基于隐马尔可夫模型的切换飞行控制系统性能分析

摘要

该文研究了一种由电磁环境对电子设备产生的数字干扰来驱动的切换飞行控制系统性能分析模型.其中,从电磁干扰的产生原理角度,采用隐马尔可夫模型(Hidden Markov Model,HMM)描述数字电磁干扰特性并对其进行建模分析,同时针对HMM参数训练算法存在对初值选择敏感的问题,提出一种快速的初值选择策略,可以在经典Baum-Welch算法迭代下达到指定的收敛精度.最后将HMM电磁干扰注入分布式飞行控制系统性能观测平台,从理论与仿真的角度对比了不同电磁环境下分布式飞行控制系统的性能下降情况.仿真实验表明:与已有的数字电磁干扰建模分析方法相比,HMM具有更高的准确度,并且仿真所得性能下降程度与理论分析一致.%This paper proposes the performance analysis model of switched flight control systems driven by the digital upsets when electronics devices are subject to electromagnetic environments. Hidden Markov Model (HMM) is used to describe the characteristics of digital upsets and construct the model based on the theory of the electromagnetic interferences. The parameter estimation algorithms of the traditional training method for HMM are sensitive to initial parameters, therefore, this paper proposes a fast initial parameter selection strategy which can also accelerate the training processes. At the end, HMM-based electromagnetic interferences are fed to the performance observation platform for the distributed flight control systems. This paper also compares multiple performance degradation results under different electromagnetic fields from theory and simulation perspectives. Simulation results demonstrate HMM model can characterize the digital electromagnetic upsets more accurately compared to the existed digital electromagnetic models, and simulation results of the corresponding performance degradation are consistent with the theoretic results.