Research On Channel Estimation And Channel Equalization Algorithms For LTE-MIMO Downlink Systems

This thesis mainly researches on channel estimation algorithm and channelequalization algorithm of MIMO-OFDM system under the wireless broadband mobilecommunication environment. Firstly, OFDM channel estimation algorithms based onthe frequency-domain pilot are discussed, including the LS algorithm, MMSE algorithm,the LMMSE algorithm by the SVD decomposition, time domain DFT transformalgorithms. An FIR low-pass filter noise reduction algorithm is proposed, which aims toimprove the performance of the conventional existing algorithms and to reduce theircomplexity. Then, various commonly used channel interpolation algorithms arediscussed, including constant interpolation, linear interpolation, quadratic interpolationand DFT interpolation. The advantages and disadvantages of them are analyzed andcompared. Finally, the performance of various channel estimation algorithms arecompared and verified through simulations.Since the introduction of the cyclic prefix in OFDM system, the system cancompletely eliminate inter-symbol interference in theory. Channel equalization can beachieved by using a simple single-tap frequency-domain filter at the receiving end.However, considering the multipath characteristics and its time variability of wirelessmobile channel, this thesis describes several traditional channel equalization algorithmsand the possibility of using them, including frequency-domain zero-forcing algorithm,traditional MAP algorithm and QR-MAP algorithm based on QR decompositionprinciple. In addition, non-linear channel equalization algorithms based on decisionfeedback principle are presented, such as QR-SIC-MAP and QR-SIC-ZF， whichbelong to serial interference cancellation algorithm。Finally a simplified MAP algorithmcalled QRM-MAP is proposed. By setting a proper number of constellation pointswhich have the minimum Euclidean distance weight, the proposed algorithm canachieve fine approximating performance.