【摘要】 在研究协方差矩阵特性的基础上 ,给出一种最小二乘辨识系统的递阶算法。该算法与自回归模型参数辨识的 Levinson方法相似 ,利用协方差矩阵的特性 ,由低阶模型参数递阶算出高一阶的参数 ,用阶次辨识方法确定模型阶次。避免了最小二乘系统辨识批处理方法中的高阶矩阵求逆及最小二乘系统辨识递推算法中的循环修正 ,提高了计算的效率。该辨识方法成功地应用于飞机颤振特性的辨识更多还原

【Abstract】 After inspecting the properties of the covariance matrix, a recursive method of system parameter estimation is derived upon LSE. Like Levinson algorithm in an autoregression model identification, this method, taking advantage of the specification of the covariance matrix, calculates the higher order parameters from the lower order parameters and identifies the model order by using the ordinary order identification methods. This method avoids the high order matrix inverse operation in batch process of LSE,providing a more effective estimation method. Finally, this method is successfully used in the identification of aircraft buffeting characteristics.更多还原