Research Of Precoding And Scheduling Algorithms In Downlink Multiuser MIMO System

MIMO (Multiple Input Multiple Output) is the very key technique applied to improve system throughput and spectral efficiency. It is widely used in the4th Generation mobile communication system, for example the LTE-Advanced and WiMAX (802.16m). A variety of related promising technologies have emerged, such as the transmission mode selection:"multiplexing" or "diversity", Space Time Coding/Space Frequency Coding, Cyclic Delay Diversity; Antenna Selection, Power Allocation, Precoding and Scheduling. Besides, MIMO can also be combined with some other techniques, such as, Channel Estimation, Channel Coding and Decoding. Especially, the combination of MIMO and OFDM has been the foundation of the Physical Layer of4G mobile network.Currently, the research of the multiple antenna technology follows several major aspects, including the modeling of MIMO channel, the analysis of MIMO channel capacity, the precoding technology, the multi-user detection and the receiver design. All of these technologies put emphasizes on different areas, whereas, by overcoming the bad impact of the fading channel, the fundamental purpose of them is to increase the system throughput and enhance the link reliability. In this work, we develop some innovative algorithms and strategies on precoding and scheduling in downlink multi-user MIMO. When the channel correlation is considered at the transmitter and receiver, an original antenna combining algorithm is derived to maintain the performance of precoding. Besides, a low complexity-orthogonality based user/antenna scheduling scheme is also proposed. Then, a joint optimization of precoding and user scheduling algorithm outperforms some of its counterparts.Firstly, based on the theory of MIMO, we analyze some typical precoding techniques, and provide some relevant simulation results. Particularly, when channel correlation exists, the distribution of the singular value of the channel matrix disperses. Based on Block Diagonalization precoding, we develop an Antenna Combining algorithm to improve the system throughput. The essence of this algorithm is to adjust the channel space at the cost of sacrificing degrees of spatial freedom. Even under the strong correlation condition, this algorithm can still get a sum rate performance boost.Secondly, according to Block Diagonalization, we propose a low-complexity orthogonality-based user/antenna scheduling algorithm. Since the Block Diagonalization tries to orthogonalize different users’channel matrices, it will choose the users, who are more orthogonal to the others. In that case, we extend the orthogonality principle to the scheduling of Block Diagonalization precoding. Some simplifications are applied. Compared with some other schemes, this algorithm not only has much lower complexity, but also achieves almost the same sum rate performance.Finally, we propose an innovative joint optimization of precoding, scheduling and receiver processing algorithm. In this algorithm, the candidates will be evaluated in a "Greedy" manner, and when a user is newly added to the scheduling subset, its post processing matrix will also be calculated. When the scheduling is completed, the calculations of reception matrices also have been finished. Besides, a very concise and effective code-book based feedback algorithm is also proposed. Both the scheduling itself and the combination of limited feedback and scheduling outperform some other similar algorithms.