BP神经网络在缺失数据估计中的应用

由于飞行参数记录系统所记录的数据很容易被污染,所以对飞参数据进行预处理已显得十分重要,而预处理的一项重要内容就是对缺失参数数据进行合理且有效地估计,真实地反映飞行器当时的状态.通过分析神经网络理论和飞参数据特征,提出了一种基于BP神经网络的缺失数据估计的方法,有效地解决了目前飞行参数记录系统记录数据时缺失数据的问题.利用某型飞机真实的数据进行仿真,结果表明了这种方法是可行且有效的.

Application of BP neural network in predicting absent data

The preprocess of data recorded by the flight recorder system became very important because of dirty flight data.The important aspect of preprocess is to predict absent data rationally and effectively,and reflect the states of aerocraft.With the analysis of the theory of neural network and the characteristics of flight data,the method of predicting absent data based on BP neural network is presented,and the problem that flight data recorder loses data is resolved effectively.The result of simulation of real flight data of certain plane shows that the method is validated and effective.