Antenna Selection And Precoding For MIMO Communication Systems

With the quickly development of wireless communication technology, the requirement for quality of service and data transmission rate is continually improving. Multiple-input multiple-output (MIMO) can significantly increase the capacity of communication systems, also improve the system performance. So MIMO fits the high-data service in wireless communications very well. Therefore, MIMO has been one of the key technologies for next generation wireless communication systems.The technique of antenna selection (AS) can get MIMO system gain with low complexity and low cost, and so it has achieved significant attention. Group layered space-time (GLST) system, which combines with space-time block coding (STBC) and layered space-time (LST), can achieve a better diversity-multiplexing tradeoff, and it makes for meeting the diversified communication services. The efficient antenna selection for GLST systems is a problem deserving of deep study.In multi-user MIMO broadcast channels, we can make use of precoding at the base station to reduce inter-user interference, and achieve multiuser diversity gain by the efficient user selection to increase the data rate. Both schemes require channel state information at transmitter, but in practical FDD systems, channel knowledge is conveyed via uplink feedback from the users to the base station transmitter. Precoding and user selection in multi-user broadcast channels with limited feedback has been a key question in MIMO systems.In this dissertation, we focus on antenna selection algorithms for GLST systems and user selection and precoding schemes for multi-user broadcast channels. The main contents and contributions of this dissertation are as follows:Firstly, transmit antenna selection (TAS) is applied to GLST system, and a novel TAS algorithm is proposed based on the restructuring channel and QR decomposition (QRD). To reduce the complexity, a fast antenna selection algorithm is presented based on the reduced-element set (RES). According to column vectors sorted by Euclidean norm's, RES of transmit antennas is built, and the active antenna subset is obtained from RES. Simulation results show that this algorithm significantly outperforms the existing fast algorithms, and when the elements in RES are chosen properly, it has similar performance with the TAS algorithm based on the complete set.Secondly, based on QRD of the restructuring channel, a suboptimal receive antenna selection (RAS) is provided for GLST systems with ordered successive interference cancellation (OSIC) detector. A permutation matrix can be obtained by sorting according to the Frobenius norm of sub-channel instead of the Euclidean norm of the row vector in the Moore-Penrose pseudo-inverse matrix, so that computational complexity is reduced. And then, a fast algorithm based on equivalent channel is provided. The numerical simulations show that the proposed algorithms can achieve near-optimal outage capacity, and the performance of the provided ORD-based algorithm is close to that of pseudo-inverse-based algorithm. There is a slight performance gap between the two algorithms, but the fast one with lower complexity.Thirdly, multiuser random beamforming (RBF) yields degraded performance for more practically low to moderate users. A combined RBF and zero-forcing beamforrning (ZFBF) precoding scheme is presented for MIMO broadcast channels. Based on RBF, the initial scheduling is achieved, and then the selected users are requested additional feedback on quantified channel direction information (CDI) and channel quality information (CQI). The final user selection is implemented to maximize the achievable sum rate at the base station, and then we obtain the precoding design by using ZFBF. Numerical results show that the presented precoding strategy significantly improve sum-rate performance, compared to the existing enhanced RBF scenario for systems with the not so large number of users, without increasing feedback bits.Finally, for multi-user broadcast channels with the definite total feedback bits, a joint user selection and precoding scheme is proposed based on the optimum threshold and the optimum feedback rate. The algorithm starts by setting a threshold value based on the data analysis for the cumulative distribution function of channel gain information, only the users whose channel gain values are above the set threshold value are considered. Then, the algorithm randomly chooses N users from the candidate user set to feed back their CDI and CQI, and N is corresponding to the optimum feedback rate. Finally, according the feedback, user schedule and ZFBF precoding are achieved in the base station transmitter. Achieving better multi-user diversity gain, the proposed scheme can obtain more sum rate capacity than the optimum feedback rate algorithm.