基于STC-IMM结构的自适应多模型跟踪算法  

Adaptive multiple-model tracking algorithm based on STC-IMM structure

作　　者：周政[1] 刘进忙[1] 李振兴[1] 

机构地区：[1]空军工程大学防空反导学院,西安710051

出　　处：《控制与决策》2013年第8期1226-1230,共5页Control and Decision

基　　金：国家自然科学青年基金项目(61102109);陕西省自然科学基金项目(2010JM8013)

摘　　要：针对机动目标跟踪问题,基于转换时间条件交互多模型(STC-IMM)结构,提出一种转换概率自适应的STC-AIMM算法.该算法根据滤波器收敛时间预设了模型转换时间条件,保证了滤波器对目标后验状态的合理逼近,同时通过模型转换概率的自适应算法实现了模型与目标运动模式的实时最优匹配.理论和仿真分析结果表明:相比交互多模型(IMM)算法和STC-IMM算法,该算法能够发挥滤波器最优性能,实现模型概率的优化分配,对目标不同强度的机动具有良好的适应性、跟踪稳定性和更高的跟踪精度.An interacting multiple-model algorithm with switch time conditions based on adaptive transition probabilities is proposed for tracking maneuvering targets, which ensures that the filter approximates target posterior state reasonably by presetting switch time conditions of the model. The model can match the target motion well in real time by using transition probabilities adaption algorithm. The theoretic and simulation analysis show that the proposed algorithm can help the filter achieving optimal performance, make the model probability more reasonable and track the target more accurately than IMM and STC-IMM algorithms. At the same time, it performs well on the stability and adaption.

关 键 词：机动目标跟踪 多模型 转换时间条件 转换概率 自适应估计 

分 类 号：TP953[自动化与计算机技术]