Precoder Design For Multiuser MIMO Communication System

Multiple-input multiple-out(MIMO) is widely considered as an effective approach to improve the spectrum efficiency,which becomes more and more important due to the rapid growth of wireless communication. As one of the key technologies for MIMO systems, multi-user precoding technology can provide higher data rate and link reliability. The first goal for the precoding is to solve the multiuser interference problem in multiuser MIMO systems. In dealing with multiuser interference problem by precoding matrix, it can lead to transmit power enhancement. Moreover, nonlinear operations for muliti-user precoding can increase the computational complexity. Focusing on the above two issue, some precoding schemes with different structures are proposed for various scenarios. Our main contents include:1). A low complexity lattice reduction aided precoding based on orthogonality defect(od) threshold is proposed. To reduce computational complexity, the orthogonality defect threshold as an early-termination condition is introduced into lattice reduction(LR) algorithm by adaptively early terminating the LR processing. Moreover, to achieve a favorable tradeoff between performance and complexity, we define a power loss factor(PLF) to optimize the od threshold. Simulation results evidence that the proposed algorithm can achieve signi?cant complexity savings with nearly the same bit-error-rate(BER) performance as the traditional LRA precoding algorithm.2). Tomlinson-Harashima precoding algorithms with two different structures is studied in multiuser MIMO downlink. To increase the system capacity, an improved algorithm is proposed. By adopting different scaling factor to precancel the interference signals for the two basic THP structures, the entropy-value of the equivalent noise can be decreased. Thereby, the proposed algorithms can increase the system capacity, which is confirmed by the simulation results. Besides, the effect of the precoding order for central THP algorithm is also discussed. Based on the discussion, a low complexity ordering algorithm for the central THP is proposed. Results show the proposed algorithm can save complexity greatly compared to conventional ordering algorithm.3). A multi-level lattice reduction aided TVP(LRA TVP) is proposed, which introduces lattice reduction as preprocessing to avoid computational expensive sphere encoder as much as possible. First, we define search probability, which can evaluate the probability for the execution of the sphere encoder after the first lattice point is found. It is shown that LRA TVP decreases the search probability of the sphere search for a given SNR. Then, we show that LRA TVP has a lower upper bound for the volume of the search space compared that of conventional TVPs, which means search complexity of the LRA TVP can be reduced. Moreover, to further reduce preprocessing complexity, a multi-level LLL algorithm is proposed. Results show the proposed algorithms can achieve significant complexity gains compared to conventional TVP.4). Precoder design is considered for the downlink of multi-user MIMO systems where antenna arrays are used at both the base station and the mobile receiver. Different from previous schemes which completely eliminate multi-user interference(MUI), the proposed precoder obtains an optimum compromise between noise enhancement and MUI mitigation. Furthermore, by extracting diversity from each user’s subspace and reducing precoding loss, the proposed scheme can offer a significant gain in terms of average bit error rate(BER).