Beamforming For The Downlink In A Multiuser MIMO System

Multiple-input Multiple-output (MIMO) communication systems have received more and more interest because of their potential for dramatic gains in channel capacity. And it is thought to be the key technology of 3G and "Beyond-3G" mobile communications, which is supposed to provide high quality and high data rate multimedia packet transmission than current voice-oriented legacy wireless systems. In this thesis, we will discuss transmit and receive techniques in downlink and focus on multi-user MIMO (MU-MIMO) system under the limited feedback scenario.Firstly, we introduce the MIMO system capacity, MIMO space-time coding techniques, and the performance analysis of different detection algorithms including ZF, ZF-OSIC, QR-SIC, MMSE and MMSE-OSIC in V-BLAST system. It shows the performance of MMSE-OSIC is obviously better than other algorithms and ZF-OSIC is better than QR-SIC.After that, we focus on the beamforming technique for the downlink MU-MIMO system, beamforming comes from smart antenna, and it can improve signal to noise ratio (SNR) and system capacity apparently. In this thesis, we discuss zero forcing (ZF), block-diagonalization (BD) and Opportunistic beamforming. Assuming perfect channel knowledge at the transmitter, in MU-MIMO channels, ZF and BD algorithm can eliminate inter-user interference perfectly and implement space-division multiple access (SDMA). Simulation shows bit error ration (BER) of multi-users are almost the same with single user. Opportunistic beamforming can achieve high system capacity and good user fairness, its performance is close to BD under large amount number of users. To overcome the bad performance of opportunistic beamforming in small amount users' scenario, we design a system based on subspace channel selection and multi-beam opportunistic beamforming. This method can reshape small amount users MIMO system to large amount users MIMO system and improve the probability of channel parameter match. At last, the performance simulation also proves this method can apparently improve the system capacity of opportunistic beamforming in small amount users' scenario.