迭代无味卡尔曼滤波的目标跟踪算法

摘要

An improved iterative unscented Kalman filter (IIUKF) is proposed to increase the target tracking accuracy of iterative unscented Kalman filter (IUKF) by using the state augmentation technique. The method augments measurement noises into states, and a measurement function of augmented state with zero measurement noise is constructed. Then the latest posterior estimate of the augmented state is substituted into a updating function to iteratively correct state estimation using measurements. Comparing analyses with IUKF shows that IIUKF is more concise, and is more precise since it avoids the statistical non-orthogonal problem, Results of digital simulation show that there is a 20% decrease in the tracking error of IIUKF over that of IUKF.%针对目标跟踪迭代无味卡尔曼滤波(IUKF)算法中跟踪精度较差的问题,提出一种基于状态扩展技术的改进迭代无味卡尔曼滤波(IIUKF)算法.新算法首先将观测噪声扩展进状态,构造关于扩展状态的零噪声观测方程,然后在观测迭代过程中将最新的扩展状态后验估计代入更新公式,进行观测迭代更新.相比IUKF算法,IIUKF算法不仅形式上更为简洁,而且避免了IUKF算法中先验估计和观测噪声非统计正交的问题,滤波精度更高.数值仿真表明,IIUKF算法的跟踪误差比IUKF算法减小了20％以上.