstruction algorithm. The greedy iterative reconstruction algorithms are mainly studied among the reconstruction algorithm. Secondly, the SISO-OFDM compressed channel estimation model is given, and greedy iterative reconstruction algorithms are used to estimate the channel parameters. Then, supposing that the channel parameters slowly vary in a subframe, the distributed compressed channel estimation model based on distributed compressed sensing (DCS) theory is given and the corresponding estimation algorithms are studied. Further, channel estimation is investigated in several special cases:1) When the pilot number compressed channel estimation used is half of the number of the LS’s, the compressed channel estimation can still gain a better performance than the LS channel estimation algorithm; 2) Compressed channel estimation makes use of the channel response which is filtered by the windowed Discrete Cosine Transform and achieves some performance gain; 3) As there exists time correlation between the OFDM symbols in a subframe, the correlation is considered to design a two dimensional (2D) measurement matrix. Based on the 2D measurement matrix, the 2D compressed channel estimation algorithm is proposed. Simulation results show that the 2D compressed channel estimation can get a better performance than 1D channel estimation.Finally, the link adaptive transmission technique is investigated. Firstly, the feedback parameters of link adaptation are reviewed, including rank indication (RI), precoding matrix indicator (PMI) and channel quality indicator (CQI). Secondly, the method of jointly evaluating RI and PMI based on the metric of sum MMSE rate is given. Thirdly, the effective signal-to-noise-ratio (SINR) mapping methods are reviewed in order to compute the effective signal-to-interfer-ratio in the range of time and frequency. The effective SINR is used to choose the suitable CQI which meets the requirement of the codeword error rate (CWER). Last but not least, simulations about CWER and spectrual efficiency in different secenories are performed and simulation results are analysed. |
PDF Full Text Request