摘要： 为提高交互式多模型(IMM)算法对机动目标的估计精度，需要增加其模型数量，但模型过多将导致计算量大并降低估计器性能。针对上述问题提出一种基于模型集的双马尔可夫多假设IMM机动目标跟踪算法。该算法用模型集间的马尔可夫转移阵描述模型集之间的大跳变，用模型的马尔可夫转移阵描述模型集内各模型间的小跳变或慢变，以达到细化建模、提高滤波精度的目的。

关键词: 目标跟踪, 交互式多模型(IMM), 模型集, 双Markov切换, 鲁棒性

Abstract: To improve estimation precision for maneuvering target of Interacting Multiple Model(IMM) algorithm, more models are demanded while too many models tend to increase of computation and decrease of estimator performance. Aiming at the problem, this paper proposes a double Markov Model Set based Multiple Hypothesis IMM(MS-MHIMM) maneuvering target tracking algorithm. This algorithm uses Markov transfer matrix between model sets to describe large hops between model sets, adopts Markov transfer matrix of model to describe small hops or slow hops between models in the same model set. Modeling is refined and filtering precision is improved.

Key words: target tracking, Interacting Multiple Model(IMM), model set, double Markov switch, robustness

中图分类号: 

• TN911.72