The invention relates to the field of signal processing, in particular to a frequency estimation method for noisy sinusoidal signals. The object of application of the present invention is the frequency estimation of sinusoidal signal with noise, which includes the following steps: first, the complex amplitude of the sampled signal is estimated by fast Fourier transform (FFT) algorithm, and a reference signal containing only negative frequency components is constructed; secondly, the real-complex conversion is realized by subtracting the sampled signal from the reference signal to suppress the influence of negative frequency components; secondly, the frequency of the complex signal is suppressed; The spectrum is interpolated by two points to estimate its frequency residuals and complex amplitudes, and the reference signal and complex signal are reconstructed. Finally, the accurate frequency estimation is obtained by iterative calculation. The frequency estimation method of the noisy sinusoidal signal of the present invention has the advantages of simple implementation, good anti-noise performance and strong real-time performance, greatly reducing the influence of negative frequency components, and improving the frequency estimation accuracy of the signal.
【技术实现步骤摘要】
一种迭代插值的实复转换频率估计方法
本专利技术涉及信号处理领域,特别是含噪正弦信号的频率估计方法。
技术介绍
正弦信号的频率估计是从含有噪声的采样信号中检测出信号的频率值,广泛应用于雷达、通信、语音、生物医学和仪表装置等领域,具有重要的理论意义和应用价值。频率估计方法主要可以分为两大类:时域法和频域法。时域法是对采样信号进行时域变换得到频率估计值的方法,该类方法思路简单,但易受信号非整周期采样影响,抗噪性较差,计算量较大,不利于实际应用,主要包括扩展自相关法、相频匹配法等。频域法是对采样信号进行频谱分析从而获取相位差估计值的方法,容易借助硬件实现,计算速度快,实时性好,且具有更强的抗噪性,因而得到了更多的研究,主要包括加窗插值法、频谱搬移法等。(1)扩展自相关法(参考文献[1]:CaoY,WeiG,ChenFJ.Aclosed-formexpandedautocorrelationmethodforfrequencyestimationofasinusoid[J].SignalProcessing,2012,92(4):885-892.),该方法通过对采样信号进行自相关计算,得到自相关信号并生成参考信号,利用最小二乘法构造误差函数,通过最小化误差函数得到频率估计值。该方法简单,但易受信号非整周期采样影响,且在低信噪比下的估计精度不高。(2)相频匹配法(参考文献[2]:TuYQ,ShenYL.Phasecorrectionautocorrelation-basedfrequencyestimationmethodforsinusoidalsignal[J]. ...
【技术保护点】
1.一种迭代插值的实复转换频率估计方法,其特征在于:适用对象为含噪正弦信号的频率估计;该方法包括以下步骤:首先,设定频率残差初值:
【技术特征摘要】
1.一种迭代插值的实复转换频率估计方法,其特征在于:适用对象为含噪正弦信号的频率估计;该方法包括以下步骤:首先,设定频率残差初值:迭代次数:I;第一步:对采样信号x0(n)进行快速傅里叶变换,得到X(k)=FFT(x0(n)),0≤k≤N/2-1,由得到k0,并利用式估计信号复幅值;式中:n表示采样信号x0(n)时刻点,n=0,1,…,N-1,N表示信号长度,argmaxX(k)表示X(k)取最大值时...
【专利技术属性】
技术研发人员:涂亚庆,陈鹏,李明,
申请(专利权)人:中国人民解放军陆军勤务学院,
类型:发明
国别省市:重庆,50
还没有人留言评论。发表了对其他浏览者有用的留言会获得科技券。