基于证据理论的多特征联合数据关联算法研究

Reaerch on the Multiple Features Joint Data Association Algorithm Based on DS Evidence Theory

数据关联算法是雷达数据处理的核心技术之一。传统的数据关联算法利用目标位置信息进行关联,如最优邻近、概率数据关联等等。当目标环境异常复杂时,如目标在杂波区、目标密集区时,基于位置信息的传统数据关联算法性能下降,易造成目标跟踪丢失。文章根据证据理论思想,提出了一种基于目标多特征信息的联合数据关联算法。对于回波起伏缓慢的目标,将目标位置、目标方位变化率、目标能量和等特征信息代入目标点迹航迹关联过程。在杂波区、目标密集区时,利用目标点迹的特征变化来判断点迹是否为目标正确关联点迹。通过雷达实测数据验证表明,多特征联合数据关联算法可以改善目标在杂波区、目标密集区等环境下的稳定跟踪能力,具有一定的实际工程应用价值。

Data association algorithm is one of the key technologies about radar data process. Nearest neighbor algorithm and the probabilistic data association algorithm use target position as traditional data association method. When Target in the echo and cluttered environment, the capability of the traditional data association method based on the position get worst, and cause to target lost. In this paper, according to the DS evidence theory, introduce a new data association method base on the target multiple feasure. For the steady target, using position state, signal amplitude, azimuth change feasure when measurement-track data association, to speculate the possibility that the dot is from the real target. The results demonstrate that the multiple features joint data association algorithm has the value of application and improves the tracking performance in clutter area and closely spaced targets area by validating with the real radar data.