基于跨层优化的频谱感知关键技术研究

发布时间：2018-05-25 19:51

  本文选题：认知无线网络 + 频谱感知　； 参考：《解放军信息工程大学》2014年博士论文

【摘要】：认知无线网络(Cognitive Radio Networks, CRN)是解决当前“频谱资源匮乏”问题的有效方法之一,近年来已成为通信领域的研究热点。该技术的基本思想是在保证授权用户(或主用户)一定服务质量的前提下,使非授权用户(或次用户)“伺机地”接入临时空闲的无线频段,从而使无线频谱资源得到更为充分、合理的利用。认知无线网络中的一项关键技术是频谱感知,用于发现无线网络环境中的空闲频谱资源。从网络分层的角度来看,频谱感知技术可以分为物理层频谱感知和MAC.层及跨层感知两类。其中,物理层频谱感知主要关注如何通过有效的信号检测算法,对信道的占用状况做出快速、准确的判断；而MAC层及跨层感知则是借鉴了跨层优化的思想,通过对感知参数和策略的选取及优化,来进一步提高物理层频谱感知的性能。目前,由于数字信号处理技术的先期发展,使得物理层频谱感知算法已具有了深厚的理论基础。相对而言,有关MAC层及跨层感知的研究还处于起步阶段。为此本文依托国家.863探索导向类项目“基于行为预测的认知网络资源优化分配技术”,基于现有的物理层频谱感知算法,对MAC层及跨层感知中多个重要感知参数和策略的选取方法进行了优化研究,其中主要包括：感知周期、感知时间长度、感知信道策略以及协作频谱感知中的融合参数。主要研究成果包括：(1)为提高多信道条件下的频谱资源检测率,提出了一种多信道协作频谱感知周期优化算法。通过分析频谱感知过程中可能影响次用户发现或使用空闲频谱资源的两种情况,基于连续时间马尔科夫理论推导出衡量多信道频谱资源检测率的目标函数,并将各信道感知周期的选取建模为一个带约束条件的多目标优化问题。在求解该多目标优化问题的过程中,为避免传统“权重法”所固有的主观因素对解的质量影响较大这一问题,文章采用了相对复杂的遗传算法对其求解。另外,为提高遗传算法在本文应用背景下的算法性能,文章还对其进行了改进,使算法的收敛速度以及最终解的质量均得到了较大提高。所提算法与现有主流算法的区别在于：在多信道感知周期的现有研究中,大多简单假设次用户是以相同的周期对各信道进行感知,而本文算法充分考虑了各授权信道间的负载差异,分别以不同的周期对其感知,从而使多信道条件下的频谱资源检测率得到了较大提高。(2)基于协作频谱感知的认知无线网络中,已有研究表明增加参与协作频谱感知的次用户数量能够提高感知性能以及信道的吞吐量。然而,由于信道容量的限制,次用户数量的不断增加并不会使信道吞吐量无限提高,反而会使次用户平均可获得的吞吐量不断降低。因此，本文认为认知无线网络的研究不能仅以提高频谱利用率为目标,还应考虑次用户平均可获得的吞吐量大小,即以次用户平均吞吐量作为认知无线网络性能的衡量标准更能直接体现次用户的利益。为此本文以最大化次用户平均吞吐量为目标,基于k/N融合准则,对感知时间和融合参数的选取方法进行了优化研究。证明了对于任意给定的融合参数,次用户的平均吞吐量是感知时间的凸函数,并基于该特性提出了交叉迭代算法进行二维优化。仿真结果表明,当信噪比为-10dB时,相对于仅考虑感知时间或融合参数的一维优化算法,所提算法可使次用户的平均吞吐量提高20%以上。(3)认知无线网络不仅要具有自适应性,更应具备一定的智能性。在未知无线环境参数以及网络动态特性的前提下,为使次用户能够从多个授权信道中选择吞吐量回报较高的信道优先进行感知,文章将强化学习理论引入到认知无线网络中,以最大化次用户吞吐量为目标,提出了一种基于强化学习的智能信道选择算法。该算法利用了强化学习理论中的在线交互式学习技术,使次用户仅通过不断与环境进行交互学习,便能够逐步改进其行为策略,使吞吐量回报逐渐得以提高。另外,该算法还借鉴了模拟退火算法的思想对信道的选择动作进行优化,使之能够从学习阶段平滑地过渡到使用阶段。仿真结果表明,相对于现有信道选择算法,本文算法可有效提高次用户的吞吐量,并且在主用户使用规律发生变化时,能够自动实现二次收敛,可作为认知无线电系统迈向智能化的一种尝试。(4)针对认知无线网络中次用户节点能量受限问题,文章引入能量传输效率作为评价次用户能量有效性的指标,并根据主用户非时隙返回信道可能与次用户发生碰撞的特点,基于连续时间马尔科夫理论对次用户的频谱感知和接入活动进行建模,提出了一种联合考虑感知时间和接入概率的能效优化算法。文中证明了的确存在最优的感知时间,能够使次用户的能量有效性达到最优。此外,基于最优感知时间,本文还提出了一种能量有效的频谱接入策略。区别于传统接入策略,次用户在发现空闲信道的情况下并不直接进行接入,而是先根据信道空闲时间的统计规律以及次用户的接入时间,计算接入后与返回主用户的碰撞概率,再依概率进行接入,从而使次用户的能量有效性得到了较大提高。
[Abstract]:Cognitive Radio Networks (CRN) is one of the effective methods to solve the current "lack of spectrum resources". In recent years, it has become a research hotspot in the field of communication. The basic idea of this technology is to make unauthorized users (or sub users) "server" on the premise of guaranteeing the quality of service of authorized users (or main users). A key technology in cognitive wireless network is spectrum sensing, which is used to discover free spectrum resources in wireless network environment. From the point of view of network stratification, frequency spectrum sensing technology can be divided into physical layer spectrum sensing and MA. The C. layer and the cross layer perception are two categories. Among them, the physical layer spectrum perception mainly focuses on how to make a quick and accurate judgment on the occupancy of the channel by effective signal detection algorithm, while the MAC layer and cross layer perception refer to the idea of cross layer optimization, and further improve the physical layer through the selection and optimization of the perceptual parameters and