Study On Precoding Technique In MIMO System

With the rapid development of communication technology, pepole have morehigher demands for data transmission rate and Spectrum utilization rate.Multiple inputand multiple output technology can increas spectrum utilization rate exponentiallywithout increasing transmit power, so it attracts lots of attentions and studies recently.There are many limits for signal because of the complex wireless channel, precodingtechnology can compensata the channel fading and make a better match between thetransmission signal and communication channel. It is a key issue to designhigh-performance, low-complexity precoding matrix, which we should note.The thesis studied on the precoding technology in MIMO system,according tosome criterions to design,such as zero-forcing,minimum mean square error,maximumcapacity and so on.Tying to get some improvements on the bit error performance andsystem capacity. The main content is displayed as follows.①Starting from zero-forcing criterion,studied the precoding algorithms includingchannel inversion (CI),selective channel invertion(SCI),correlation rotation(CR) andjoint transmit-receive optimization. Aim at the drawback of enlarging noise by ZFcriterion, designed selective channel invertion(SCI), correlation rotation (CR) based onMMSE criterion. According to characteristics of Inter-Channel Interference(ICI) signal,proposed the algorithm combined CR(SCI) with transceiver optimization, and gived thedesign processses of various linear precodings under imperfect channel stateinformation. Finally, the simulations of these algorithms were given to prove thesuperiority of the designed algorithms. Simulation results shows that the proposedalgorithms has obviously advantage.②Studied on some precoding technology based on channel decomposition;Thisthesis gived design processes of Singular Value Decomposition(SVD), Geometric MeanDecomposition(GMD) and Uniform Channel Decomposition(UCD). Aim at thedrawback of neglecting the power of interference signal of GMD(UCD), designedGMD（UCD）algorithm jointed with SCI and CR technology. Simulation results showsthat the proposed method can improve the system BER performance about2dB.③Aim at the drawback of error transfer effect and more capacity loss of linearprecoding, studied the nonlinear procoing called Tomlinson-Harashima proding(THP).Analyzed the factors that affect the BER performance, designed three improved methods, including combined with diversity coding, combined with power allocation,combined with GMD decomposition. Studied the factors that affect capacity, gived outthe capacity computing method of MMSE-THP system, and capacity formulas ofZF-THP and MMSE-THP system under imperfect channel state information. Aimed atthe capacity loss reason, designed MMSE-THP system combined with beamformingbased on maximum capacity criterion and proved by theoretical analysis and somesimulations.