论文部分内容阅读
本文在正弦信号频率估计的线性预测方法基础上,充分利用正弦信号的特殊性信息,给出了一种正弦信号频率估计的快速算法。文中首先利用线性预测参数的对称性将原线性预测方程的阶数降低一半求解,使其计算量减少到原来的八分之一。提取正弦信号的频率,还要求解一高次特征多项式方程,利用其系数的对称性,经过巧妙的根的变换,还可把该高次复根多项式方程的求解转化为次数减半的实根多项式方程的求解。该实根多项式方程的求解可此原复根多项式方程求解的速度提高八倍。文中最后给出计算机模拟结果,并与原线性预测方法和前后向线性预测方法加以此较。
Based on the linear prediction method of sinusoidal signal frequency estimation, this paper makes full use of the special information of sinusoidal signal and gives a fast algorithm of frequency estimation of sinusoidal signal. Firstly, the order of the linear prediction equation is reduced by half using the symmetry of the linear prediction parameters to reduce the computation amount to one eighth. Extracting the frequency of sinusoidal signal, we also need to solve a high-order eigenvalue polynomial equation. Using the symmetry of its coefficients, through the clever transformation of the root, the solution of the high-order complex root polynomial equation can be transformed into the real root of the number of halves Solving the Polynomial Equation. The real root polynomial equation can be solved by the original complex root polynomial equation to improve the speed of eight times. Finally, computer simulation results are given and compared with the original linear prediction method and the forward-backward linear prediction method.