基于逐波束同步的大规模MIMO传输技术研究

发布时间：2018-09-11 12:49

【摘要】：如今,大量无线终端的接入使得数据流量面临指数级的增长,这要求未来移动通信技术实现新的突破。其中,大规模多输入多输出(MIMO,multiple-inputmultiple-output)无线传输技术能够有效地提升系统吞吐量,被认为是5G的核心技术之一。毫米波由于其丰富的频谱资源以及易于与大规模MIMO相结合等优势,也引起研究者的广泛关注。然而将毫米波应用到移动通信系统最大的问题是由于频点较高而导致的严重的多普勒效应。正交频分复用(OFDM,orthogonal frequency division multiplexing)因其较高的频谱效率,和大规模MIMO相结合的MIMO-OFDM技术适用于高速率的宽带系统传输。然而,如果大规模MIMO-OFDM系统仍采用传统的正交导频,导频开销将成为限制系统性能的重要因素。本论文针对以上问题,研究了基于逐波束同步的大规模MIMO传输技术。首先,本文针对大规模MIMO波束域信道进行了建模和深入的研究。对大规模MIMO空间信道进行建模,研究了空间域信道与波束域信道之间的转换关系,进而提出波束域传输方法。在此基础上,揭示了波束域信道的稀疏特性,为本文基于压缩感知的信道信息获取方法提供了理论依据。接着,本文推导了波束域信道的近似表达,揭示了波束域信道多普勒扩展和时延扩展的带状特性,具体而言,与空间域信道相比,波束域信道的多普勒扩展和时延扩展将大幅减小,且多普勒扩展减小的倍数近似等于移动台天线数目,这为基于逐波束同步传输方法的研究提供了重要的理论依据。基于上述信道特性,本文针对移动场景下的毫米波大规模MIMO系统提出了基于逐波束同步的传输方法。考虑到移动台具有多波束的分集度,提出了基于波束域统计信道信息的贪婪用户波束调度算法。在基本时频同步技术的基础上,提出波束域时频同步方法,仿真实验说明了该方法参数估计的精度大大提高。由于波束域信道的多普勒扩展和时延扩展都大幅减小,本文提出了基于逐波束同步的上行联合接收和下行联合发送的传输方法,该方法能够有效地减小信道在频域和时域的色散,缓解毫米波大规模MIMO系统的多普勒效应。理论推导和数值仿真都表明,同步之后的等效信道在频率域和时间域都具有慢变特性,因此信道的相干时间和相干带宽都会增加,从而带来较大的性能增益。仿真实验验证了本文所提出的基于逐波束同步传输方法的性能显著优于传统的天线域同步传输方法,尤其当移动性增强或者移动台天线数目较多时,其性能优势更加明显。在逐波束同步基础上,本文研究了大规模MIMO-OFDM系统上行链路的导频设计与信道估计方法,以解决导频开销较大的问题。通过利用信道在波束时延域的稀疏性,将上行多用户系统的信道估计问题转化成一个压缩感知模型。在压缩感知的框架下,本文提出一种循环移位的导频设计,该导频被证明能满足有限等距性质,其中有限等距性质是压缩感知能够成功重建信号的必要条件。由于信道的稀疏性被充分利用,因此本文所提出的方法能大幅减小导频开销。进一步的,提出基于近似消息传递(AMP,approximate message passing)的信道估计方法,该方法能够以较低的复杂度进行近似贝叶斯推断。仿真结果说明,本文所提出的导频方案和基于AMP的信道估计方法显著优于其它方法,尤其在低信噪比时。此外,相较于传统微波大规模MIMO系统,毫米波大规模MIMO系统的性能优势更加显著。
[Abstract]:Nowadays, a large number of wireless terminal access makes the data flow face exponential growth, which requires a new breakthrough in future mobile communication technology. Among them, large-scale MIMO (multiple-input multiple-output) wireless transmission technology can effectively improve system throughput, which is considered as one of the core technologies of 5G. Due to its abundant spectrum resources and easy integration with large-scale MIMO, it has attracted wide attention of researchers. However, the biggest problem of applying millimeter wave to mobile communication system is the serious Doppler effect caused by high frequency point. Orthogonal frequency division multiplexing (OFDM) is the main reason. However, if the large-scale MIMO-OFDM system still uses the traditional orthogonal pilot, the pilot overhead will become an important factor limiting the performance of the system. Modular MIMO transmission technology. Firstly, this paper studies the modeling and in-depth study of large-scale MIMO beam-domain channel. The modeling of large-scale MIMO spatial channel is carried out, and the conversion relationship between space-domain channel and beam-domain channel is studied. Then, the approximate expression of the beam-domain channel is deduced, and the band-like characteristics of the Doppler spread and the delay spread of the beam-domain channel are revealed. The reduction of Doppler spread is approximately equal to the number of antennas of the mobile station, which provides an important theoretical basis for the research of beam-by-beam synchronous transmission method. Based on the above channel characteristics, a beam-by-beam synchronous transmission method is proposed for millimeter wave large-scale MIMO systems in mobile scenarios. A greedy user beam scheduling algorithm based on statistical channel information in beam domain is proposed for multi-beam diversity. On the basis of basic time-frequency synchronization technology, a beam-domain time-frequency synchronization method is proposed. Simulation results show that the accuracy of parameter estimation is greatly improved. In order to reduce the Doppler effect of millimeter-wave large-scale MIMO systems, a new method based on beam-by-beam synchronization is proposed, which can effectively reduce the channel dispersion in frequency domain and time domain and mitigate the Doppler effect of millimeter-wave large-scale MIMO systems. The simulation results show that the performance of the proposed method based on beam-by-beam synchronization is significantly better than that of the traditional antenna-domain synchronization transmission method, especially when the mobility is enhanced or the number of mobile antennas is increased. On the basis of beam-by-beam synchronization, this paper studies the pilot design and channel estimation methods for uplink of large-scale MIMO-OFDM systems to solve the problem of high pilot overhead. By utilizing the sparsity of the channel in the beam delay domain, the channel estimation problem of uplink multiuser systems is transformed into that of uplink multiuser systems. In the framework of compressed sensing, a cyclic shift pilot design is proposed. The pilot is proved to satisfy the finite equidistant property, in which the finite equidistant property is a necessary condition for compressed sensing to successfully reconstruct a signal. Furthermore, a channel estimation method based on approximate message passing (AMP) is proposed, which can perform approximate Bayesian inference with low complexity. Simulation results show that the proposed pilot scheme and AMP-based channel estimation method are significantly superior to other methods, especially. In addition, compared with traditional microwave large-scale MIMO systems, millimeter-wave large-scale MIMO systems have more significant performance advantages.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3

【相似文献】