基于预测判定和无味滤波多目标无源跟踪算法

Multi-target Passive Tracking Based on Prediction Judge and Unscented Filtering

多传感器多目标无源交叉定位时存在虚假点的问题,且随着传感器和目标数量的增加,虚假点的数量也急剧增加。针对这个问题,提出了一种对目标无源定位跟踪的新方法,即首先通过判断预测点到传感器与目标构成的传感器目标测向方程的最小距离,来选取传感器目标测向方程,该算法可以有效避免大量虚假点的产生,以传感器目标测向方程的交点作为目标量测状态,再通过无味滤波(UF)算法得到目标点的位置。此方法解决了多传感器多目标定位跟踪时大量虚假点的存在对目标定位的影响。研究表明:本算法大大降低了运算量,且提高了关联正确率以及对目标的定位跟踪精度。

The ghost problem exists in multi-sensor and multi-target passive crossing-location. The number of ghosts increases rapidly with the number of sensors and targets. Aiming at this problem, an improved way to target passive location and tracking is presented. First, chose sensor-target function of measurement direction by judging the smallest distance between predicted position and sensor-target function of measurement direction, which can effectively avoid producing a lot of ghosts. Second, treat the point in the sensor-target function of measurement direction as target measurement states. Last, we obtain the target positions through unscented filter （UF） algorithm. The algorithm can effectively deal with the effect of the ghosts. The research shows that the proposed algorithm has the following attributions： low computational load, high correct association probability and good location and tracking precision.