正弦信号频率估计的改进高阶自相关算法

Modified high-lag autocorrelation estimation method for frequency estimation of sinusoidal signal

针对混有加性高斯白噪声的正弦信号,利用正弦信号的线性预测性质和高阶自相关函数,提出了新的频率估计算法。新算法与多种算法进行了计算复杂度比较,同时理论推导得到新算法的频率估计方差的闭合表达。新算法平衡了估计性能和计算量之间的矛盾。在仿真实验中,与改进协方差(MC)算法、Rim算法、Pisarenko谐波分解(PHD)算法及其多种改进型算法进行比较。结果表明:本文算法总体优于各对比算法,特别在在信号序列较短和中高信噪比情况下,性能接近克拉美罗界。

Based on the linear prediction（ LP） property and high-lag autocorrelation of sinusoidal signals,a new frequency estimation algorithm for real sinusoid signals in additive white Gaussian noise is proposed. The computational complexity and theoretical variance expression of the frequency estimation algorithm are given. The new estimator can reach a compromise between estimation performance and amount of computation. Computer simulations were performed to validate the performance of the proposed algorithm via comparison with the Cramer-Rao lower bound（ CRLB） and several conventional frequency estimation algorithms,including modified covariance（ MC）,Rim,Pisarenko harmonic decomposition（ PHD）,and their modified algorithms. The results show the proposed algorithm is superior to the other methods,and its performance is close to that of CRLB for short data lengths and large SNR.