认知无线电中基于声望和重叠式联盟博弈的频谱感知和资源分配算法

Reputation-based Overlapping Coalitional Game for Spectrum Sensing and Resource Allocation Algorithm in Cognitive Radio

在认知无线电网络中,次用户频谱感知和接入会受到多径衰落和阴影衰落等因素的影响.为了提高频谱感知准确度和资源分配效率,将多个次用户合作频谱感知和接入问题建模为重叠式联盟博弈模型,每个次用户可以加入多个联盟来提升自己的期望收益.为了提高全局有效吞吐量和资源分配公平性,引入声望机制来设计联盟资源分配规则,提出了基于声望值的重叠式联盟形成(R-OCF)算法.仿真结果表明:与无声望机制算法和分离式联盟形成(DCF)算法相比,R-OCF算法的资源分配效率和公平性更高;同时,次用户的期望收益和自身声望值相关,次用户的声望值越高,获得的期望收益越大.

In cognitive radio networks, single-user spectrum sensing and access can be affected by multipath fading and shadow fading. In order to improve spectrum sensing accuracy and resource allocation efficiency, the multiple-user cooperative spectrum sensing and access problem was modeled as an overlapping coalitional game model. Each secondary user can join multiple coalitions to improve their own expected payoffs. In order to improve the global effective throughput and the fairness of resource allocation, a reputation mechanism was proposed to design the coalitional resource allocation rule, and a reputation-based overlapping coalition formation(R-OCF) algorithm was proposed. The simulation results show that the R-OCF algorithm is more efficient and equitable than non-reputation algorithm and disjoint coalition formation(DCF) algorithm. Meanwhile, the secondary user’s expected payoff is related to its reputation value. High reputation brings high payoff.