认知无线电网络动态频谱接入方案的分析.pdfVIP

摘 要 随着无线通信技术的发展，人们对频谱的要求越来越迫切，频谱资源也越来 越紧缺。为了缓解低频谱利用率和高频谱需求这一矛盾，认知无线电技术应运而 生。认知无线电以认知能力为核心，能够智能地发现和利用频谱空洞，明显的提 高了频谱利用率。 本文主要针对认知无线电网络的关键技术之一，即动态频谱接入技术，进行 了深入的研究。相比于传统的动态频谱接入模型，具有在线自主学习能力的Q 学 习算法能够更好的解决动态频谱的接入问题。因此，本文对基于Q 学习的动态频 谱接入技术进行了深入研究，分析基于 Q 学习的动态频谱接入技术的可行性和优 势。在此基础上，本文从网络整体性能和用户需求两个角度出发，将 Q 学习算法 和动态频谱接入进行了参数映射，在中心式架构的认知无线电网络下提出一种基 于Q 学习的动态频谱接入优化算法。该算法通过认知用户根据Q 学习算法提出信 道申请和控制节点根据网络整体性能处理申请两个主要步骤的实施，完成动态频 谱的接入。仿真与分析表明该优化算法在满足用户频谱需求的同时使得网络整体 性能得到明显的改善。 关键词：认知无线电网络 动态频谱接入 Q 学习 万方数据 Abstract With the development of wireless technologies, the demond of spectrum becomes more and more urgent while the resource of spectrum becomes more and more scarce. In order to alleviate the constradiction between the low spectrum efficiency and high spectrum needs, the cognitive radio technology emerges. Cognitive radio whose core is the cognitive ability can discover and use the spectrum hole intelligently and it can improve the spectrum efficiency significantly. This paper researches on dynamic spectrum access acted as one of the key technologies in cognitive radio networks. Compared to the traditional dynamic spectrum access models, the Q- learning algorithm can solve the problem of dynamic spectrum access effectively with online autonomous learning ability. Therefore, this paper studies dynamic spectrum access technology based on Q-learning. And the feasibility and advantages are analyzed for the dynamic spectru