论文部分内容阅读
本文提出一种估计自回归AR参数的新算法。新算法采用递推Householder变换算法。文中给出了ARMA(4,4)仿真计算例子及两个正弦加白噪声的仿真计算结果,并与最小二乘法的计算结果进行了比较。结果表明新算法在分辨率和估计质量方面均优于最小二乘法和已有的谱估计方法,也说明用提高算法稳定性的方法可解决负谱问题和提高谱估计质量。
This paper presents a new algorithm for estimating AR parameters. The new algorithm uses recursive Householder transform algorithm. The simulation results of ARMA (4,4) simulation and two sinusoidal white noise are given in the paper and compared with the results of least square method. The results show that the new algorithm is better than the least square method and the existing spectral estimation methods in terms of resolution and quality of estimation. It also shows that the negative spectral problem can be solved and the spectral estimation quality can be improved by improving the stability of the algorithm.