【作者】 赵滨；

【导师】 关新平；

【作者基本信息】 燕山大学 ， 控制理论与控制工程， 2011， 博士

【摘要】 由于受到电磁波传播特性等限制,目前能够用来进行无线通信的频谱资源非常有限。但随着无线通信应用大量增加,特别是近几年来无线互联网应用的爆发式增长,对无线频谱资源提出了大量的需求。而当前普遍采用固定分配制度,各国的可用无线电频段几乎已经分配殆尽。认知无线电技术能够在避免对授权用户干扰的同时,利用授权用户频谱进行传输。认知无线电技术的提出,为解决应用增长和资源有限的供需矛盾问题,提供了一种有效的解决途径。本论文针对认知无线电网络中的主用户定位和位置信息应用问题,围绕主用户位置信息的获取、利用以及主用户定位误差对从用户网络性能的影响进行了深入的研究。首先,在研究传感器网络定位算法的基础上,根据认知无线电网络的特点,研究了基于接收信号强度的主用户定位算法和节点选择算法。在主用户发射信号强度已知情形下,提出了一种无需迭代的定位算法,并对此算法线性化过程中的模型误差与接收信号强度之间的关系进行了分析。考虑定位算法的几何结构性能,基于估计理论,分析了节点分布的几何结构与估计性能下界之间的关系,进而,基于模型误差和几何结构分析,提出了一种节点分区选择算法。该算法不但能够在降低模型误差的前提下避免较差的几何结构,提高定位的精确性。而且通过节点选择降低了计算的规模,减少了定位算法的计算量,从而降低了节点能量消耗,提高了网络生存时间。其次,针对一类在无线传感器网络定位中广泛使用的,基于多次最小二乘的定位算法的模型误差问题,分析了模型误差与接收信号强度的关系。同时考虑估计性能下界与节点选择数量之间的关系,提出了一种节点选择算法,可以在模型精确性和几何结构性能上获得一个良好的折中,提高了原算法的定位精度,并降低了原算法计算复杂度。进一步研究了基于位置信息并区分信道特征的频谱分配算法。针对认知无线电网络可利用频谱资源间的频率差别较大,导致频率成为影响用户间干扰关系主要因素的问题,提出了根据频率进行信道分类的标准。这种标准更适用于认知无线电网络“频谱稀缺”的环境。在资源分配优化问题中,利用主用户位置信息进行了空间重用,有效地提高了频谱利用效率。最后,研究了主用户定位误差对认知无线电网络性能的影响。考虑一个基于OFDMA技术的认知无线电网络的下行通信场景中,主用户定位误差对资源分配算法以及从用户网络性能的影响。建立了包含主用户位置不确定性的最优资源分配鲁棒优化模型,并将其转换为一个普通优化问题进行求解。基于功率控制的“注水原理”,研究了主用户定位误差对从用户网络性能不同影响程度的条件,以及对从用户网络性能影响的上界。提出了一种可行的判断定位误差对从用户网络影响程度的方法。同时也证明了之前提出的主用户定位和节点选择算法适合在基于位置信息的资源分配问题中应用。更多还原

【Abstract】 The available spectrums are very limited due to the character of electromagnetic wave itself. But along with the increasing applications of wireless communication, especially with the sharp increasing applications on the wireless internet in recent years, large amount of wireless spectrum resource is required. Because of the fixed allocation rule which is currently wide used, the available radio spectrum has been assigned almost completely in every country. Cognitive radio technology can utilize the licensed users’frequency for transmission and avoid interferences on licensed users. This technology provides an effective solution to solve the problem between the increasing wireless applications and the limited resource. In this thesis, we have studied the location information collection and utilization including the primary user localization algorithm, location-based spectrum allocation algorithm, and the impacts of the localization error on the secondary network performance.First, motivated by the sensor network localization algorithms, considering the requirements of localization in cognitive radio networks, we have studied the received signal strength based localization algorithm and the node selection algorithm. A localization algorithm without requirement of iteration is proposed under the condition that the transmitting signal strength is known. The relationship between model error brought in the process of linearization and received signal strength is analyzed. Based on estimation theory, the relationship between the lower bound of estimation performance and geometry structure is analyzed with the consideration of the geometry structure performance in localization. Then, based on the model error and geometry analysis, a node selection algorithm is proposed to reduce the error and avoid the poor geometry. Thus, the localization accuracy is improved. On the other hand, the proposed algorithm greatly reduces the computational scale and the computation amount is largely reduced. Therefore, node energy consumption is reduced and the network lifetime is prolonged.Second, we have studied the localization problem for a class of localization algorithms widely used in wireless sensor networks without primary user’s emission signal strength. The model error is analyzed. Taking into account the relationship between the lower bound of the geometry structure and the estimated lower bound, the relationship between the model error and the geometry structure is studied. A node selection algorithm is proposed to achieve a good compromise between the model accuracy and the geometry performance. Thereby, the localization accuracy is improved and the complexity of the algorithm is reduced.Third, we have studied the location-based spectrum allocation for heterogeneous channels. Considering the distinction between the available spectrums and the difference on the radius of transmission and interference between the spectrum frequencies may be large, a more applicable standard for the "spectrum scarcity" character in cognitive radio networks, which classify channels according to the channel frequency, is proposed. In resource allocation problem, taking into account the spatial reuse, we proposed a location based method and it can improve the efficiency of spectrum utilization.Finally, the impacts of the location error on the cognitive radio network performance have been studied. By considering the downlink transmission in an OFDMA based cognitive radio network, we analyze the performance changes of resource allocation and secondary user networks when there exists the location error on the primary use. We focus on the effects of localization error for interference radius. Based on the "water filling" theory in power allocation problems, the upper bound on the impacts of localization error on secondary users are derived under the condition that location error interval is known. The conditions are proposed for the assessment of effects on the secondary uses networks. A feasible method to judge the impact degree of location error is given based on our proposed conditions. It is also proved that our localization and node selection method are applicable in spectrum spatial reusing.更多还原