低信噪比下一种基于优选法的正弦信号频率估计算法

针对插值FFT算法在低信噪比下估计精度不高,二分搜索法及其改进算法计算量较大或需要初始化引导等问题,提出了基于优选法的正弦信号频率估计算法,给出了测试点的设计原则和计算流程,分析了算法的频率分辨率和计算量,最后通过多次蒙特卡洛仿真对算法的有效性进行了验证.仿真结果表明,该算法通过合理设计测试点,能够在低信噪比且不需要初始化引导的情况下以较小代价达到参数估计误差下限(CRLB),此外,算法对相对频率偏差也具有较好的适应性.

A New Frequency Estimator of Sinusoid Based on Optimum Seeking Method at Low SNRs

To counter the defects of interpolated FF'I' method which has poor accuracy at low SNRs,dichotomous search method and its modification forms which have either heavy computational load or an initialization need,a new frequency estimator of sinusoid based on optimum seeking method is put forward. The rules of designing test points and the procedure of working axe presented,and then the analyses of the frequency resolution and computational load are also provided. Performance is contrasted and validated through Monte Carlo simulations. Simulation results indicate that through designing test points in reason,this method is shown to have less computational load and a performance that is comparable with the CRLB without any initialization at low SNRs. Moreover,the proposed method also shows resilience to relative frequency deviation.