strategies. The performance of spectrum sensing. At present, due to the advance of digital signal processing technology, the spectrum sensing algorithm of physical layer has a profound theoretical basis. Relative, the research on MAC layer and cross layer perception is still in the initial stage. This paper relies on the national.863 exploration oriented project "cognitive network based on behavior prediction" Based on the existing physical layer spectrum sensing algorithm, this paper optimizes the selection methods of several important perceptual parameters and strategies in MAC layer and cross layer perception, which mainly include the perception period, the perception time length, the perceptual channel strategy and the fusion parameters in the cooperative spectrum sensing. The results include: (1) a multi channel cooperative spectrum sensing cycle optimization algorithm is proposed to improve the spectrum resource detection rate under multi channel conditions. By analyzing two situations that may affect the secondary user discovery or the use of free spectrum resources in the spectrum sensing process, the multichannel spectrum is derived to measure the multichannel spectrum based on the continuous temporal Markov theory. The objective function of the detection rate of resources is modeled as a multi-objective optimization problem with constraint conditions. In order to avoid the problem that the subjective factors inherent in the traditional "weight method" have great influence on the quality of the solution, the article adopts a relatively complex heredity in solving the multi-objective optimization problem. In addition, in order to improve the performance of the algorithm in the background of the genetic algorithm, the paper also improves the algorithm, which makes the convergence speed of the algorithm and the quality of the final solution greatly improved. The difference between the proposed algorithm and the existing mainstream algorithm is that in the existing research of the multi-channel perception period, the algorithm is mostly simple. It is assumed that the sub user is aware of the channels in the same cycle, and this algorithm fully considers the load differences between the authorized channels and perceiving them in different cycles, so that the detection rate of the spectrum resources under the multi channel condition has been greatly improved. (2) in the cognitive wireless network based on cooperative spectrum sensing, the research has already been studied. The increase in the number of sub users participating in cooperative spectrum sensing can improve the perceptual performance and the throughput of the channel. However, the increase in the number of sub users will not increase the throughput of the channel indefinitely because of the limitation of the capacity of the channel. On the contrary, the throughput of the sub users will be reduced. The research of line network should not only aim at improving the spectrum utilization, but also take into account the average throughput size of the secondary users. That is, the average throughput of the sub user as a measure of the performance of the cognitive wireless network can directly reflect