大规模MIMO系统容量优化理论方法研究

随着天线技术以及集成工艺的发展，大规模MIMO系统成为可能并且将在未来无线通信网络中发挥重要作用。本课题以大规模MIMO系统为研究对象，拟提出更为实际的系统模型并针对实际模型进行信道容量分析并提出适用于大规模MIMO系统的信道估计和多用户检测方法，具体内容包括：(1)通过理论分析评估实际系统信道容量，根据分析结果优化天线配置、载频复用系数等参数。充分利用大规模天线资源、提高频谱效率。(2)根据信道矩阵在频域和空间域的稀疏特性，提出基于压缩感知的信道估计方法，提高信道估计精度、节约计算成本。(3)针对上行多用户信号检测计算复杂度大的问题提出稀疏检测和次优检测相结合的检测方法，在不损失检测精度的前提下降低计算复杂度。另外，我们将搭建实验系统来验证研究结果。本课题兼有理论研究价值和实际应用价值，研究成果可为实际系统提供理论指导，并为系统参数设计及优化提供依据。

Massive MIMO has been proposed recently and has attracted a lot of interests from both academic world and industry. It has been proved that under the assumption of unlimited number of antennas at the base station, the spectrum efficiency can be increased greatly while on the other hand, the transmit power can be significantly reduced without sacrificing the quality of service. Due to these facts, massive MIMO has been considered as a powerful candidate for the future wireless communication system. However, there remain a lot of open issues about massive MIMO. The capacity analyses in the literatures used an ideal system model and important parameters such as correlation among antennas and co-channel interference have not been addressed properly. The channel estimation algorithms for the downlink communication assume pilot-aided channel estimation at mobile terminals and feedback of the compressed channel state information. Such algorithms utilize the valuable bandwidth resource and suffer from the additional noise at the uplink receiver. As a result, the channel estimators become inaccurate and spectrally inefficient. In the uplink transmission, multi-user detection becomes extremely computational complicated due to the large number of users and low-complexity detection algorithms have not yet been reported according to the applicant's knowledge. Therefore, the focus of this project is on massive MIMO system by using a more practical system model. System modeling by taking into consideration the parameters such as antenna correlation and frequency reuse factor will be carried out at first to setup a more practical system model for massive MIMO system. Based on this practical system model, three problems will be solved: (1) System capacity analysis. The system capacity will be re-formulated as a closed-form function of all the system parameters. In addition, the system parameters such as antenna separation and frequency reuse factor can be optimized based on the analytical results to satisfy the system requirement as well as to maximize the spectrum efficiency. (2) Compressive sensing based downlink channel estimation. Due to the fact that correlation exists among antennas, the components in the channel matrix will exhibit sparsity after proper transforming. In this project, reciprocity between uplink and downlink is used and a channel estimator based on compressive sensing theory will be proposed to achieve high accuracy and low computational complexity as well. (3) Low-complexity multi-user detection for uplink transmission. Based on the fact that the channel matrix is sparse in frequency-domain or space-domain, the time-domain multi-user detection can be transformed and reformed as a sparse detection problem. In addition, sub-optimal time domain detection algorithms such as QR decomposition based M-algorithm can also be used together with the sparse detection to further reduce the computational complexity.