修正的马尔科夫转移矩阵自适应IMM算法  

Improved Adaptive Markov Matrix IMMAlgorithm

作　　者：李昊润 卜凡康 周剑雄[1] LI Hao-run;BU Fan-kang;ZHOU Jian-xiong(ATR Key laboratory,School of Electronic Science,National University of Defense Technology,Changsha 410000,China)

机构地区：[1]国防科技大学电子科学学院ATR重点实验室,长沙410000

出　　处：《火力与指挥控制》2021年第9期118-124,132,共8页Fire Control & Command Control

摘　　要：机动目标跟踪算法中,交互式多模型(IMM)算法是一种相当有效的方法。但该算法中存在Markov矩阵据先验知识给定且不变的问题。针对此问题,提出一种Markov矩阵自适应的IMM算法。该算法利用后验信息实时修正Markov矩阵,改善了位置和速度的滤波结果,使IMM算法的模型概率变化更加符合目标真实运动。经仿真验证,提出的算法具有更好的跟踪性能。Standard Interacting Multiple Model(IMM)algorithm is an effective approach of tracking maneuvering target.For the problem of constant Markov matrix in standard Interacting Multiple Model(IMM)algorithm,a new improved adaptive Markov matrix IMM algorithm is proposed.The proposed algorithm can modify the Markov transition probabilities adaptively through modifying posterior information,which makes the model probabilities changing more real and improves the filtering accuracy.The simulation results show that the proposed algorithm has better tracking performance compared with standard IMM algorithm.

关 键 词：交互式多模型(IMM) Markov矩阵 机动目标跟踪 后验信息 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]