论文部分内容阅读
众所周知,有色噪声可能会降低跟踪算法的性能。一种常见的补救办法是按自回归(AR)过程建立有色噪声的模型,并应用量测差分方法。该方法存在的问题是AR参数通常是未知的。本文提出一种新的方法来自适应地估计AR参数。该方法简单可行。我们将该方法与交互多模型(IMM)跟踪算法结合起来,并说明其性能几乎与已知参数的情况一样好。
It is well known that colored noise may reduce the performance of the tracking algorithm. A common remedy is to model colored noise by autoregression (AR) and to apply the measurement difference method. The problem with this method is that the AR parameters are usually unknown. This paper presents a new method to estimate AR parameters adaptively. This method is simple and feasible. We combine this approach with the Interactive Multi-Model (IMM) tracking algorithm and show that its performance is almost as good as the known parameters.