残差滤波及其在密集多回波环境下机动多目标跟踪中的应用

 修正概率数据关联滤波（MPDAF）是目前解决密集多回波环境下机动多目标跟踪较为有效的方法,但当回波密度增高时,该方法容易失跟。本文针对此特点,在MPDAF基础上,提出了残差滤波（RF）方法,分别从理论和Monte carlo仿真两方面揭示了RF与数据关联的内在机理,结果表明该方法能大幅度提高跟踪滤波器捕捉目标和剔除关联域内多余回波的能力以及跟踪的性能,是一种解决高密集多回波环境下机动多目标跟踪的有效方法。

RESIDUAL FILTER AND ITS APPLICATION TO TRACKING MULTIPLE MANEUVERING TARGETS IN DENSE MULTI-RETURN ENVIRONMENT

Modified probabilitic data association filter (MPDAF) is now a more efficient approach to tracking multiple maneuvering targets in dense multi-return environment. But, when the density of returns increases, such approach will lose targets easily. To overcome this shortcoming, this paper puts forword an approach of risidual filter (RF) on the basis of MPDAF and reveals its inherent mechanism between RF and data association from two aspects of theoretical analysis and Monte Carlo simulation. The results show that the approach improves the catching ability of the tracker, the ability of rejecting extraneous returns in correlation regions, and the tracking performance greatly. The new approach is an efficient method for tracking multiple maneuvering targets in higher dense multi-return environment.