【作者】 罗海燕；

【导师】 张朝阳； 李式巨；

【作者基本信息】 浙江大学 ， 通信与信息系统， 2010， 博士

【摘要】 随着无线通信技术的蓬勃发展,日益增长的无线通信需求与有限的无线频谱资源之间的矛盾越来越明显。在授权频谱利用率低下和非授权频谱趋于饱和的矛盾下,认知无线电技术应运而生,通过大力发掘授权频谱的复用潜力,有效地改善了当前固定频谱分配方式的不合理性。基于认知无线电的动态频谱接入机制,能够大大提高频谱资源的利用率,在一定程度上缓解了频谱资源的紧张程度。而基于认知无线电的频谱共享机制,则可以实现异构无线网络和谐共存的目标,在满足各种业务需求的同时为用户提供更好的体验。本学位论文首先分别针对支持非时延敏感业务和时延敏感业务的认知无线电系统,就认知无线传输链路的优化控制进行了分析和研究。接着创新性地提出了基于认知中继协作的频谱共享机制,并深入研究了其中的资源联合优化问题。主要从以下几个方面展开深入研究,概括如下：研究了针对非时延敏感业务的认知无线传输链路的优化控制问题。在认知OFDM系统中,在主用户即时信息速率不变的约束下,认知用户发送端必须拿出一部分功率合作传输主用户信号以消除自身传输对主用户链路的干扰,导致每个子载波上分配到的功率和实际获得的比特速率不一致。在离散比特约束下,给出了一种功率和比特的转化方法,从而将功率和比特联合优化分配问题简化为单变量优化问题,在此基础上提出了贪婪比特优化分配算法,并证明了其最优性。对于主用户自适应调整其调制编码方式的认知无线电系统,在训练序列阶段,根据主用户传输所允许的最小平均频谱效率确定认知用户在该阶段采用的最佳发射功率。在数据传输阶段,根据认知用户所能获得的主用户链路即时信息(如调制编码方式,接收信干噪比等)的程度,在不违反主链路和认知链路要求的误包率约束下,分别提出了三种基于AMC技术的自适应链路优化控制策略,尽力提高认知链路在单位频段上的平均频谱效率。如当主用户链路即时信道状态信息未知时,给出了认知用户在固定发射功率下,用来确定AMC模式的SINR阈值的最佳选择。当主用户链路即时AMC模式已知时,联合考虑低灵活度的功率优化控制和随机接入策略。而当主用户链路即时接收SINR已知时,则可以采用高灵活度的功率优化控制和确定性接入策略。研究了针对时延敏感业务的认知无线传输链路的优化控制问题。针对ON/OFF两状态授权信道模型,认知用户的数据包到达率受到两方面的约束：时延QoS保障下的有效容量约束和来自主用户方的传输冲突概率约束。分别提出了η随机接入策略和感知参数调整策略以缓解上述两个约束不匹配导致的频谱资源非有效利用问题,同时在满足时延QoS约束下尽可能地提高认知用户的数据包到达率。针对瑞利衰落信道模型,在主用户发射功率固定的情况下,将主用户要求的传输中断概率约束放松为平均干扰约束。而在主用户自适应地根据灌水法则调整发射功率的情况下,相应地将传输中断概率约束放松为即时干扰约束。接着分别就以上两种情况,联合平均发射功率约束,在凸优化理论的指导下,研究了如何实时调整发射功率以最大化认知用户的有效容量,并分析了QoS指数对功率优化控制策略以及认知用户吞吐量的影响。研究了基于认知中继协作的频谱共享机制中的资源联合优化问题。首先创新性地提出了基于认知中继协作的频谱共享机制,使得非认知的从用户系统在不需要更新设备的情况下可以实现和主用户系统共享频谱的目标。该频谱共享机制充分结合了系统内合作频谱共享和系统间合作频谱共享的优势,使得认知中继在为从用户创造频谱接入机会的同时,有效减少了对主用户系统的干扰并提高了从用户系统的吞吐量。根据认知中继协作对象的不同,分别给出了该机制下三种可能存在的传输模式。为了实现最小化主用户传输中断概率以及最大化从用户平均传输速率的双赢目标,提出了基于模式切换的机会式认知中继策略,并研究了每种模式下从用户和认知中继的功率联合优化问题。当直传链路不存在时,考虑了更为复杂的认知中继同时协作主从系统传输的问题。认知中继首先侦听来自主用户和从用户发送端的信号,接着分别进行译码,将主从信号编码再生后转发给对应的接收端。当认知中继分别采用逐次译码和叠加编码来处理接收信号和发送再生信号时,研究了从用户发送端的功率控制以及认知中继端的逐次译码顺序的联合优化控制,并在叠加编码时考虑了功率优化分配。为了获得更好的性能,提出认知中继应该联合从用户发送端采用机会式干扰消除的方式,创造更多对主用户信号译码的机会,同时提高从用户的传输速率。此外,在发送再生信号时,认知中继应该结合脏纸编码和叠加编码,使得从用户接收端可以获得无干扰的传输速率。更多还原

