一种新的机动目标跟踪的多模型算法  

A New IMM Method for Tracking Maneuvering Target

作　　者：巴宏欣[1] 赵宗贵[2] 杨飞[1] 刘海燕[1] 

机构地区：[1]解放军理工大学指挥自动化学院,南京210007 [2]中国电子科技集团公司第二十八研究所,南京210014

出　　处：《电子与信息学报》2005年第1期13-16,共4页Journal of Electronics & Information Technology

基　　金：国家部级基金资助课题

摘　　要：设计了一种仅仅使用两个模型实现对机动目标精确跟踪的多模型算法,采用了含有法向和切向加速度的加速度均值自适应的当前统计模型和扩展后的常速模型进行交互。该算法不受目标转弯率大小和变化的限制,对目标运动模式的未知参数变化的适应性较强。仿真结果表明,该算法对目标的跟踪精度明显优于传统的使用3个以上模型交互的IMM-CV/CT算法。由于本算法能够估计出目标的法向和切向加速度,进行适当的模型集设计后,可以实现对复杂、快速机动目标的全过程跟踪,具有可扩展性的应用前景。A new IMM algorithm for tracking maneuvering target is proposed, which only uses two models that consist of a CS model and an augmented CV model for interaction. The CS model which uses the state variable with normal acceleration and tangent acceleration is fit for the accurate tracking of coordinated turn movement. This method is not subjected to the limit of the magnitude and change of turn rate, and has more flexibility to uncertain parameters. The simulation shows that this algorithm is superior to the conventional IMM-CV/CT algorithm which uses more than three models for interaction in terms of the accuracy of tracking and adaptation. Because this algorithm can estimate the normal and tangent acceleration, it may be used to implement the tracking for the complexly and highly maneuvering target after appropriate model-set design.

关 键 词：多模型 目标跟踪 模型集设计 自适应估计 

分 类 号：TN953[电子电信—信号与信息处理]