论文部分内容阅读
针对杂波环境下的机动目标跟踪问题,给出了目标状态与类型的联合概率密度表示,并证明了目标的状态量测预测密度函数服从高斯混合分布。根据这一性质,为每一类目标分别建立对应的类跟踪门,从而实现更有效的量测到航迹的数据关联。滤波过程使用高斯混合加权卡尔曼滤波器,避免了机动目标跟踪过程中的机动检测问题。在仿真实例中,对比三种算法的跟踪结果进一步显示了本文算法的有效性。
Aimed at the maneuvering target tracking problem in clutter environment, the joint probability density representation of target states and types is given. It is also proved that the target state predictive prediction density function obeys Gaussian mixture distribution. According to this property, a corresponding class-following gate is established for each type of target, so that the data association of the track can be more effectively measured. The Gaussian mixture-weighted Kalman filter is used in the filtering process to avoid the maneuvering detection problem during maneuvering target tracking. In the simulation example, comparing the tracking results of the three algorithms further shows the effectiveness of the proposed algorithm.