基于OFDM的认知无线网络多播资源分配

发布时间：2018-06-02 18:20

本文选题：认知无线网络 + OFDM　；参考：《北京邮电大学》2014年博士论文

【摘要】：智能终端的大量引入以及各式应用的快速涌现,给无线网络带来了前所未有的机遇与挑战。频谱利用率低下与频谱资源稀缺的矛盾日益激化,亟需全新的频谱分配与利用技术更智能地使用频谱资源。认知无线电(Cognitive Radio, CR)很好地契合了该需求,引起了学术界和工业界的广泛关注。认知获得的频谱具有零散性、碎片化特征,而正交频分复用技术(Orthogonal Frequency Division Multiplexing, OFDM)技术在资源利用方面具有的灵活性刚好满足了认知无线网络的技术需求。同时越来越多的应用需要向特定的用户群发送特定的信息,此时采用多播技术,将对同一内容感兴趣的用户作为一个多播组,为其提供多播业务,大大降低了连接用户数与所需系统资源的线性依赖性,使得相同的无线网络资源能够服务更多的用户,获得运营商和用户双赢的局面。因此,在基于OFDM的认知无线网络中有效开展多播业务是未来无线通信系统发展的一大趋势,具有重要的意义。尽管认知无线网络与多播技术具有诸多优点,但是实际性能的提升则主要取决于资源分配方案的设计与选择。因此,高效而实用的资源分配方案是基于OFDM的认知无线多播网络提高系统频谱效率、保证用户QoS性能的关键。目前,在基于OFDM的认知无线网络方面,大部分研究针对单播资源分配展开；少数文献研究了传统多播资源分配,得到了一些有用的结果,但是传统多播的性能受限于多播组中性能最差的用户,不能充分发挥无线多播的技术优势。而编码多播,将多播数据进行统一的信源编码,使得编码后的数据具有质量可扩展性,接收到的数据越多,最终解码以后恢复出来的数据质量越高,从而其性能不再受限于最差用户,基站可以根据用户信道条件的差异实现弹性多播,为用户提供更优质的多播业务,在通信系统中具有良好的应用前景。而目前针对编码多播资源分配研究几乎属于空白。考虑到认知无线网络面临的全新挑战,以及编码多播的技术难点,传统的资源分配方案已经难以满足其需求,如何设计新的方案极具挑战。本文将认知无线蜂窝网络与OFDM技术以及编码多播技术结合,针对基于OFDM的认知无线网络编码多播资源分配问题,按从简单到复杂、从理想到实际的循序渐进方式逐步深化系统资源分配模型与方法的研究,取得了一些具有一定理论价值与实际意义的研究成果。第一,本文在第三章分析了认知无线网络编码多播的极限性能。由于在大多数场景下,考虑无线传输的复杂性,以及认知无线网络的特殊性,精确的性能分析可能会推导出极其复杂的数学表达式,很难从中发现有用的规律,获得有用的指导信息,因此本文利用极值理论,对认知无线网络场景下采用单播、传统多播和多描述编码多播传输公共数据业务的渐进吞吐量性能进行了分析和比较。仿真结果表明,即使在用户数不是很大的情形下渐进分析结果也很精确。同时,理论分析和仿真结果都表明在认知场景下,多描述编码多播相比单播和传统多播具有巨大的优势,因此在认知无线网络中使用多描述编码多播进行公共数据的传输具有很大的意义,进一步验证了本论文研究工作的重要性,也为设计资源分配算法提供了理论指导。第二,本文在第四章设计了基于统计信道状态信息的多描述编码多播方案。在之前的理想模型中,假设次基站在执行资源分配时总是知道所有信道的精确信息,但是信道的估计误差、反馈时延、量化误差等不完美因素将制约系统性能的提升,另外在多播用户数较多时,进行频繁地信道信息反馈将不切实际。本文针对认知无线网络的特点,研究了非理想信道下的编码多播资源分配,利用概率统计、随机过程、序统计理论进行理论推导,设计了一种次基站仅有统计信道状态信息的多描述编码多播方案,所提方案能以低复杂度逼近最优性能,具有较好的实用价值。第三,本文在第五章和第六章设计了多小区编码多播的资源分配方案。由于实际系统中,一个小区并非孤立地存在,多小区共存是目前主流蜂窝系统的组网方式,因此有必要研究多小区场景的多播资源分配。对于基于OFDM的认知无线网络,多小区场景需要考虑小区间的同频干扰问题,相比单小区场景更加复杂,目前公开发表的文献鲜有涉及多小区多播资源分配。本文首先分析了多小区编码多播的理论性能,接下来借鉴博弈论、几何规划、对偶分解等方面的理论和方法,结合本文特定的场景,提出了两种低复杂度分布式算法,既避免了集中资源分配的高计算复杂度,又减小了小区间的反馈开销,在性能与复杂度之间取得了良好折中。
[Abstract]:The large number of intelligent terminals and the rapid emergence of various applications have brought unprecedented opportunities and challenges to wireless networks. The contradiction between low frequency spectrum utilization and the scarcity of spectrum resources is becoming more and more urgent, and a new spectrum allocation and utilization technology is urgently needed to use the spectrum resources more intelligently. Cognitive radio (Cognitive Radio, CR) is very good. The demand has attracted wide attention from the academia and the industry. The spectrum of cognitive acquisition has fragmentary and fragmented features, and the flexibility of Orthogonal Frequency Division Multiplexing (OFDM) technology has just met the technical requirements of the cognitive wireless network. The more applications need to send specific information to a specific user group, using multicast technology to provide multicast service to the users of the same content as a multicast group, which greatly reduces the linear dependence of the number of connected users and the required system resources, making the same wireless network resources more capable of serving more. Users can win a win-win situation between the operators and the users. Therefore, the effective development of multicast service in the OFDM based cognitive wireless network is a major trend in the development of future wireless communication systems, which is of great significance.
Although the cognitive wireless network and multicast technology have many advantages, the improvement of actual performance mainly depends on the design and selection of the resource allocation scheme. Therefore, the efficient and practical resource allocation scheme is based on the OFDM based cognitive wireless multicast network to improve the efficiency of the system spectrum and ensure the performance of the user QoS. At present, it is based on OFDM In the cognitive wireless network, most of the research focuses on unicast resource allocation; a few documents study the traditional multicast resource allocation and get some useful results. However, the performance of the traditional multicast is limited to the worst performance users in the multicast group, and it can not give full play to the technical advantage of wireless multicast. According to the unified source code, the encoded data has the quality extensibility, the more the received data, the higher the quality of the data recovered after the final decoding, and the performance is no longer limited to the worst users. The base station can implement elastic multicast according to the difference of the user channel conditions and provide the users with better multicast. Business has a good application prospect in the communication system. At present, the research on the allocation of coded multicast resources is almost blank. Considering the new challenges faced by the cognitive wireless network and the technical difficulties of coding multicast, the traditional resource allocation scheme has been difficult to meet its needs. It is very challenging to design a new scheme.
