基于压缩感知的循环平稳特征检测方法

Cyclostationary feature detection method based on compressed sensing

在认知无线电中,传统的循环平稳特征检测技术为了达到理想的感知效果,需要大量的数据采样点,导致其感知过程复杂度大,感知时间长。针对此问题,提出了一种基于压缩感知的改进循环平稳特征检测方法,该算法利用信号循环自相关函数(Cyclic Autocorrelation Function,CAF)的稀疏特性,基于分段平均的时变自相关函数估计值,通过压缩感知技术重构二维CAF矩阵,再根据重构结果实现循环平稳特征检测。该方法不仅可有效降低计算复杂度和检测时间,而且提高了二维CAF的估计精度。仿真结果表明该方法的检测性能优于基于经典CAF估计的循环平稳特征检测技术。

In the cognitive radio,the traditional cyclostationary feature detection technology needs a large number of data samples to achieve ideal perceptual effects,which results in high complexity and long sensing time.To solve this problem,an improved cyclostationary feature detection method is proposed based on compressed sensing.It takes advantage of the sparsity of the cyclic autocorrelation function （CAF） and uses the time-varying autocorrelation estimated by segmental averaging to reconstruct the 2D CAF with compressed sensing,then the reconstructed CAF is employed to detect the cyclostationary features.This approach not only reduces computation complexity and sensing time but also improves the estimation accuracy of 2D CAF.Simulation results indicate that the detection performance of the proposed algorithm is superior to traditional cyclic feature detection method based on the classical CAF estimation.