下降进近阶段飞行安全分析与建模  

Analysis and modeling of flight safety in decent and approach

作　　者：李晓明[1] 曹力[1] 王旭辉 林静[1] 夏大旺 

机构地区：[1]南京航空航天大学民航学院,南京210016 [2]中国民航科学技术研究院,北京100028

出　　处：《信息技术》2015年第1期7-10,15,共5页Information Technology

基　　金：国家自然科学基金(61179066)

摘　　要：下降进近阶段,飞行姿态不当造成不稳定进近,容易引起飞行事故,准确控制飞机姿态变化能够避免事故的发生。通过分析飞机受力情况,确定了影响下降进近阶段飞行安全的因素,构架了包括飞行环境、飞行状态和飞行员操作等因素的"人-机-环"的综合分析环境,并对飞行环境进行了分类;在分类环境下,以QAR飞行数据为基础,建立了基于飞行状态与飞行员操作的神经网络"人-机"响应模型,优化了网络结构,并通过实验验证了模型能够有效预测飞机姿态变化。During descent and approach, improper aircraft attitude causes approach unstable, which leads to flight accidents. Accurate control of the variation of aircraft attitude can prevent such accidents. Factors which affect flight safety of descent and approach are conformed through aircraft＇s force analysis. An integrated analysis system of ＂ pilot-aircraft-environment＂ which includes flight environment, aircraft conditions and pilot operations is built and flight environment is classified into different kinds. In each kind of flight environment a ＂pilot-aircraft＂ response model of aircraft conditions and pilot operations is established with the neural network based on QAR flight data. And with the optimization for the network structure, the experiments show that the model can predict the variation of aircraft attitude effectively.

关 键 词：QAR 下降进近 飞行安全 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]