一种基于循环谱特征识别的频谱感知算法

Spectrum Sensing Algorithm Based on Cyclic Spectrum Feature Recognition

针对传统循环谱计算复杂度高的问题,提出一种基于循环谱特征识别的频谱感知算法,该算法采用自相关矩阵替代频域平滑方法计算信号的谱特征,在降低算法复杂度的同时,能够准确区分噪声与有用信号,利用软件无线电构建测试平台,用以验证算法的有效性。实验结果表明,该算法能够准确检测出频谱空穴,准确度大于90%,并能降低时间复杂度,可以满足认知无线电中对频谱感知的性能要求。

Aiming at the high computiaion complex for traditional cyclic spectrum,this paper presents a cyclic spectrum recognition based spectrum sensing algorithm,which uses correlation matrix instead of frequency domain smoothing to compute spectrum feature.This algorithm greately reduces the complexity while distinguish between noise and useful signal accurately.A test platform is constructed based on software defined radios to demonstrate the effectiveness of the method.Experimental results show the spectrum holes of different primary users can be detected with high accuracy greater than 90% and low complexity.The performance can satisfy the requirement of spectrum sensing in cognitive radio.