大规模MIMO组网场景中的JSDM用户分组技术研究

发布时间：2018-07-22 17:23

【摘要】：多输入多输出(MIMO)技术从诞生开始,就因其可以增加空间自由度,提升移动通信系统的性能增益而备受关注。然而,随着智能终端及其应用的快速发展,海量的联网终端所产生的巨大数据流量对现有的移动通信网络造成了巨大的挑战。传统MIMO技术已逐渐无法满足需求。大规模MIMO技术作为传统MIMO技术的延伸,在继承了传统MIMO技术的优点以外,还极大地提高了系统的频谱效率,显著地降低了能源消耗、硬件成本以及空口延迟等等,成为下一代无线蜂窝通讯系统的热门候选技术之一。但是,目前大规模MIMO技术的实用性较差,仍然有许多问题亟需解决。其中,随着天线数目与用户数目的增加,大规模MIMO系统进行信道估计时所需的下行训练与反馈的开销也将剧增。如何减少这些开销,是大规模MIMO传输技术走向实用首要解决的问题。联合空分复用(JSDM)下行传输方案的主要思想是利用信道的空间相关性对各小区内的用户进行分组,然后对各用户组分别进行两级的混合预编码,能够有效地减少频分双工(FDD)模式下的大规模MIMO系统在进行信道估计时的开销。其中,用户分组是实施JSDM方案的第一步,且传统JSDM方案是在单小区场景中进行的,所以本文将着重研究多小区场景中的JSDM用户分组技术与干扰协调技术。在已有的JSDM用户分组算法中,K-means分组算法因简单、有效而使用广泛,但是该算法采用随机选取初始用户组中心点,这对最终的分组结果影响很大,容易产生局部最优解。针对此缺陷,本论文提出一种改进的K-means用户分组算法,采用最大距离法选取初始中心值,并使用加权似然准则替换原算法中的欧式距离准则对用户的差异性进行度量。在经过多次迭代后,算法收敛,用户可被充分分组。但是,在多小区场景中,该改进的用户分组算法无法消除或抑制小区间干扰。所以,本文后续又提出了一种多小区下的JSDM干扰协调方案。该方案在JSDM技术的基础上,利用了信道的零空间预编码设计原理,将蜂窝网络中的每一个用户的有效信号投射到其余全部用户的信道零空间中,继而可以有效抑制多小区间干扰。通过修改已有的长期演进(LTE)系统仿真平台,对本论文所提出的改进用户分组算法及零空间干扰协调方案的性能进行了全面验证和深入分析。仿真结果表明,在大规模MIMO组网场景中,改进的JSDM分组算法能够有效地解决由传统K-means算法随机初始化导致的局部最优解问题,使得用户分组更为彻底,JSDM下行传输方案的性能提升了近20%。此外,基于JSDM方案的零空间干扰协调方法有效地抑制多小区间的干扰,从而小区边缘5%频谱效率和系统的总体吞吐率最高均能提升2.2倍左右。
[Abstract]:Multiple input multiple output (MIMO) technology has attracted much attention since its birth because it can increase the degree of freedom of space and improve the performance gain of mobile communication systems. However, with the rapid development of intelligent terminals and their applications, the huge data traffic generated by massive networked terminals has posed a great challenge to the existing mobile communication networks. Traditional MIMO technology has been unable to meet the demand. As an extension of the traditional MIMO technology, the large-scale MIMO technology not only inherits the advantages of the traditional MIMO technology, but also greatly improves the spectral efficiency of the system, and significantly reduces the energy consumption, hardware cost, empty port delay and so on. It has become one of the popular candidate technologies for the next generation wireless cellular communication system. However, there are still many problems to be solved because of the poor practicability of large-scale MIMO technology. With the increase of the number of antennas and the number of users, the downlink training and feedback overhead of large scale MIMO systems will be greatly increased. How to reduce these overhead is the most important problem for large-scale MIMO transmission technology. The main idea of the joint space division multiplexing (JSDM) downlink transmission scheme is to use the spatial correlation of the channel to group the users in each cell. It can effectively reduce the channel estimation overhead of large scale MIMO systems in FDD mode. User packet is the first step to implement JSDM scheme, and the traditional JSDM scheme is carried out in a single cell scenario. Therefore, this paper will focus on the JSDM user packet technology and interference coordination technology in multi-cell scenarios. Among the existing JSDM user packet algorithms, K-means grouping algorithm is widely used because of its simplicity and efficiency. However, the algorithm adopts random selection of initial user group center points, which has a great influence on the final grouping results and is easy to produce local optimal solutions. To solve this problem, this paper proposes an improved K-means user grouping algorithm, which uses the maximum distance method to select the initial center value, and uses the weighted likelihood criterion to replace the Euclidean distance criterion in the original algorithm to measure the difference of the user. After several iterations, the algorithm converges and users can be fully grouped. However, in multi-cell scenarios, the improved user packet algorithm can not eliminate or suppress inter-cell interference. Therefore, a JSDM interference coordination scheme for multi-cell is proposed in this paper. Based on the JSDM technology, the scheme uses the zero-space precoding design principle of the channel to project the effective signal of each user in the cellular network into the channel null space of all the other users. Then it can effectively suppress multi-cell interference. By modifying the existing long-term evolution (LTE) system simulation platform, the performance of the improved user packet algorithm and the zero-space interference coordination scheme proposed in this paper are verified and analyzed thoroughly. Simulation results show that the improved JSDM packet algorithm can effectively solve the local optimal problem caused by random initialization of traditional K-means algorithm in large-scale MIMO networking scenarios. The performance of the JSDM downlink transmission scheme has been improved by nearly 20%. In addition, the zero-space interference coordination method based on JSDM scheme can effectively suppress the multi-cell interference, thus the spectral efficiency of 5% of the cell edge and the overall throughput of the system can be increased by about 2.2 times.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3

【相似文献】