论文部分内容阅读
本文给出目标跟踪问题的一种一般化的最优滤波解。先定义一个广义参数,即跟踪指数。它正比于由目标机动引起的位置不确定性与由敏感器测量引起的位置不确定性之比。把最优滤波理论用于目标跟踪问题,发现跟踪指数这个广义参数在随机调节跟踪问题的最优稳态解和跟踪初始化过程中都具有重要的作用。对于不同阶数的跟踪模型,跟踪指数解给出一组关于广义跟踪增益、最佳参数关系和跟踪性能的封闭形式的、一致的公式。应用跟踪指数参数和一种递推形式的初始化和跟踪过程,就可以用一种象熟知的α-β滤波器或α-β-γ滤波器(这取决于跟踪的阶数)那样简单的算法达到卡尔曼滤波器的精度。
In this paper, a generalized optimal filtering solution to the target tracking problem is given. First define a generalized parameter, that is, tracking index. It is proportional to the ratio of the position uncertainty caused by the target maneuver to the position uncertainty caused by the sensor measurement. Applying the optimal filtering theory to the target tracking problem, we find that the generalized parameter of the tracking index plays an important role in the optimal steady-state solution and tracking initialization of the stochastic tracking problem. For different orders of tracking model, the tracking index solution gives a closed set of consistent formulas for generalized tracking gain, optimal parameter relationship and tracking performance. Using a tracking index parameter and a recursive form of initialization and tracking process, a simple algorithm such as the well-known α-β filter or α-β-γ filter (depending on the order of tracking) can be used Achieve Kalman filter accuracy.