基于LE和IIN算法的正弦信号频率估计算法

摘要

Aiming at the problem that when the frequency locates in the central region of two adjacent discrete frequency points, the variance of iterative interpolation (IIN) based estimator becomes larger and the estimation precision of linear equation (LE) algorithm decrease when the frequency is near to the discrete frequency points, a new algorithm which based on LE and IIN algorithm was proposed in this paper.Firstly, a coarse frequency was estimated by the LE algorithm.Then move the original signal to the discrete frequency point via frequency shift modification.Finally, the second iteration of IIN algorithm was used to get a fine signal frequency estimate.Monte-Carlo simulation showed that the root mean square error (RMSE) approached to Cramer-Rao lower bound(CRLB) throughout whole frequency range.The performance of the proposed algorithm was better than R-IIN algorithm, M-LE algorithm and IIN algorithm both in accuracy and stability, and the computation was smaller than IIN algorithm(two iterations) and M-LE algorithm.%针对信号频率位于两个相邻离散频率点的中心区域时,迭代插值(Iterative Interpolation Based Algorithm, IIN)算法估计误差较大,而当信号频率位于离散频率点附近时,线性方程(Linear Equation, LE)算法估计误差较大的问题,提出了一种基于LE和IIN算法的正弦信号频率估计算法.该算法运用频谱搬移的思想,首先利用LE算法进行频率粗估计,然后将原信号往离散频率点附近频移,再利用IIN算法进行二次迭代.Monte-Carlo仿真结果表明,该算法频率估计的均方根误差在全频段内逼近克拉美-罗界(Cramer-Rao Lower Bound,CRLB),精度和稳定性皆优于R-IIN算法、M-LE算法和IIN算法,并且运算量小于IIN算法(二次迭代)和M-LE算法.