【作者】 周娟；

【导师】 唐友喜； 沈莹；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2013， 博士

【摘要】 认知无线电通过动态频谱共享，提高频谱资源利用率，可解决当前很多国家遇到的频谱稀缺问题，是未来无线通信最有前途的技术之一。然而，传统认知无线电获取无线场景信息的能力有限。以频谱“侦听”为代表的认知无线电“侦听式”技术，被动监听无线电磁环境，感知频谱占用信息。这种被动监听缺少全景多维信息。通过“听”方式认知的参数局限在信号射频特征上，无法掌握用户周围的建筑楼群密度、用户密集程度和用户移动趋势等无线通信环境。视觉认知无线电可以解决上述问题，得到更多环境信息，推动无线通信系统环境自适应能力的全面提升。视觉认知研究基于视觉图意的移动无线网络认知理论及关键技术。“视觉图意”认知获取无线信道环境的图像，进行场景分析、资源决策等认知过程，解析出无线场景的通信资源信息，即图的含意“图意”，优化无线网络系统配置，达到提升网络整体性能的效果。位置是视觉认知的核心成分之一，本文围绕次用户感知天线位置优化、次用户基站位置优化与位置辅助协同频谱感知展开研究。首先，研究了认知无线电次用户感知天线的位置优化问题：(1)推导单天线下，复合瑞利-阴影信道的频谱感知性能闭合表达式。分析路径损耗和主用户位置布局对感知性能的影响，以最小化感知错误概率为准则，求得次用户感知天线的最优位置。(2)采用与次用户单天线最优位置相似的研究步骤，对次用户多天线的最优位置进行了研究。(3)通过理论分析发现，次用户单天线的最优位置位于小区中心；次用户多天线的最优位置关于小区中心对称分布，天线之间的距离由信道质量和小区大小决定，可应用凸优化理论计算出最优位置。理论和仿真结果表明，能量检测门限T=26，能量检测采样间隔N=10，小区半径R=100m，噪声方差σ~2=100dBm，路径损耗指数η=4，阴影标准差δ=6dB，阴影系数α=0.12时，两根天线之间的最优距离为33.6m。其次，针对采用并发接入与主用户网络实现频谱共享的广播次用户网络，研究次用户基站的最优位置问题：(1)考虑并发接入频谱共享下主用户的干扰约束以及次用户的服务质量约束，计算出次用户网络覆盖范围。以次用户网络覆盖范围最大化为准则，推导出次用户基站的最优位置。(2)将上述研究扩展到障碍存在于主用户和次用户基站之间的场景，推导了障碍存在场景下的次用户基站最优位置。(3)通过理论分析发现，随着主用户与障碍之间的距离增加，次用户基站最优位置与主用户之间的距离将不断增加，直至收敛到障碍不存在情况下的最优位置。理论与仿真结果表明，无障碍存在场景下，主用户和次用户基站天线高度h0=20m，主用户网络小区半径r=10,000m，主用户最大发射功率Pp=30dBm，次用户基站最大发射功率Ps=10dBm，主用户干扰门限I th=40dBm，次用户信干比门限th=10dB，求得最优次用户基站位置为8,200m。在障碍距离主用户为300m时，最优位置为600m，当障碍距离主用户大于800m后，最优位置收敛到8,200m。最后，考虑到位置优化在一些情况下分析困难或无解，从另一个角度即位置辅助应用进行研究：(1)提出了位置辅助的协作频谱感知方法。本地采用能量检测得到一个硬判决结果，并将判决结果传递到数据融合中心，数据融合中心利用位置信息对本地感知结果加权合并，得到全局判决结果。(2)对给出方法进行性能分析。根据本地频谱感知可靠性和判决结果传输可靠性，推导了位置辅助频谱感知的全局性能，即全局检测概率和全局虚警概率的精确计算闭合表达式；由于全局性能精确计算复杂，提出高斯近似方法计算全局性能；以全局感知错误概率最小化为准则，推导出本地检测最优门限值。理论和仿真结果表明当平均本地信噪比分别为12dB、8dB、4dB时，位置辅助协作频谱感知方法的全局错误概率为0.103，而传统方法的全局错误概率为0.447。(3)位置信息估计不可避免的会存在误差，从而对提出方法进行了误差分析，给出了误差衡量准则。论文工作为视觉认知无线电技术的深入研究提供了理论和方法支撑，具有较好的理论和实用价值，位置优化可应用于无线通信系统的规划布局，位置辅助应用可改善无线通信系统性能。更多还原

