论文部分内容阅读
通过观测获得的AR序列,常常由于混有观测噪声而使AR谱估计的分辨率和估计质量大大下降。本文通过构造高阶相关阵,用householder变换递推算法做AR谱估计,得到了AR参数的无偏估计,从而剔除了噪声对AR谱的污染,保证了较好的估计质量。
By observing the AR sequences obtained, the resolution and the estimated quality of the AR spectral estimation are often greatly reduced due to the mixing of the observed noise. In this paper, by constructing higher-order correlation matrix and using householder transform recursive algorithm to do AR spectrum estimation, unbiased estimation of AR parameters is obtained, thus removing the pollution of AR spectrum by noise and ensuring a better estimation quality.