基于二分法峰值的频率搜索算法

The Peak Frequency Search Algorithm Based on Bisection Method

用最大似然估计单个正弦信号频率的时候,涉及对连续频谱峰值搜索问题,传统的样点补零法运算量较大,峰值相邻点比幅法运算量小并且精确,但其要求特定的窗函数,并且数据适应性较差,因此提出将二分法用于峰值频率点搜索,其速度快、精确并且适应性强,也可用于多信号的频率搜索。

A problem of peak search of continuous spectrum is involved in sinusoidal frequency estimation based on maximum likelihood estimator, the traditional method of filling zero has a large amount of calculation; the method of ratio of peak amplitude to the adjacent points has a small calculation and high-accuracy, but it requires specific window function and has poor adaptability. A bisection search algorithm is applied for peak frequencies search, it has a high speed, high accuracy and good adaptability, and it also can be applied for multi-frequency signals search.