【作者】 单杭冠；

【导师】 王宗欣；

【作者基本信息】 复旦大学 ， 电路与系统， 2009， 博士

【摘要】 协作传输作为一种新型的传输技术，让用户间共享天线，构成虚拟多天线系统，在不增加硬件复杂度的情况下获得了空间分集增益，有效地克服了多径传播导致的信号衰落，实现了提高传输性能的目的，因而受到了广泛的关注和重视。本文从物理层的资源分配和上层的协议设计两个角度分析和研究了协作在通信场景中的性能增益。首先，我们针对基于CRC检验的选择解码转发(CDF，CRC-based selectiondecode-and-forward)和放大转发(AF，ampliy-and-forward)两种协作协议，分析了它们的中断概率，指出编码、调制对于以上两种协议的不同影响；并在考虑QoS(Quality of service)要求的前提下，以最大化协作网络的传输速率为目标，联合基于最小化中断概率的功率分配和自适应编码调制等技术，提出了一种扩展的链路自适应传输技术，研究表明：该技术可以有效地提高系统的传输速率和扩大系统的覆盖范围。其次，我们把协作传输与多载波调制技术相结合，针对放大转发型多载波协作传输系统，分别以最小化传输错误概率和最大化传输速率为目标，联合子载波和功率分配以及自适应编码调制等技术，提出了最优和次优的资源分配算法。研究表明：以最小化传输错误概率为目标所提出的资源分配算法不仅可以有效地降低传输误码率，还可以提高系统的分集阶数，并且当系统已知节点的位置信息时，只要通过设定次优资源分配算法的复杂度参数，就可以在获得最佳系统性能的同时降低计算复杂度；而以最大化传输速率为目标所提出的资源分配算法具有低实现复杂度和可靠提高传输速率等优点。然后，针对分布式无线通信系统，把物理层的协作传输与MAC(mediumaccess control)层的协议设计相结合，研究了如何通过跨层设计来提高分布式协作传输系统的吞吐量和保证更加可靠的通信质量。我们先从MAC的角度分析和研究了协作在分布式网络中所会遇到的各种困难和挑战。然后，针对全连通网络，侧重于协作节点的选择问题，基于跨层的思想提出了一种802．11 DCF(Distributed coordination function)构架的协作MAC协议。再在此基础上，研究了多跳网络的协作MAC设计：首先，在理论上提出了一种两跳的干扰模型，从空间频率复用的角度分析了协作对于这类网络的影响，通过分析我们发现协作传输对于使用协作的链路本身具有抑制干扰的好处，但是也会造成被阻塞的节点个数增加而可能导致网络整体性能的下降；然后，为了证实协作传输对于该类网络可能带来的影响，提出了一种适用于多跳网络的基于Busy-tone的跨层协作MAC协议，为了有效地解决由协作带来的新的隐蔽站问题，该协议中加入了目的节点辅助的无冲突竞争机制，同时为了进一步提高协作节点的竞争效率，我们也提出使用分组避退的竞争方法，最后就节点密度和网络流量等因素对于多跳协作网络的性能影响进行了仿真。最后，我们把协作扩展到应用层，以无线传感器网络WSN(Wireless sensornetwork)的定位问题为例，研究了节点间的协作对于无线传感器网络定位的帮助。提出了一种利用节点间协作的，适用于基于测距的分布式定位方法——3/2-NANDB，它可以充分利用节点间的冗余信息，完全确定只有两个邻居节点的节点位置，而那些即使只有一个邻居节点的待估节点也可能被完全定位，或者最大可能性地估计出它们的位置，从而可以最大化网络内部可定位节点数目，扩大网络观察范围，减少GPS携带节点数量，延长网络存活时间，同时该方法也可以用来提高其他现有的基于三邻居节点定位的算法的总体性能。更多还原

【Abstract】 Cooperative communication is a promising diversity technique that has attractedsignificant research interests,where mobile nodes share their information andtransmit cooperatively,forming a virtual antenna array and thus providing diversityto combat signal fading due to multipath propagation without requirement ofadditional antennas at each node.In this dissertation,we investigate the problem that how to improve theperformance of cooperative communication under various scenarios by resourceallocation on physical layer and higher layers’ protocols design.Firstly,we analyze the outage probability of two cooperative protocols,CRC-based selection decode-and-forward (CDF) and amplify-and-forward (AF),andillustrate the different impacts they suffer from modulation and coding schemes.Furthermore,considering quality of service (QoS) requirement,an extended linkadaptation (ELA) technique is proposed for transmission rate maximization.Outageprobability optimization based power allocation and adaptive modulation and codingtechnique are jointly designed to improve its performance.Simulation results showthat with the ELA,the two merits of cooperation,improved transmission rate andincreased coverage,can be further effectively enhanced.Secondly,cooperative communication is integrated with the multicarrier systems.For the AF-based multicarrier cooperative system,both the optimal and suboptimalresource allocation algorithms are proposed for bit error rate (BER) minimizationand rate maximization,respectively,utilizing joint power allocation,subcarrierreassignment design and adaptive modulation and coding techniques.Simulationresults show that the proposed BER-minimization algorithms not only reduce BEReffectively,but also improve the diversity order.In addition,with the locationinformation knowledge,the optimal BER performance can be achieved just bysetting an appropriate complexity coefficient of the suboptimal algorithm.Simulation results of the rate-maximization algorithm verify its advantage of efficient date rate improvement under very low implementation complexity.Thirdly,for the distributed wireless communication systems,cross-layeroptimization between cooperative communication on physical layer and protocoldesign on medium access control (MAC) layer is considered.We first discussvarious issues and challenges on designing an efficient cooperative MAC.Then,focusing on helper selection,a cooperative MAC based on IEEE 802.11 DCF isproposed under the scenario of single hop networks.After that,we extend our workto multihop network scenario and propose a two-hop interference model from aMAC perspective to analyze the performance of a cooperative network in term ofspatial and frequency reuse.It is found that with cooperation,a cooperative linkbenefits from less interference due to more suppressed interference sources,however,the increased number of blocked traffic nodes can degrade the network performancedue to a lower spatial frequency reuse.To verify the influence of cooperativecommunication on such networks,a cross-layer designed busy-tone basedcooperative MAC scheme is proposed.Furthermore,destination-assistedcollision-free piggyback mechanism and group-based helper contention are proposedin the new MAC scheme to solve the new hidden terminal problem caused bycooperation and to further improve the efficiency of helper selection,respectively.Simulation of the new cooperative MAC scheme is done to investigate theinfluenced factors including node density and traffic load.Last but not least,cooperation is further extended to application layer.Takinglocation problem as an example,we investigate the problem that how to make use ofcooperation between neighbor nodes to help to solve location in wireless sensornetworks (WSNs).A distributed location method called 3/2-NANDB based ondistance measurement is proposed according to the principle of minimal hardwarecost.Due to the information sharing between neighbors,3/2-NANDB cansufficiently mine redundant information among nodes,and effectively eliminate thedilemma of node’s position.It ascertains the positions of the nodes having twoneighborhoods completely and a part of nodes which only have one neighborhood as well.It achieves the goals of reducing the number of nodes equipped with GPS,maximizing the number of localizable nodes in networks,extending the networks’observation area and prolonging the lifetime of WSNs.Using the node’s dilemmaelimination algorithm,it also can assist other algorithms which are based on 3-NAlocation algorithm to improve their location performance.更多还原