【作者】 韩亮；

【导师】 唐友喜； 邵士海；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2013， 博士

【摘要】 分布式多天线主导的网络架构作为未来移动通信网的显著特征，可以提高系统容量、降低发射功率和扩大网络覆盖。对于分布式多天线系统而言，由于天线分布在不同的地理位置，首先要解决的问题就是如何优选天线的位置。然而，目前对于移动通信网分布式天线位置优选的研究还很不充分。本文针对两种典型的分布式多天线系统——“分布式发射天线系统”和“协同中继系统”展开研究。针对分布式发射天线系统，从误码率角度研究了选择部分天线发射策略下的分布式天线位置优选和室内三维环境中的天线位置优选；针对中继系统，从中断概率角度研究了相关阴影衰落信道下的中继位置优选和协同自动重传请求(ARQ:Automatic Repeat reQuest)中继系统的中继位置优选。具体来说，全文包含以下四个创新点：首先，在半径为R的圆形小区内，针对圆形布局的分布式发射天线系统，推导了下行选择部分天线发射策略下的误符号率，提出了最小化小区平均误符号率的分布式天线位置优选方法，并且以选择1根发射天线和选择2根发射天线两种情况为例对所提方法做了说明。数值和仿真结果表明最优的天线半径与路径损耗因子、总天线数目和选择天线数目有关。特别的，当总天线数为6，选择天线数为2，路径损耗因子为3时最优的分布式天线半径是0.65R。其次，研究了室内三维环境中分布式发射天线空时分组码(STBC: Space TimeBlock Coding)系统的天线位置优选技术。在考虑信号经历路径损耗、阴影衰落和莱斯衰落的情况下，首先推导了用户在室内的平均误符号率，然后通过最小化室内平均误符号率得到了最优的发射天线位置。理论分析和仿真表明优选的发射天线位置与房间的尺寸和移动台的高度有关，当房间的长a、宽b、高c和移动台高度h满足a2b212(c h)2≤0时，分布式发射天线退化为集中式发射天线，即两根发射天线集中放置在屋顶中心；当a2b212(c h)2>0时，两根天线分布在屋顶沿着房间长方向的中轴线上，并且关于屋顶中心对称，距中心距离为(a2-b2)12-(c-h)2。再次，研究了相关阴影信道下放大转发(AF: Amplify-and-Forward)中继系统的功率分配和中继位置优选问题。在考虑路径损耗、小尺度瑞利衰落和相关阴影信道的基础上，首先推导了3点中继系统在高信噪比下的近似中断概率，然后以最小化近似中断概率为目标，分别研究了固定中继位置下的功率分配优化，固定发射功率下的中继位置优化，以及功率分配和中继位置的联合优化。理论分析和仿真表明：阴影衰落的标准差和相关系数会显著影响最优的功率分配和中继位置，因此在实际系统中不能忽略相关阴影衰落的影响；在考虑阴影衰落相关系数与位置有关时最优的中继位置不一定在源节点到目的节点的连线上；相比固定中继位置下优化功率分配和固定发射功率下优化中继位置，联合优化能得到更好的性能。最后，研究了协同ARQ中继系统的功率分配和中继位置优化。首先推导了快衰落瑞利信道下3点解码转发(DF: Decode-and-Forward)ARQ中继系统在高信噪比下的近似中断概率，然后以最小化近似中断概率为目标，分别研究了固定中继位置下的功率分配优化，固定发射功率下的中继位置优化，以及功率分配和中继位置的联合优化。理论分析和仿真表明：当中断概率为10-3时，协同ARQ相对于直接传输有大约4dB的性能增益；最大重传次数和路径损耗因子会影响最优的功率分配和中继位置。论文在分布式天线位置优选方面做出的上述研究成果，从误码率和中断概率角度给出了分布式多天线系统位置优选的方案，可以作为分布式移动通信网天线选址的重要参考。更多还原

