| Multiple antenna techniques have been under hot discussion for more than 10 years in the wireless communication area. By utilising spatial resources, mutiple antenna system can provide higher diversity and multiplexing gains without requiring more bandwidth or time. The potential gains make MIMO (Multiple Input Multiple Output) be the core technique in the physical layer of IMT-Advanced. The rapidly growing techniques of precoding and user scheduling are key research points in all MIMO systems, from single cell single user MIMO to single cell multiuser MIMO, and now multiple cells multiuser MIMO.In this dissertation, from the perspectives of matrix analysis, information theory and signal processing theory, several hot topics in the MU-MIMO research field are investigated, including the user scheduling with novel criterion, the cross-layer design of precoding and user scheduling with limited feedback, the cooperative precoding in multi-cell scenario, and the highly efficient prediction-based CSI (Channel State Information) feedback scheme. These reseaches strive to explore the essence of signal transmission, and achieve the potential of system performance. For feasibility in practical engineering and compatibility with industry standards, the proper tradeoff between performance and complexity is taken into account as another research goal. The main content of the dissertation is summarized as follows:Chapter 1 briefly reviews the research background of MIMO precoding and user scheduling, including the concept definitions and classifications, which are relevant to the research work of this dissertation.Chapter 2 introduces a leakage based user scheduling in MU-MIMO broadcast channel. In the proposed scheme, a user's leakage power to other users is an important factor to be used for determining whether the user should be selected. The accuracy of closed-form solution of SLNR (signal to leakage plus noise ratio) makes the proposed scheme outperform the traditional scheduling method based on SINR. The proposed algorithms can work in case of perfect or limited feedback CSI at transmitter side. Relationship among the feedback overhead, the number of users to be selected and the number of transmit/receive antennas is analyzed, which can be seen as a design guideline.In chapter 3, a joint user pairing and precoding scheme in MU-MIMO broadcast channel with limited feedback is investigated. In the proposed scheme, each user feeds back its preferable precoding vector index as well as two scalars denoting quantization error and the largest singular value of channel matrix, respectively. Base station (BS) selects the first user who has highest estimated SINR, and generates a unitary precoding matrix based on the first user's preferable precoding vector. Then, BS selects users one by one. In each step, BS calculates the estimated SINR for each un-selected user, using the generated precoding vector, and then chooses the user with largest estimated SINR.Chapter 4 proposes a decentralized cooperative precoding scheme in multi-cell MIMO system. In the system model, each user feeds back its CSI to all neighboring BSs, and each BS uses the CSI to generate its precoding vector independently. Three classical single cell precoding methods are extended to the multi-cell scenario, and based on them, three iterative updating transceiver design algorithms are proposed. Chapter 5 investigates a prediction based CSI feedback scheme in multi-user MIMO system. In the proposed scheme, base station and user equipments predict channel vectors. Users only quantize and feedback the error between predictive value and real value of channel. A codebook design using the idea of Lloyd algorithm is proposed, with the objective function of minimizing the mean value of prediction error. BS uses the quantized error index and its own predictive value to reconstruct the real time channels.Finally, the contributions in this dissertation, as well as suggestions for future work, are summarized in chapter 6.
|
PDF Full Text Request