the interests of the secondary users. Therefore, the aim of this paper is to maximize the average throughput of the user and based on the k/N fusion. According to the criterion, the selection method of perceptual time and fusion parameter is optimized. It is proved that the average throughput of the sub user is the convex function of the perceptual time for any given fusion parameter, and the cross iteration algorithm is proposed for two-dimensional optimization based on this characteristic. The proposed algorithm can improve the average throughput of the secondary users by more than 20%. (3) the cognitive wireless network should not only have adaptability, but also have a certain intelligence. Under the premise of unknown wireless environment parameters and network dynamic characteristics, the sub user can get from multiple authorized channels. In this paper, the channel preference for higher throughput is preferred, and the reinforcement learning theory is introduced into the cognitive wireless network. In order to maximize the secondary user throughput, an intelligent channel selection algorithm based on reinforcement learning is proposed. This algorithm makes use of the online interactive learning technology in the reinforcement learning theory to make the sub user. Through interactive learning with the environment, the behavior strategy can be improved gradually and the throughput returns are gradually improved. In addition, the algorithm also uses the thought of simulated annealing algorithm to optimize the channel selection action, making it smooth transition from the learning stage to the use stage. There is a channel selection algorithm. This algorithm can effectively improve the throughput of the secondary users, and can automatically achieve two convergence when the rule of the main user changes, which can be used as an attempt to intelligentize the cognitive radio system. (4) the energy transfer is introduced in this paper for the problem of the energy limitation of the sub user nodes in the cognitive wireless network. Efficiency is an index to evaluate the energy efficiency of secondary users, and based on the characteristics of the collision between the main users and the secondary users in the non time slot. Based on the continuous time Markov theory, the spectrum perception and access activities of the secondary users are modeled, and an energy efficiency optimization algorithm which combines the perception time and the access probability is proposed. It is proved that there is an optimal perception time that can make the energy efficiency of the secondary user optimal. In addition, based on the optimal perception time, this paper also proposes an energy efficient spectrum access strategy, which is different from the traditional access strategy. The statistical law of the idle time of the channel and the access time of the sub user are used to calculate the collision probability with the return of the main user after the access, and then the access is carried out according to the probability, thus the energy efficiency of the secondary users is greatly improved.
【学位授予单位】：解放军信息工程大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN925

