【摘要】 利用滤波误差能量给出了一个新的最小二乘快速递推算法。数值试验表明，该算法计算稳定性好，对噪声不敏感，能快速递推到高阶模型，并将该算法同Ｍａｒｐｌｅ算法等作了比较。更多还原

【Abstract】 AR estimation is a very important algorithm for model spectral estimation. This paper gives a new least squares fast recursive algorithm using filtering error energy. Numerical tests demonstrate that the algorithm has good computation stability and low sensitivity to noise. It can be deduced very fast to a high order AR model. It has been compared with the Marple algorithm and several numerical instances are given.更多还原