基于CAE与AdaBoost的频谱感知

摘要

针对无线信道环境中各低信噪比情况下主用户信号检测率较低的问题，提出一种基于循环自相关估计（CAE ）和AdaBoost的认知网络频谱感知算法。对信号采用循环自相关估计算法进行特征参数提取，生成训练样本和待测样本，采用AdaBoost算法分别对有无主用户情况下的信号进行分类检测。仿真结果表明，与人工神经网络和最大最小特征值算法相比较，该算法在各低信噪比情况下，具有较高的分类检测性能，有效地实现了对主用户信号的感知。%Aiming at the problem of the low accuracy rate of the primary user detection under the circumstance of low signal to noise ratio in the wireless channel environment,a method based on cyclic auto-correlation estimation (CAE)and AdaBoost for the primary user spectrum sensing was proposed.A set of cyclic spectrum features were calculated,and the cyclic auto-correla-tion estimation algorithm was applied to extract the characteristic parameters of the received signal.The feature vectors were formed with the characteristic parameters as training samples and testing samples for classification.Finally,the trained Ada-Boost was utilized to detect the primary user signal.Test result shows that the proposed algorithm has good performance for classification and detection under the circumstance of low signal to noise ratio compared with artificial neural network and maxi-mum-minimum eigenvalue.