民用飞机着陆距离预测研究

Prediction of landing distance for civil aircraft

为防止飞机着陆时冲出跑道,采用支持向量机(SVM)模型预测飞机着陆距离。基于机场、气象以及飞机自身等3方面影响因素,选取B737-800为参考机型。利用波音公司的LAND软件采集相关运行数据。通过选择误差最小、精度最优的径向基核函数(RBF)构建最有效的SVM模型。探讨网格参数算法、遗传算法(GA)和粒子群优化(PSO)算法对最佳惩罚函数c和核函数参数g的影响。结果表明,预测着陆数据与实测着陆数据吻合较好——最大绝对误差在20 m范围内,最大相对误差为1%。

In order to prevent aircraft from running out of runway,the paper was aimed at predicting the aircraft landing distance by means of an SVM model. B737-800 was taken as the reference type on the basis of considering specific factors influencing the distance,namely those relating to the airport,the weather and the aircraft. The operation data were collected by using Boeing LAND software. The radial basis function（ RBF） kernel function was chosen by selecting the minimum error and the optimal accuracy. The best penalty function c and the kernel function parameter g were optimized by using grid parameters,genetic algorithm and particle swarm optimization algorithm. The results show that the prediction of landing distance conforms with the measured data,the maximum absolute error is 20 meters,and the maximum relative error is 1%.