Abstract: Mixed sharing manners of spectrum are efficient for the cognitive networks in which the secondary users have different geographical locations. And as one of the key issues for wireless networks, channel allocation has been widely studied in recent years. The centralized algorithms are commonly used, but it is uneasy to implement centralized algorithms in distributed systems like cognitive radio networks. In this paper, the channel allocation in cognitive networks with mixed sharing manners is formulated as a one-to-one matching game, and a distributed user-channel matching algorithm is proposed with low complexity, which can achieve the stable matching state. Simulation results show that the proposed algorithm has the features of short convergence time, and the transmission rate of stable matching state approximates the rate of optimal channel allocation algorithm which uses the Hungarian method in the cognitive network with mixed sharing manners, much better than the rate of the random access algorithm.

Key words: cognitive radio, mixed spectrum sharing manners, channel allocation, matching algorithm

摘要： 混合频谱共享是适应认知用户不同地理分布的有效频谱共享模式。信道分配是无线通信网络中关键的问题之一，近年来得到了广泛研究。集中式信道分配算法是最常用的算法形式，但在认知无线电网络这种分布式系统中，集中式算法不易实现。将混合共享认知无线网络的信道分配问题构建为一对一的匹配博弈，提出了分布式用户-信道匹配算法。该算法数学复杂度低，且能够达到稳定匹配。仿真结果表明，算法收敛时间短，稳定匹配状态下的平均传输速率与使用匈牙利算法的最优分配算法所获得传输速率相接近，远优于随机分配算法的传输速率。

关键词: 认知无线电, 混合频谱共享模式, 信道分配, 匹配算法