作为一项很有发展前景的无线通信技术，认知无线电（Cognitive Radio, CR）近年来被学者们广泛研究，用于提高频谱利用效率，满足日益增长的无线设备接入需求。在认知无线电网络中，频谱感知和共享是两个研究热点话题。为充分挖掘频谱接入机会并延长无线网络寿命，本文研究具有能量收集功能的协作认知无线电网络，提出了两个分别发生在频谱感知阶段和频谱共享阶段的主次用户合作机制。 第一个合作机制解决次级网络中的感知节点分配问题。网络系统模型中包含多个主用户和多个次用户。现有的相关工作大多没有考虑主用户接入频谱的统计属性，而在一些考虑到统计属性的研究中，系统模型未涉及次用户接入频谱的传输过程。本文的研究则同时考虑了主用户接入频谱的历史信息和次级网络的数据传输阶段，并把节点调度决策看成部分可观测马尔可夫决策过程（Partially Observable Markov Decision Processes，POMDP）中的行动（action）。另外，本文的次级网络的能量主要来源于能量收集，而其在数据传输阶段利用到了协作通信，这对次级网络的性能起到了积极的作用。最后次级网络中的基站负责执行贪心的感知节点调度算法，获得了最优的期望吞吐量。 第二个机制考虑协作认知无线电网络中的中继选择。协作认知无线电网络中主次用户的合作对双方均有益。近年来许多中继选择方法被相继提出，然而目前大多数工作只允许参与中继的次用户获得信道接入机会，这可能会引起频谱资源的浪费和次级网络吞吐量的减少。另一方面，多中继选择能够进一步挖掘主次用户之间的合作传输带来的性能提升。为了克服上述问题并充分利用合作通信传输，本文提出一个基于Lyapunov优化理论的GMLDF合作机制，实现了主用户的多径传输，给所有次用户提供了频谱接入机会，最终使频谱的利用效率得到提高。 仿真结果表明，本文提出的感知调度算法相比随机感知节点分配算法，能获得更高的次级网络吞吐量，并验证了GMLDF合作机制比已有的中继选择机制更能提高频谱利用效率。

As one of the promising approaches, cognitive radio technique has been proposed to enhance spectrum utilization and satisfy the ever-increasing demand of wireless services. As the two important functions in cognitive radio networks(CRNs), spectrum sensing and sharing are the hotpots in related research. To fully explore the spectrum access opportunities and improve the energy efficiency, this paper studies cooperative cognitive radio network with energy harvesting and proposes two cooperation schemes during spectrum sensing and sharing. Secondary users(SUs) sensor allocation is considered in an energy harvesting enabled cognitive radio network with multiple primary channels and multiple secondary users in this paper. Lots of existing work studying the problem does not take the statistical property of the primary users(PUs) access pattern into count, and other work’s system models only consider the spectrum sensing process. To enhance the throughput of the secondary network, this paper exploits the history information fully by using POMDP framework, and the scheduling policy is transformed as actions in the system. In the meanwhile, using the energy harvesting to power the SUs and perform cooperative transmission has significant impact on the performance of the secondary network. By perform greedy sensors scheduling algorithm, the optimal expectation of secondary network throughput can be achieved. The second scheme considers relay selection for cooperation transmission in cooperative cognitive radio networks(CCRNs). Cooperation in the network can benefit both the PUs and SUs, achieving a “win-win” situation. In recent years, a number of relay selection schemes have been proposed. However, most of the existing work only permits relay SUs have the opportunities to access primary user’s channel, which may incur spectrum resource waste and the throughput reducing of the secondary network if the relays have little data to send or the channels are worse. In addition, multi-relay selection can further exploit cooperation transmission between the PUs and the SUs. To deal with this issue while exploiting cooperation transmission fully, this paper proposes a novel scheme GMLDF-cooperation to achieve multi-path cooperation transmission of primary transmitters and provide all SUs with a chance to send their own data. The simulation results show the proposed method of sensing scheduling method achieves a significantly higher secondary network throughput than the random sensors allocation scheme, and validates the effectiveness of the proposed GMLDF algorithm compared with the existing algorithm.