目前,在进行液压系统辨识时,往往是人为地根据理论模型来确定系统的阶次,然后辨识参数。这首先要对被辨识的对象有较深入的了解,而且理论定阶得到的阶次往往高,在实际应用时,特别是用于数字控制器设计,是困难的。本文提出用受控的时间序列CAR 模型来描述电液伺服系统,采用AIC 准则定阶,用递推辅助变量法辨识参数。再根据被辨识参数的渐近正态性确定模型的子阶 At present, in the hydraulic system identification, it is often artificial to determine the order of the system according to the theoretical model, and then identify the parameters. This first of all have to have a better understanding of the object being identified, and the order of the theory is often high orders, in practical applications, especially for digital controller design is difficult. In this paper, a controlled time series CAR model is proposed to describe the electro-hydraulic servo system. The AIC criterion is used to rank and the recursive auxiliary variable method is used to identify the parameters. Then the sub-order of the model is determined according to the asymptotic normality of the identified parameters