基于HMM的低空目标航迹威胁识别  

Threat identification for low-altitude target track based on HMM

作　　者：余晓洁 魏嵩 盛佳恋 张磊 YU Xiaojie;WEI Song;SHENG Jialian;ZHANG Lei(School of Electronics and Communication Engineering,Sun Yat-sen University,Shenzhen 518107,China;Shanghai Radio Equipment Research Institute,Shanghai 201109,China)

机构地区：[1]中山大学电子与通信工程学院,广东深圳518107 [2]上海无线电设备研究所,上海201109

出　　处：《系统工程与电子技术》2023年第5期1399-1408,共10页Systems Engineering and Electronics

基　　金：航空科学基金(2019200M1001);广东省自然科学基金(2021A1515011979)资助课题。

摘　　要：稳健的低空目标威胁识别是低空域安全防护的重要任务。传统的多属性决策方法对目标运动参数的量测精度要求较高,忽略了目标运动的时序关联信息,在实际应用中缺乏噪声稳健性和动态分析能力。因此,在多属性决策方法的基础上引入了隐马尔可夫模型,提出了一种动态稳健的低空目标威胁等级识别方法。通过建立隐状态与威胁等级、威胁数值之间的内在联系,将威胁识别问题转化为隐马尔可夫模型的状态解码问题。相比于常规算法,所提方法能够有效地抑制量测噪声干扰并具有一定的威胁预测能力。仿真实验验证了所提方法的有效性和稳健性。Robust low-altitude target threat identification is an important task of low-altitude security protection.The traditional multi-attribute decision-making(MADM)method requires high measurement accuracy of target motion parameters,neglects the temporal correlation information of target motion,and lacks noise robustness and dynamic analysis ability in practical applications.Therefore,the hidden Markov model is introduced on the basis of the MADM method,and a dynamic and robust method for low-altitude target threat level identification is proposed.By establishing the internal relationship between hidden state,threat level and threat value,the problem of threat identification is transformed into the problem of state decoding of the hidden Markov model.Compared with conventional algorithms,the proposed method can effectively suppress measurement noise interference and has certain threat prediction ability.Simulation experiments verify the effectiveness and robustness of the proposed method.

关 键 词：低空目标 威胁识别 多属性决策 隐马尔可夫模型 动态识别 

分 类 号：TN959.17[电子电信—信号与信息处理]