【作者】 郑长亮；

【导师】 周正；

【作者基本信息】 北京邮电大学 ， 电路与系统， 2013， 博士

【摘要】 认知无线电的动态频谱接入能够大幅度提高频谱的使用效率,因而受到人们的普遍关注,认知无线网络的路由及传输是其中的重要关键技术。本文选题来源于国家自然科学基金等项目,具有重要的理论意义和广阔的应用前景。本文在深入研究认知无线网络路由算法及传输方案基本原理的基础上,主要完成了以下具有创新性的研究成果：针对动态频谱接入快速变化的环境,本文提出了一种多信道频谱意识(spectrum aware)机会路由算法,即不依赖于任何事先确定的路由,仅由接收数据的候选转发节点动态决定数据转发；主要利用多信道的优势,有效降低链路时延及端到端的时延；并利用无线信道的广播特性,通过侦听加速数据包的传输；此外,还采用网络编码控制数据包重复问题以增加网络吞吐量。仿真结果表明,该算法有效降低了网络的端到端时延,显著提高了最大支撑网络负载,并能补偿信道可用概率降低导致的性能损失。针对信道衰落单跳认知无线网络环境,本文提出了一种基于随机线性编码的多信道batch(分批,一组数据包)传输方案,即基于编码的一种数据重传替代传输方案,但它比基于多信道自动请求重传(ARQ)的传输方案效率更高；这是因为,基于编码的方案能够模糊数据包信息,使得各个编码数据包的信息没有差别,从而能够减少多个可用信道上的数据包重复问题；此外,其所需要的ACK确认消息较少,可使其对反馈信道的依赖性更低。基于机会路由策略和网络编码,本文还提出了一种两跳认知无线网络的多信道batch传输方案,即同时利用数据包转发和数据包侦听进行传输,并在多个动态可用信道上广播不同的数据包；使用一个简化的两阶段传输模型,本文推导了此两跳batch传输方案的batch时延。分析和仿真结果表明,本文的传输方案能够显著减少batch传输时延,并有助于更好地理解和设计认知无线网络中的机会路由。针对频谱变化较慢的单收发机认知无线网络,本文提出了一种基于最大流片段(maximum flow-segment)的信道分配和路由算法；给出了最大流片段的定义及其构建过程,提出了一种信道分配方法,即通过选择能够连接路由路径上最多数目节点的可用信道,最小化信道切换次数,并通过使用一个简化的信道分发算法,进一步整合到按需路由算法之中。仿真结果表明,本文方案可显著降低端到端时延,而且在网络负载较重的条件下具有较高、稳定的网络吞吐量。本文最后对所做工作进行了总结,并对认知无线网络路由设计的后续工作进行了展望。更多还原

【Abstract】 Cognitive radio networks (CRNs), enabling dynamic spectrum access, can improve spectrum utilization significantly and thus attract many researchers’interests. Routing and transmission are key technologies in CRNs. The topic of this thesis comes from the National Natural Science Foundation of China and the topic plays an important role in both theory and practice.This thesis makes the following innovative contributions, based on the comprehensive study of the principles of routing algorithms and transmission schemes in CRNs.For the CRNs with fast varying spectrum availability, we propose a multi-channel spectrum aware opportunistic routing algorithm. The proposed scheme does not depend on pre-setup routes and it fully exploits the benefits of multiple channels. Moreover both packets forwarding and overhearing are used. Also network coding is exploited to enhance the network throughput. Simulation results show that the proposed scheme delivers good end to end performance in CRNs.We propose a random linear coded scheme for batch transmission in single hop CRNs with lossy channels, which can be considered as an alternative to retransmission. The proposed scheme can blur packet information and make the transmission more efficient than that in ARQ based schemes. The batch delay is derived theoretically. The analysis is further validated by simulations. The coded scheme requires less ACK message and thus it is less dependent on feedback channels.Combining opportunistic routing strategy and network coding, we further design and analyze the batch transmission in two hop CRNs. The scheme exploits both packet forwarding and overhearing. The scheme broadcasts different packets over multiple available channels. We derive the batch delay. Simulation results show the good performance of our scheme. The analysis here is helpful to protocol design in CRNs.For single transceiver CRNs with quasi-static spectrum, we propose a maximum flow-segment (MFS) based channel assignment and routing algorithm. We introduce the definition of MFS and the procedure to build MFS. The MFS based channel assignment minimizes the number of channel switches and it is integrated into the routing algorithm with an efficient channel information dissemination method. Simulation results show that the MFS scheme can reduce end to end delay and achieve high and stable throughput.更多还原