In this paper, we combine the cognitive wireless cellular network with OFDM technology and coded multicast technology. In view of the OFDM based multicast resource allocation problem in cognitive wireless network, the research on the system resource allocation model and the square method is gradually deepened from the ideal to the practical step by step, from the simple to the complex, and some theoretical prices have been obtained. Research results of value and practical significance.
First, in the third chapter, the ultimate performance of the cognitive wireless network coding multicast is analyzed. In most scenarios, considering the complexity of wireless transmission and the particularity of the cognitive wireless network, accurate performance analysis may derive extremely complex mathematical expressions. It is difficult to find useful rules from which useful fingers can be obtained. In this paper, we use extreme value theory to analyze and compare the incremental throughput performance of the unicast, traditional multicast and multi description multicast transmission public data services under the cognitive wireless network scenario. The simulation results show that the incremental analysis results are very accurate even if the number of users is not very large. Both analysis and simulation results show that multi description coding multicast has great advantages over mono and traditional multicast in cognitive scene. Therefore, it is of great significance to use multi description coding multicast in cognitive wireless network to transmit public data. It further validates the importance of this research work and the design resources. The matching algorithm provides theoretical guidance.
Second, in the fourth chapter, a multi description coding multicast scheme based on statistical channel state information is designed. In the previous ideal model, it is assumed that the secondary base station always knows the exact information of all channels in the execution of resource allocation, but the imperfect factors such as channel estimation error, feedback delay, quantization error and so on will restrict the performance of the system. In addition, when the number of multicast users is more, it is unrealistic to carry out frequent channel information feedback. In this paper, based on the characteristics of cognitive wireless networks, this paper studies the allocation of coded multicast resources under non ideal channels, and uses probability statistics, random process, order statistics theory to deduce the theory, and designs a secondary base station with only statistical channel state. A multi description coding multicast scheme for information is presented. The proposed scheme can approach the optimal performance with low complexity and has good practical value.
Third, this paper designs a multi cell coding and multicast resource allocation scheme in the fifth and sixth chapters. Because of the actual system, a cell is not isolated in isolation, multi cell coexistence is the networking mode of the current mainstream cellular system. Therefore, it is necessary to study multicast resource allocation in multi cell scenes. For OFDM based cognitive wireless networks, Multi cell scene needs to consider the same frequency interference between communities, which is more complex than single cell scene. At present, the published literature rarely involves multi cell multicast resource allocation. Firstly, the theoretical performance of multi cell coding multicast is analyzed, and then the theory and method of game theory, how to plan and dual decomposition are used for reference. In this particular scenario, two low complexity distributed algorithms are proposed, which not only avoid the high computing complexity of the centralized resource allocation, but also reduce the feedback overhead between the communities, and achieve a good compromise between the performance and the complexity.
【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5

【参考文献】