基于BP神经网络的飞行训练品质评估

为了弥补现有飞行训练品质评估系统存在的不足,本文提出利用飞行员生理信号和飞行操作参数来构建评估指标体系。建立了基于BP神经网络的飞行员飞行训练品质评估模型,通过已完成训练的网络的权值分布计算出各输入指标对最后评分结果的影响,并通过算例分析检验了该模型的可靠性。检验结果表明:经BP神经网络模型训练得到的结果和样本的专家评分基本吻合,选取的评价指标有效,该评估模型能够有效地将飞行员生理信号与飞行训练参数相结合,对飞行员的飞行训练品质进行评价。

Quality assessment of flight training based on BP neural network

To overcome the existing shortage of the model for flight training assessment,we propose in this paper the meth-od that makes use of pilot physiological signals and flight parameters to construct the evaluation system based on the BP neu-ral network.The influence of the input indexes on the grading result is calculated according to the weight distributions of the trained network.The reliability of this model is tested by analysis of the numerical example.The validation result shows that the result obtained according to the model proposed is consistent with the sample results of expert evaluation.The proposed model can effectively combine pilot physiological signals with flight training parameters to evaluate the quality of pilot train-ing.