【Abstract】 As a significant feature of the future mobile communications networks, distributedmultiple antennas systems can achieve high channel capacity, decrease the transmitpower and enhance the cell coverage. Therefore, distributed multiple antennas systemshave become research hotspot in recent years. For the distributed multiple antennassystems, antennas are placed in different locations, so the first problem to be solved ishow to place antennas to obtain the best performance. However, research in this area isstill insufficient.We focus on two kinds of typical distributed multiple antennas systems: distributedtransmit antenna systems and cooperative relay systems. For the distributed transmitantenna systems, we investigate the antenna location optimization for antenna selectionsystem and indoor distributed space time block coding (STBC) system from theperspective of bit error ratio (BER). For the cooperative relay systems, we investigatethe relay location optimization for amplify-and-forward (AF) cooperative relaying overcorrelated shadowed fading channels and decode-and-forward (DF) automatic repeatrequest (ARQ) relaying over fast Rayleigh fading channels from the perspective ofoutage probability. The details of the investigation are as follows:Firstly, we present the downlink antenna selection strategies and derive the symbolerror ratio for the distributed transmit antennas systems, considering a circular cell withthe radius of R and with antenna circular layout. Then optimal distributed antennaradius can be obtained by minimizing cell averaged symbol error ratio (CASER). Wetake selecting one antenna and selecting two antennas for example to illustrate ourproposed method. Numerical and simulation results show that the optimal antennaradius depends on the path loss exponent, the total number of distributed antennas andthe number of selected antennas. In particular, when the total number of antennas is6,the number of selected antennas is2, and the path loss exponent is3, the optimal radiusof the distributed antenna is0.65R.Secondly, the area averaged symbol error ratio (AASER) for indoor STBC downlinkwith distributed transmit antennas is derived, considering the effects of path loss,shadow fading and Rician fading. Then optimal antenna location can be obtained byminimizing AASER. Theoretical analyses show that the optimal antennas location isrelated to the room length a, width b, height c and the height of the mobile station h.When a2b212(c h)2≤0, the distributed antennas degenerate into the centralizedantennas. When a2b212(c h)2>0, two antennas should be located on the longitudinaldirection of the central axis of the room, and should be symmetrical about the center ofthe roof. Furthermore, the distance from the distributed antenna to the center of the roof Thirdly, we investigate the relay placement and power allocation for AF cooperativerelaying. Considering the joint effects of path loss, correlated shadowing, and flatRayleigh fading, we first derive the approximate outage probability in the highsignal-to-noise ratio (SNR) regime. Then three optimization problems are formulated tominimize the obtained approximate outage probability, namely, optimal relay placementwith fixed power allocation, optimal power allocation with fixed relay location, andjoint optimization of relay placement and power allocation. It is shown by the analyticaland numerical results that the correlation coefficients and the standard deviations ofshadowing have significant impacts on the optimal relay placement and powerallocation. As such, the impact of the correlated shadowing cannot be ignored. Theoptimal relay position may not be on the line connecting the source and the destinationfor certain scenarios. Furthermore, the joint optimization obtains the best outageperformance.Finally, we investigate the power allocation and relay placement for cooperative ARQrelaying over fast Rayleigh fading channels, where the channel gains are constantduring each ARQ transmission round and change independently from round to round.We derive the asymptotically tight approximation of outage probability in the high SNRregime. Furthermore, three optimization problems are formulated to minimize theobtained asymptotic outage probability, namely, optimization of the power allocation atthe source and the relay with fixed relay placement, optimization of the relay placementwith fixed power allocation, and joint optimization of the power allocation and relayplacement. It is shown that the cooperative ARQ relaying has4dB gain at outageprobability of10-3compared to direct transmission, and the maximum number of ARQrounds and the path loss exponent will impact the optimal power allocation and relayplacement.The research results in this dissertation provide feasible schemes for distributedantennas placement from the perspective of system BER and outage probability, whichcan be applied to the future distributed mobile communications networks.更多还原