基于递推最小二乘估计的动力学模型参数识别

Parameter Recognition of Dynamic Model on the Recursive Least-square Estimation

动力学建模过程中,模型参数的不确定性和参数的未知变化使得建模精度大为降低。通过对非线性最小二乘估计理论的研究,提出动力学模型参数的最小二乘估计,并设计出相应的估计器,基于现场数据给予模型较为准确的定论。应用于曲柄滑块机构中滑块与滑动面间的摩擦系数的估计,经过较少的迭代次数获得满意的估计效果。为动力学模型参数识别,建立更精确模型提供一个有效的途径。

It is hardly to obtain a accurate model in the course of dynamic modeling for the uncertainty and the unknown changes of model parameters. In this paper, the recursive last- square estimation for dynamic model has been proposed and the corresponding estimator has also been designed after studing on the non - linear teast squares estimation theory. Then a more accurate conclusion for model parameter can be abained basing on- line data. Finally, it has bee successfaly applied to estimate the friction coefficient between the slider and the sliding surface of slider- crank mechanism and a satisfactory value has been gained after less numerous iteration times. Actually it will provide an effective way to identify dynamic model parameters and build a more accurate model.