一种基于极限学习机的频谱感知方法  

Spectrum Sensing Method in Cognitive Radios Based on Extreme Learning Machine

作　　者：刘龙飞[1] 孙学斌[1] 周正[1] 

机构地区：[1]北京邮电大学信息与通信工程学院,北京100876

出　　处：《无线电工程》2015年第3期25-29,共5页Radio Engineering

基　　金：国家自然科学基金资助项目(60902046)

摘　　要：为了改善认知无线电系统中频谱感知的检测概率和检测速度,提出了基于极限学习机的频谱感知方法。极限学习机仅需要配置隐层节点个数,参数选择相对较容易,在算法执行过程中可以产生最优的唯一解,且训练速度快。计算出样本信号的能量值和循环谱值,对模型进行训练,使之具有快速判决测试信号的最佳性能。通过仿真结果可以看出,其在检测性能上优于传统的频谱感知算法和支持向量机感知算法,并在训练速度上比支持向量机感知算法有大幅度的性能提升。To improve the detection accuracy rate and speed of cognitive radio spectrum sensing system,a spectrum sensing method based on Extreme Learning Machine（ ELM） is proposed. The ELM only needs to configure the number of nodes in the hidden layer,the parameter selection is relatively easy,and during the implementation of this method,a unique optimal solution and fast training speed can be generated. The energy value and cyclic spectrum value of the sample signal are calculated. These values are used to train the proposed model to make it have optimal performance for fast judgment of test signal. The simulation results show that the proposed method has better performance compared with the traditional methods and SVM-based sensing method,and the training speed is improved significantly compared with SVM-based sensing method.

关 键 词：认知无线电 频谱感知 极限学习机 支持向量机 

分 类 号：TN911.7[电子电信—通信与信息系统]