【相似文献】

中国期刊全文数据库 前10条

1 刘婷婷;王建新;束锋;;合作频谱感知吞吐量和感知时间关系的研究[J];现代雷达;2009年05期

2 常帅;孔凡军;陈娟;郑学强;;协同频谱感知中的认知用户数优化[J];军事通信技术;2010年03期

3 汪一鸣;周刘蕾;;基于追踪定位的认知频谱感知解决方案[J];通信学报;2010年11期

4 岳文静;郑宝玉;孟庆民;崔景伍;解培中;;衰落信道下提高协作频谱感知能力的方法[J];中国科学:信息科学;2011年02期

5 韩勇;陈强;王建新;;合作频谱感知安全技术研究[J];通信技术;2011年06期

6 韩勇;王放;陈强;王建新;;一种信任度模糊分配的合作频谱感知算法[J];信号处理;2011年06期

7 蒋金波;王可人;陈小波;金虎;;频谱切换中基于频谱感知的链路保持概率[J];探测与控制学报;2011年06期

8 刘颖;杨震;;协作频谱感知系统中的结盟技术研究[J];南京邮电大学学报(自然科学版);2012年05期

9 秦臻;薛峰;梁继民;;最大系统效用合作频谱感知优化算法[J];信号处理;2013年02期

10 江晓林;顾学迈;何晨;;基于压缩感知的联合协作频谱感知算法[J];上海交通大学学报;2013年07期

中国重要会议论文全文数据库 前10条

1 岳文静;郑宝玉;;一种基于信道可靠性的协作频谱感知算法[A];2009年通信理论与信号处理学术年会论文集[C];2009年

2 张培;孟庆民;朱卫平;;一种可以减少感知时间的频谱感知方法[A];2010年通信理论与信号处理学术年会论文集[C];2010年

3 郝博雅;周辉;孙斌;;基于权值的合作频谱感知算法[A];第六届全国信号和智能信息处理与应用学术会议论文集[C];2012年

4 吴昊;赵杭生;;基于三阶累积量的频谱感知技术研究[A];2009年全国无线电应用与管理学术会议论文集[C];2009年

5 张冰玉;郑宝玉;岳文静;;一种基于串行网络的协作频谱感知方案[A];2010年通信理论与信号处理学术年会论文集[C];2010年

6 冯子木;漆春梅;王军;李乐民;;协同频谱感知算法综述[A];2008年中国西部青年通信学术会议论文集[C];2008年

7 王翊;胡艳军;唐希雯;;一种改进的基于多阈值的协作频谱感知方法[A];2011年通信与信息技术新进展——第八届中国通信学会学术年会论文集[C];2011年

8 郑仕链;何斌;杨伟程;;认知无线电宽带频谱感知试验研究[A];浙江省信号处理学会2011学术年会论文集[C];2011年

9 李佳俊;黄清;孔勇;;基于小波变换的认知无线电宽带频谱感知方法[A];中国电子学会第十六届信息论学术年会论文集[C];2009年

10 杨家胜;刘光斌;程俊仁;;基于压缩采样技术的认知无线电宽带频谱感知方法[A];中国通信学会第六届学术年会论文集（上）[C];2009年

中国博士学位论文全文数据库 前10条

1 肖淑艳;认知无线电中频谱感知及资源优化分配的研究[D];中国矿业大学;2015年

2 卿浩博;面向无线通信系统的频谱感知理论与技术研究[D];北京邮电大学;2015年

3 杨雪洲;认知中继网络频谱感知与资源分配研究[D];电子科技大学;2014年

4 卢剑奇;认知无线电频谱感知关键技术研究[D];电子科技大学;2015年

5 刘洋;基于跨层优化的频谱感知关键技术研究[D];解放军信息工程大学;2014年

6 齐佩汉;稳健精细抗差异性频谱感知技术研究[D];西安电子科技大学;2014年

7 乔晓瑜;基于频谱感知的动态频谱管理研究[D];北京交通大学;2013年

8 李含青;基于压缩感知的宽带频谱感知技术研究[D];哈尔滨工业大学;2014年

9 欧扬;选择性协同频谱感知技术研究[D];苏州大学;2014年

10 刘思杨;下一代无线通信系统中高效传输及频谱感知关键技术的研究[D];北京邮电大学;2009年

中国硕士学位论文全文数据库 前10条

1 王艳君;认知无线网络中基于多节点的协作频谱感知技术的研究[D];昆明理工大学;2015年

2 谷凯;基于GNU Radio的动态频谱感知及数据传输[D];海南大学;2015年

3 孟晨;基于稀疏组套索的宽频带频谱感知研究[D];燕山大学;2015年

4 魏亮;认知车载网中频谱感知算法仿真研究[D];西南交通大学;2015年

5 周瑞;基于信任度的认知无线电联合频谱感知研究[D];湖南工业大学;2015年

6 钱陆;基于数据并发传输的频谱感知系统的研究[D];北京邮电大学;2014年

7 栾红志;基于能量检测的协作频谱感知技术研究[D];解放军信息工程大学;2014年

8 王勇锋;多天线认知无线网络频谱感知技术研究[D];解放军信息工程大学;2014年

9 邵长友;认知无线网络中宽带压缩频谱感知关键技术[D];南京邮电大学;2012年

10 林高河;认知无线网络中基于联盟博弈的频谱感知算法研究[D];北京邮电大学;2012年

，

本文编号：1934463