认知用户动态分组下的认知网络能效提升算法

Cognitive network energy efficiency improvement algorithm based on dynamic grouping of cognitive users

提出一种动态交错分组算法来提升认知无线电网络的能量效率。该算法先对认知用户的可靠性进行评估,在认知网络的路由器中建立了训练模型,通过训练得到认知用户的可靠度初始值。以训练得到的初始值作为基准参数,认知网络的路由器采用监督学习的方法对认知用户的可靠度进行周期性更新,在每个周期内选择可靠度排名前列的偶数个认知用户参与协作感知,将选中的用户交错分为各项指标近似相等的两组,两组用户交替探测目标频段,实现在有效保护主用户的同时提高认知网络的能量效率。实验结果证明,在信噪比为-20 dB时,所提算法的能量效率较传统算法高19.1%,且高检测概率能有效保护主用户免受干扰。

A dynamic interleaving grouping algorithm is proposed to improve the energy efficiency of cognitive radio networks.The algorithm first evaluates the reliability of cognitive users,and establishes a training model in the router of the cognitive networks to obtain the initial reliability of cognitive users through training.Taking the initial reliability value obtained from training as the benchmark parameter,the router of cognitive network uses supervised learning method to update the reliability of cognitive users periodically.In each cycle,an even number of cognitive users with top reliability are selected to participate in collaborative perception,and the selected users are interlaced into two groups with approximately equal indexes.The two groups of users alternately detect the target frequency band,so as to effectively protect the primary user and improve the energy efficiency of cognitive network.The experimental results show that the energy efficiency of the dynamic grouping algorithm in this paper is 19.1%higher than that of the traditional algorithm when the SNR is−20 dB,and the high detection probability can effectively protect the primary user from interference.