基于RBF神经网络与Markov组合的飞行风险预测研究

为了更加准确地对飞机飞行安全性做出评估,在径向基(radial basis function,RBF)神经网络的基础上,通过引入马尔可夫链的方式进行误差修正,建立了一种拟合程度高且无需反复调整权值的新型评估模型.并以国内某航空公司A320机型近20年来发生的飞行机械故障为基础数据对模型进行训练、拟合、修正.将修正结果与单一RBF神经网络评估方法相比较,分析二者差异后得出误差降低的结论.为管控飞机飞行风险提供了方法拓展.

Flight risk prediction based on RBF neural network and Markov chain

In order to evaluate the flight safety of aircraft more accurately, a new evaluation model with high fitting degree and no repeated adjustment of weight is established on the basis of radial basis function (RBF) neural network and through error correction by introducing Markov chain. The model is trained, fitted and corrected based on the data of the flight mechanical failure of the A320 aircraft of domestic airline in the past 20 years. Comparing the correction results with the single RBF neural network evaluation method, the conclusion of error reduction is obtained after analyzing the difference between the two, providing a way to expand the control of aircraft flight risk.