多传感器融合系统的可靠性模型研究
李 军, 刘君华
(西安交通大学电气工程学院, 710049 , 西安)
摘要: 为有效评价多传感器融合系统硬件的可靠性,提出了一种新的三模冗余结构(TMR) 半马尔可夫过程模型. 该模型以系统单元寿命服从负指数分布、系统单元修复时间服从广义任意分布(修复率是时间的函数)为前提,通过引入补充变量,可以转化为马尔可夫过程模型进行求解,从而获取系统有效性和可靠性的一些重要指标. 理论分析和数值结果表明,传统可靠性建模的方法可以作为新模型简化的特例,利用新模型的TMR 冗余结构多传感器融合系统与无冗余的单传感器系统相比,系统的平均无故障时间(MTBF) 提高了两个数量级. 因此,提出的新模型可以准确地描述系统的实际工作状况,为有效评估多传感器融合系统可靠性提供了理论依据.
关键词: 多传感器融合系统;半马尔可夫过程;有效性;可靠性
中图分类号: TP27414 文献标识码: A 文章编号: 0253 – 987X(2004) 08 – 0775 – 04
Study of Reliability Model for the Multi2Sensor Fusion System
Li Jun , Liu Junhua
(School of Electrical Engineering , Xi′an Jiaotong University , Xi′an 710049 , China)
Abstract : A novel semi2Markov process model based on three2model redundant (TMR) architecture is proposed to effec2
tively evaluate the hardware reliability for multi2sensor fusion system. The semi2Markov process mode is established under
the condition that the natural life of each unit of the system is assumed to confirm wih a negative exponential distribution
and repair time of each unit with generalized arbitrary distribution where the repair rate is considered as a function of
time. Simultaneously , the novel model is converted to Markov process model by supplementary variable , and then some
universal reliability and availability indices of the system model are obtained. Theoretical analysis and numerical result
confirm that the traditional method can be regarded as a special case of the proposed model . Furthermore mean time be2
tween fail (MTBF) of the TMR system employing proposed model for the multi2sensor fusion two order of magnitude in2
creases over the single sensor system. The semi2Markov process model illustrates more accurately the operating character2
istics of the system and provides theoretical basis to evaluate effectivly the reliability of the multi2sensor fusion system.
Keywords : multi2sensor fusion system ; semi2Markov process ; availability ; reliability
多传感器融合系统是指对来自多个传感器的数据进行多级别、多方面、多层次的处理与综合,从而获得更精确、更可靠的有用信息与结果. 与只采用单一传感器的系统相比,来自多个传感器的信息具有冗余性、互补性、关联性,基于分布式多传感器的融合系统具有很强的容错性能,可靠性高. 评价多传感器融合系统的硬件可靠性对于多传感器融合系统的配置具有一定指导意义[1 – 3 ] . 本文对多传感器融合
系统硬件冗余配置的可靠性建模进行了研究,并针对典型的三模冗余(TMR) 结构配置,给出表征系统可靠性的半马尔可夫过程模型[4 ,5 ]及相应的可靠性指标,为该类系统的可靠性评价和建模提供理论指导.
1 系统模型建立
多传感器融合系统的硬件冗余结构如图1 所示. 在该传感器融合系统中,根据Marzullo[6 ]的结论,认为只要有一半以上的传感器正常工作,该系统就能得到准确的输出. 因此,该系统可靠度的模型为( N , N/ 2 + 1) 并联结构, 即在由这N 个传感器构成
的冗余系统中,当有N/ 2 + 1 个或N/ 2 + 1 以上个传感器正常工作时,系统就可以正常工作. 基于这一表
决系统模型,我们可以对系统的硬件可靠性进行分析.
以具有代表性的TMR (三取二) 冗余表决结构配置的多传感器系统为例,建立多传感器融合系统硬件可靠性的一种模型.
