基于贝叶斯准则的航天器故障判决方法  

Spacecraft Error Judgment Method Based on Bayes Judgment Rule

作　　者：徐冰霖[1] 李战怀[1] 

机构地区：[1]西北工业大学计算机学院,陕西西安710072

出　　处：《计算机仿真》2013年第9期117-120,共4页Computer Simulation

基　　金：国家自然科学基金资助项目(51005179)

摘　　要：研究航天器故障准确识别问题,由于多种因素干扰造成的航天器遥测参数观测值误差,导致故障诊断系统误判航天器状态。如何正确地从噪声信号中提取出关键的故障信号,并且根据故障发生概率及影响因素对故障进行判决,是目前航天器故障诊断领域面临的一大难题。在可获得先验的故障概率和代价系数的条件下,通过对高斯环境中故障信号误差的分析,建立了错误判决代价的贝叶斯准则判决方法,设计了高斯白噪声情况下采用贝叶斯准则进行故障判决的算法。通过某航天器实际遥测数据的实验结果表明,改进方法可显著提高系统的故障诊断准确性。There are possibly many errors in the telemetry because of a lot of the influence factors,which maybe make mistakes in the fault judgment of the spacecraft.How to separate the fault patterns from the mix signals and judge the faults according to the fault probability and influence is the key problem in the spacecraft fault diagnosis.In this paper,the fault signals were analyzed,and Bayes judgment rules based on calculating the error judgment costs were presented.The simulation results show that,given the fault probability and cost coefficients,the proposed method can improve the performance of fault diagnose system.

关 键 词：数据处理 贝叶斯准则 故障判决 错误代价 

分 类 号：V411.8[航空宇航科学与技术—航空宇航推进理论与工程]