【作者】 韩宇辉；

【导师】 张晔； 孟维晓；

【作者基本信息】 哈尔滨工业大学 ， 信息与通信工程， 2012， 博士

【摘要】 在无线通信系统中，多径传播所引起的信号衰落严重影响了传输的质量，并极大限制了系统的容量。多天线分集技术是一种可以有效对抗衰落的方法，然而由于体积、功耗以及硬件复杂性等方面的限制，一些通信系统的移动终端很难配置多个天线。在这种背景下，协作分集技术被提出并得到了日益广泛的关注。协作分集在一些文献中也称为协作式MIMO，是一种新的空域分集方式，使具有单天线的移动终端通过共享彼此的天线也可以获得分集的效果。由于移动终端间的距离通常远远大于数倍的波长，信道间的相关性大大降低，从而可以使信道容量最大化。然而，移动终端的多天线化是发展的必然趋势，在多天线终端的无线网络中仍然存在协作问题，具有多天线终端的协作式MIMO可使系统性能得到进一步的提升。本文围绕采用解码转发方式的多天终端协作式MIMO，对其协作方法、功率分配、协作伙伴选择等问题进行了研究。研究了多天线终端协作式MIMO的协作方法，提出了两种适用于不同情况的自适应协作算法。在对采用解码转发方式的多天线终端协作式MIMO误比特率性能进行详细分析的基础上，给出了协作方式误比特率性能优于非协作方式时各个链路瞬时信噪比所需要满足的条件。对于网络中通信量较小，空闲信道资源较多的情况，可以采用最小化误比特率的自适应协作算法，根据各个链路的实时信噪比选择直接传输模式或解码转发协作模式，使系统的误比特率最小，同时消耗的能量和占用的信道资源相对于一般的协作方式也较少。对于网络中通信量较大，空闲信道资源较少的情况，可以采用最小化占用信道资源的自适应协作算法，在满足一定误比特率要求的前提下，优先选择直接传输模式，从而大大减少了消耗的能量和占用的信道资源，并降低了系统的中断概率。对多天线终端协作式MIMO的功率分配方法进行了研究，提出了基于黄金分割迭代的功率优化分配算法以及两种简化的功率分配算法。这几种算法将最小化能量消耗作为优化目标，满足系统一定误比特率要求的前提下，根据各个链路的实时信噪比，在源节点与中继节点间动态地进行功率分配，使系统总的发射功率最小，即能量增益最大。基于黄金分割迭代的功率优化分配算法可以通过迭代使包含功率分配系数最优解的区间逐渐缩小，从而可以获得满足一定精度要求的最优解；而两种简化功率分配算法分别通过设置协作门限、缩小初始区间以及直接利用最优功率分配系数近似值公式等方法，使计算的复杂度大大降低，也可以获得与基于黄金分割迭代的功率优化分配算法相近的能量增益。对多天线终端协作式MIMO系统的协作伙伴选择方法进行了研究。在比较经典的最差链路优先算法的基础上进行改进，提出了三种适用于不同条件的伙伴选择算法：最小化误比特率的伙伴选择算法、等功率分配的最大化能量增益伙伴选择算法以及结合功率分配的伙伴选择算法。这三种算法根据对系统误比特率及能量增益性能的分析结果，通过设置协作门限以及通过简单计算获得能量增益或功率分配因子估算值，从而大大降低了计算的复杂度。仿真结果表明本文提出的三种协作伙伴选择算法均可获较高的性能，同时计算复杂度远远小于一般的最差链路优先算法。更多还原

【Abstract】 In a wireless communication system, the presence of signal fading arising frommultipath propagation degrades the system performance and severely limits the systemcapacity. Multiple-antenna diversity technology is an effective method that combatsfading. However, it is difficult for many systems to equip a mobile terminal withmultiple-antenna due to size, power or hardware complexity. In this circumstance,cooperative diversity technology was proposed and attracted increasing attention.Cooperative diversity is also called cooperative MIMO in some literatures, it is a newform of spatial diversity, which is a method to allow a single-antenna mobile to shareother mobiles’ antennas and thus achieve transmit diversity. Due to the long distanceamong the terminals, much larger than several times of wavelength, the channels haveless correlation with each other and hence maximize the channel capacity. The use ofmultiple-antenna terminals, however, has been identified as a certain developmenttendency. In wireless networks with multiple-antenna terminals, the problem ofcooperation still exits, cooperative MIMO with multiple-antenna terminals can improvethe system performance further. This dissertation focuses on decode-and-forwardcooperative MIMO with multiple-antenna terminals. Cooperation strategies, powerallocation algorithms and partner selection algorithms have been studied.The cooperation strategies of cooperative MIMO with multiple-antenna terminals arestudied and two adaptive algorithms are proposed. Based on the detailed bit error rateanalysis of decode-and-forward cooperative MIMO, the conditions of cooperation’s biterror rate performance being better than non-cooperation that the instantaneous signal tonoise ratios of different links should satisfied are given. For networks with light trafficand much unoccupied channel resource, adaptive algorithm with minimized bit error ratecan select direct transmission mode or decode-and-forward cooperation mode accordingto the instantaneous signal to noise ratios of different links, minimizes the system biterror rate, the energy consumption and channel resource occupancy are also less thantraditional cooperative diversity. For networks with heavy traffic and little unoccupiedchannel resource, adaptive algorithm with minimized channel resource occupancy givesdirect transmission mode priority on the premise of bit error rate request can be satisfied.In this way, the energy consumption and channel resource occupancy can be decreasedgreatly and the outage probability is lower as well.Power allocation algorithms of cooperative MIMO with multiple-antenna terminals arestudied, an optimal power allocation algorithm based on golden section iterative methodand two simplified algorithms are proposed. Minimizing energy consumption is taken asthe optimization objective in these algorithms, transmission power are allocated dynamically among the source and the relay according to instantaneous signal to noiseratios of different links, minimizing the total transmission power and maximizing theenergy gain on the premise of bit error rate request can be satisfied. The optimal powerallocation algorithm based on golden section iterative method can reduce the interval thatcontains the optimal solution gradually through iterations and thus get the optimalsolution that satisfies some accuracy demand. By setting cooperation thresholds,reducing initial interval or directly using the formula of the optimal power allocationcoefficient approximation, the two simplified algorithms can decrease the computationalcomplexity greatly and get similar energy gain as the optimal power allocation algorithmbased on golden section iterative method.Partner selection methods for cooperative MIMO with multiple-antennaterminals are studied. Three improved worst link first partner selection algorithms areproposed for different scenarios: partner selection algorithm with minimized bit errorrate, partner selection algorithm with equal power allocation and maximized energygain and combined power allocation and partner selection algorithm. According tothe analysis of system bit error rate and energy gain, these three algorithms decreasedcomputational complexity greatly through setting cooperative thresholds and usingestimated energy gains or power allocation factors. Simulation results show thatthese three algorithms can get high performance, and the computational complexitiesare much lower than ordinary worst link first algorithm.更多还原