【作者】 苏曦；

【导师】 沈树群；

【作者基本信息】 北京邮电大学 ， 光学工程， 2010， 博士

【摘要】 认知无线电技术通过对频谱资源的二次利用实现了授权系统与认知系统之间的频谱共享,从而有效提高了频谱利用率,缓解了目前无线通信系统中频谱资源短缺的现象。因此,认知无线电技术作为提高频谱利用率的最佳方案而受到广泛的重视。然而,如何实现认知无线电系统对已授权频谱资源的有效利用是亟待解决的重要问题。首先,传统的动态频谱分配机制侧重于避免认知网络之间干扰的研究,虽然能够实现认知网络之间的频谱共享,但却无法保证认知系统的可靠性；其次,传统的频谱接入机制研究在频谱检测不理想时及在干扰温度限制条件下的接入机制,旨在避免对授权系统的干扰,但却忽略了对认知系统的频谱接入机会及正常通信的保障；最后,认知系统中频谱资源具有多维特征,但传统的资源优化机制仅利用频谱资源的单个特征实现频谱资源与业务的匹配,然而联合考虑多维特征才能实现频谱资源与业务的最佳匹配。因此,本文以实现认知系统对频谱资源的有效利用为研究目标,研究了上述三种机制,取得了如下的创新性成果。本文在动态频谱分配机制方面的创新成果如下：针对认知网络中频谱资源的不稳定性而引起的掉话率现象,本文首次提出了基于风险分担机制的动态频谱分配机制,通过多个认知网络或小区分担丢失频谱资源的风险,从而降低了认知系统中掉话率,提高了整个认知系统的可靠性。为了降低上述机制在认知宽带系统中的实现复杂度及系统信令开销,本文权衡了认知系统的可靠性与频谱连续性,设计出了自适应动态频谱分配机制,该机制通过衡量认知系统中业务类型的分布,设定认知系统可靠性门限,采用最优化方法选择出满足可靠性门限的连续频谱资源。针对频谱不稳定性而引起的认知系统中的频繁频谱切换,本文提出了基于业务分裂机制及频谱映射的网内动态频谱分配机制,该机制将业务流分裂为基础层数据流与提高层数据流,并映射到具有不同稳定度的频谱资源上,从而提高认知系统的鲁棒性,降低频谱切换率。本文在频谱接入机制方面的创新成果如下：针对认知网络中频谱检测过程存在的检测误差现象,本文提出了基于频谱预测的频谱接入机制,该机制提出了一种新的基于SVM的频谱预测方法,利用该方法实现有效的频谱预测,从而对检测结果的正确性进行实时评估并修正；该机制适用于认知系统利用授权系统在时域-频域上的频谱空洞的场景。针对授权系统对频谱资源的干扰温度限的要求,本文分别提出了考虑多用户功率控制机制的授权系统上行链路以及下行链路的干扰模型,用于指导不同地理位置的认知终端的频谱接入及发射功率,该机制有效避免了对授权系统产生恶劣干扰,并保证了认知系统的正常通信,适用于认知系统利用授权系统空域-频域上的频谱空洞的场景。本文在资源优化方面的创新成果如下：针对认知系统中频谱资源的多维性特征以及业务的多样性,本文提出了基于频谱多维特征联合描述的资源优化方法,建立了频谱资源与认知终端承载业务的匹配关系,在这个匹配关系中充分考虑到了业务类型、认知终端的地理位置、授权系统的干扰温度限要求、认知终端与授权终端之间的频谱间隔及频谱的暂间性特征,实现了认知系统对频谱资源利用的有效性。论文最后对全文进行了总结,并指出了今后的研究方向。更多还原

【Abstract】 Cognitive Radio is intended to improve spectrum efficiency by reusing the spectrum resource shared between the primary system and the cognitive system. Therefore, cognitive radio has been thought to be the best way to relief the problem of spectrum shortage and has been paid more and more attentions. However, how to make cognitive system have an effective utilization of the spectrum resource becomes an important priority.Firstly, traditional dynamic spectrum allocation mechanisms focus on avoiding interference among cognitive networks in order to share spectrum effectively among these cognitive networks, but they do not guarantee the reliability of the cognitive system; secondly, traditional spectrum access mechanisms are applied under spectrum sensing errors or interference temperature limit required by the primary system, these mechanisms are designed to avoid interference to the primary system while ignoring system performance of cognitive system; thirdly, there are multi-dimensional characteristics of spectrum resource in cognitive system, but traditional resource optimization methods just consider parts of these characteristics, how to combine these multi-dimensional characteristics and realize an optimum match between spectrum resources and radio services need to be deeply researched. Therefore, in order to make cognitive system re-use spectrum resources effectively, the thesis researches these three mechanisms mentioned above and makes some innovations listed as follows. The innovations about dynamic spectrum allocation mechanisms are shown as following:To deal with the call dropping problem caused by the instability of spectrum resource in cognitive network, a dynamic spectrum allocation mechanism based on risk-sharing has been proposed in this thesis. By this mechanism, the call dropping rate is reduced while the stability of the cognitive system is improved, in a way of sharing the risk of spectrum reoccupation by the primary system.When the above mentioned mechanism is applied to the broadband cognitive system with various channel bandwidths, the cognitive system will have more overhead. Therefore, a self-adaptive dynamic spectrum allocation mechanism is proposed. The mechanism can determine the threshold of system stability for each cognitive network, according to the distribution of services types, and then optimization theory is referred to achieve a better trade-off between spectrum stability and continuity. The mechanism can improve system stability as well as reduce system overhead.To deal with the frequent spectrum handoff problem caused by the instability of spectrum in cognitive network, a new spectrum assignment mechanism intra cognitive cell is proposed in this thesis. The mechanism is based on adaptive radio multi-homing and mapping scheme between traffic flows and spectrum resource.The innovations about spectrum accessing mechanisms are shown as following:For sensing errors in spectrum sensing mechanism, a spectrum access mechanism based on spectrum prediction is put forward. In this mechanism, a new method of spectrum prediction based on Support Vector Mechanism (SVM) is proposed. This method can predict spectrum occupied or idle more effectively, and judge the correctness of spectrum sensing results accordingly, in order to eliminate the harmful effect caused by sensing errors in the cognitive network. This mechanism is applied to the scenario in which cognitive and primary networks share the primary spectrum in temporal.For the interference temperature limit required by the primary system, a spectrum access mechanism based on uplink and downlink interference models and power control on multi-users is put forward, it can guide cognitive terminals in different locations to access the spectrum and to transmit in a certain power. It can suppress interference to the primary system and guarantee the communication in the cognitive system. This mechanism is applied to the scenario in which cognitive and primary networks share the primary spectrum in spatial.The innovation about resource optimization method is shown as following:For multi-dimensional characteristics of spectrum and diversity of services in cognitive system, a resource optimization method based on joint descriptions of spectrum characteristics has been given. The method can establish relational matching between spectrum resource and services, and in this matching, service type, location of cognitive terminals, interference temperature limited, spectrum space between the cognitive transmitter and the primary transmitter, and spectrum instability are considered comprehensively. The method can make cognitive system use spectrum resource effectively.A summary is given at the end of this thesis, where the future research directions related to this thesis are also pointed out.更多还原