【Abstract】 Cognitive Radio (CR) can improve the spectrum utilization efficiency through theway of dynamic spectrum share, which addresses the spectrum scarcity problem that isencountered in many countries. CR is widely regarded as one of the most promisingtechnologies for future wireless communications.However, the ability of wireless scenario acquisition is limited in traditionalcognitive radio. The cognitive radio “monitoring” technique is based on spectrumsensing, which is always sensing the spectrum usage by monitoring the radioelectromagnetic environment passively. It is lack of the panoramic andmultidimensional radio environment. The parameters cognized by “monitoring” arelimited on the signal characters. The buildings around the users, the intensity of theusers, and the mobile trend of the users are not available. In order to obtain moreenvironment information for improving the environmental adaptability of the wirelesscommunication system, visual cognitive radio is proposed.Visual cognitive radio studies the cognitive theory and key technologies of themobile wireless cognitive radio network based on the visual image meaning.“Visualimage meaning” cognition captures the image of the wireless channel environment,cognizes the image by wireless scenario analysis and resource decision, recognizes thecommunication resource information about the wireless scenario, namely the meaningof the image--“image meaning”, optimzes the configuration of the wireless networksystem, and so as to improve the network performance.Location is one of the most important components in visual cognitive radio. In thispaper, two key technologies around the location are studied, which are locationoptimization and location appilication.Firstly, the optimal location of the sensing antenna for the sencondary user (SU) incognitive radio is studied.(1) A closed-form expression of the spectrum sensing performance is derived in the composite Rayleigh and shadowed fading environment for the SU with onesensing antenna. The effects on the performance from the pathloss and thedistribution of primary users (PUs) are further analyzed. The optimal locationsof the sensing antennas are derived so as to minimize the spectrum sensingerror probability.(2) Following the similar procedures, the locations optimization is studied for twoantennas.(3) The theory analysis shows that, the optimal loation is the cell center for oneantenna. As for two antennas, they should be placed symmetrically. Thedistance between the two antennas depends on the channel condition and thecell range, which can be determined by the application of the convexoptimization theory. Both theory analysis and simulation results show that, theoptimal distance between two antennas is33.6m if the energy detectorthreshold, the interval of the energy detector, the cell radius, the noise variance,the pathloss exponent, the standard deviation of shadow, and the shadowcoefficience are26,10,100m,－100dBm,4,6dB, respectively.Secondly, the optimal location of the secondary base station (SBS) is studied forbroadcasting cognitive radio networks with spectrum underlay.(1) Considering the interference constraints for the PU and the quality of service(QoS) constaints for the SU, the coverage area of the SU network is calculated.The optimal SBS location is derived so as to maximize the coverage area ofthe SU network.(2) The analysis is extended to the scenario with an obstacle located between thePU and SBS, the optimal SBS location is discussed in this scene.(3) The theory analysis shows that, the optimal location of the SBS will be largerand will converge to a fixed point with the increase of the distance between thePU and the obstacle. This fixed point is the optimal location of the SBSwithout the obstacle. Both theory analysis and simulation results show that, theoptimal location of the SBS is8,200m without obstacle, under the conditionthat the antenna heights of the PU and SBS are20m, and the cell radius of the PU network is10,000m, the maximum of the PU transmit power is30dBm,the maximum of the SBS transmit power is10dBm, the interference thresholdof the PU is40dBm, and the receive SIR threshold of the SU is10dB. Theoptimal SBS location is600m if the distance between the PU and the obstacleis300m. The optimal SBS location will converge to8200m if the distancebetween the PU and the obstacle is larger than800m.At last, location usage is studied, since the location optimization in most cases isdifficulty for analysis or can not be solved out.(1) A location-based cooperation spectrum sensing scheme is proposed. Energydetection is performed to obtain a local hard decision, which is thentransmitted to the data fusion center. Combining the weighted local decisionswith location information, a global decision is obtained at the data fusioncenter.(2) The performance on the proposed scheme is analyzed. The global performance,namely the global detection probability and the global false alarm probability,are derived based on the local sensing reliability and the transmissionreliability; The Gaussian approximation method is introduced to calculate theglobal performance, since the exact calculation is too complicated. When theaverage local SNRs are12dB,8dB, and4dB, the global error probability inthe location-based cooperation spectrum sensing scheme proposed is0.103,while it is0.447in the conventional scheme.(3) The location estimation error is inevitable, therefore, the error analysis of theproposed scheme is given and the error measurement rule is described.The work in this paper provides the theory and method for the deep study in visualcognitive radio. It is full of theory and practical value. The location optimization can beused for the plan of wireless communication system. The location application can beintroduced to improve the performance of wireless communication system.更多还原