User selection schemes for downlink multiuser mimo with codebook based precoding

Abstract

In this thesis, we deal with the user selection schemes for the downlink multiuser MIMO systems with codebook based procoding schemes. In multiuser MIMO system, active users who share the spectral resources experience the inter-user interference. It is generally assumed that each user does not cooperate to each other. Therefore, it is required that the base station eliminates the interference using the precoding techniques such as dirty paper coding (DPC), channel inversion, vector perturbation, and so on. In general, the base station apply the precoding schemes based on the downlink channel state information (CSI). Although the perfect CSI guarantees the good performance, for example, the DPC can acquire the theoretical sum-capacity of the multiuser MIMO systems, the feedback overhead wastes the uplink resources. Thus, the limited feedback schemes have been researched. In addition, it is an important procedure to select the users group of which the users at least interfere to each other. Per user unitary rate control (PU2RC) and semi-orthogonal user selection are well known schemes to support the user selection of the multiuser MIMO with the codebook based precoding. They guarantee to minimize the inter-user interference using the orthogonal or semi-orthogonal codeword grouping and orthogonal precoding (channel inversion). After briefly examining two algorithms, we shows that these systems have limitations in performance of sum-capacity and average BER. We also explain the the trade off between the codebook size and the channel quantization error which degrades the performance of channel inversion precoding. As the larger size codebook makes the less quantization error, the performance of system is improve using it. However, when the number of total users who request the service is less than the codebook size, the base station does not efficiently use the spatial resources. That is, there is capacity loss in the large codebook with the referential scheme...