论文部分内容阅读
本文建立了在线计算目标运动要素的时间序列分析方法。它无需目标运动假定、无需量测噪音假定,亦无需观察时间假定。当给定量测的激励白噪音平方和最小准则下,在给定模型的阶(1、n、m)时,最优估计参数()。一旦估计参数()给定后,用二维搜索法估计(1、n、n-1)的阶,进而找到模型的阶(1、n、m)。文章指出了本方法等价于一类有色噪音滤波——参数识别模型。对两种建模方法进行了稳定性、可测性、逆转性的条件分析。最后用两个例子说明了两种建模的关系。
This paper establishes a time series analysis method to calculate the target motion elements online. It requires no target motion assumptions, no measurement noise assumptions, and no need to observe time assumptions. The optimal estimation parameter () is given for the order (1, n, m) of a given model given the minimum criterion for the square sum of excitation white noise for a given measurement. Once the estimated parameters () are given, the order of (1, n, n-1) is estimated by two-dimensional search and the order (1, n, m) of the model is found. The article points out that this method is equivalent to a kind of colored noise filtering - parameter identification model. The stability, testability and reversibility of the two modeling methods are analyzed. Finally, two examples illustrate the relationship between the two models.