针对某些工业过程控制参数高度时变的特点和LS拟合算法存在病态估计的现象，提出了一类基于 修正目标函数的时变参数估计算法，并给出递推辅助变量算法。此类算法不仅具有较强的实时跟踪能力和较高 的估计精度，而且能克服病态估计，摆脱了伪随机信号在工程中带来的麻烦。新算法没有增加计算量，非常适 合于工程应用，仿真及实际应用结果都证明新算法是有效的。文中给出了有关新算法的定理及其证明。

A new estimation method based on modified performance index is proposed for time-varying process to avoid the ill-condition of least squares method. Based on the new method,a series of new recursive algorithms such as the new recursive least squares algorithm,the instrumental variable algorithm are constructed. The new identification method can ease the ill-condition and improve the identification accuracy without using the additional signals. Moreover, without additional computation,it is suitable for engineering practice very much. Both simulation results and practical operation show the new identification algorithm is effective. In this paper, the justification of this modified algorithm is proved.