【Abstract】 With the development of cognitive radio technology, the current state of spectrum scarcity and spectrum under-utilization can be greatly alleviated, by allowing the unlicensed users to opportunistically access to the licensed spectrum without interrupting the transmissions of the licensed users.Thanks to the cognitive radio technology, dynamic spectrum access is proposed to improve the spectrum utilization, in which optimal link control of the unlicensed cognitive user is extremely important. In addition, cognitive radio based spectrum sharing is able to re-alize heterogeneous network coexisting, while incorporating with cooperative relay may further enhance the system performance. In this dissertation, the optimal link control and cooperative relay mechanism in wireless cognitive networks are intensively investigated, and the contents of this work are listed as follows:For delay insensitive service, the optimal link control mainly focuses on how to reduce the interference introduced to the primary link and how to combat the fading channel of the secondary link. In the context of cognitive OFDM system, we investigate the bit and power allocation problem while guaranteeing the primary link against information rate loss. We find the conversion relationship between the unmatched transmit power and bit rate, and then propose a greedy bit allocation algorithm and prove its optimality. In the context of the cognitive radio network where both the primary and secondary users can operate with Adaptive Modulation and Coding (AMC) technology, we aim to maximize the average spectral efficiency of secondary link under the constraint of average spectrum efficiency and packet error rate of the primary link. For different cases of availability of primary link information to the secondary user, i.e.,1)it has no any primary link information,2) it knows the transmission mode of primary transmitter, and 3) it knows the SINR at primary receiver, three AMC-based adaptive link control policies for the secondary user are proposed.For delay sensitive service, the optimal link control in wireless cognitive networks becomes even more challenging.Considering an ON/OFF primary channel, the arrival rate of secondary user with delay related QoS requirement, is limited by both the effective capacity provided by primary channel,and the packet collision probability constraint of primary link. Depending on whether the sensing parameters could be adjusted or not, two spectrum utilization approaches, i.e.,η-probability random access and sensing parameter adjustment are proposed to fully utilize the transmission opportunities.Considering the Rayleigh fading channel,we investigate the optimal power control to improve the effective capacity of the secondary user. When the primary user adopts fixed transmit power, the restricted transmission outage probability is relaxed to an average interference constraint for the secondary user. On the contrary, if the primary user adaptively adjusts its power according to the water-filling policy, then the transmission outage probability is relaxed to an instantaneous interference constraint. Finally, we find the optimal power control policies for both cases based on convex optimization, and investigate the impact of QoS exponent on them.Finally, a cognitive relay assisted spectrum sharing architecture is proposed to enable the legacy non-cognitive secondary system to coexist with the primary system. Based on that, an opportunistic cognitive relaying (OCR) scheme is specially devised, in which the cognitive re-lay opportunistically switches among three different working modes, i.e.,relay for primary link (RPL),relay for secondary link (RSL) or relay for neither of the links (RNL), respectively, based on the channel-dependent observation of both systems. In addition, the transmit power for cog-nitive relay and secondary transmitter in each mode are optimally determined by maximizing the transmission rate of secondary system while keeping or even reducing the outage probability of primary system. In the case of no direct links, the cognitive relay has to simultaneously assisting the transmissions over both primary and secondary links.Two decode-and-forward schemes with different levels of complexity, i.e.,SD-SC (successive decoding then superposition coding) and OIC-DPC(opportunistic interference cancellation then dirty paper coding) are presented respec-tively. In order to maximize the transmission rate of secondary link while providing adequate primary link protection, the transmit power for the secondary transmitter and the decoding order for the relay, as well as the power division for the regenerated signals are optimally determined in both schemes.更多还原