基于HMM的多态系统状态识别模型研究  

Study of Multi-State System States Recognition Model Based on HMM

作　　者：黄景德[1] 郝学良[1] 王明[2] 

机构地区：[1]海军大连舰艇学院舰炮火控教研室,辽宁大连116018 [2]中国白城兵器试验中心,吉林白城137001

出　　处：《测试技术学报》2012年第2期154-157,共4页Journal of Test and Measurement Technology

基　　金：国家自然科学基金资助项目(60971092)

摘　　要：在分析装备三级工作模式基础上,建立了多态系统的可靠性向量模型,基于隐马尔可夫模型(HMM)原理分析了多态系统的状态转移过程,并建立了多状态系统HMM模型,在此基础上利用MATLAB对系统隐藏状态转移和观察状态之间的关系进行了仿真.仿真结果表明,在一定观察序列的情况下,该模型可以实现较高的状态识别率和状态预测精度,为科学合理地诊断多状态系统的潜在故障提供了技术支持,具有重要的理论意义和应用价值.Based on the analysis of equipped with three operating modes, a multistate system reliability vector model is established. According to Hidden Markov Model （HMM） theory the state transition process of the multi - state system is analyzed, and a HMM model of the multstate system is built up. On this basis, the relation between hidden state transition and observation state is simulated by using MATLAB. Simulation results show that, in the case of a certain observation sequence, the model can achieve higher recognition rate and the state prediction accuracy and provide a technical support for the scientific and rational diagnosis of po- tential faults of multistate system.

关 键 词：多状态 状态识别 HMM 可靠性分析 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]