基于机器学习的离港航空器滑出时间预测

Departure Flights'Taxi-out Time Prediction Based on Machine Learning

准确地预测航空器的滑出时间对于提升机场场面运行安全和效率至关重要。基于机场场面运行态势分析,获得进/离港航空器滑行的时空分布特征,从而准确定义同时段离港航空器数量、进港航空器数量、起飞队列长度。基于影响因素进行相关性结论分析,构建了基于机器学习的航空器滑出时间预测模型,并使用中南某枢纽机场2周的实际运行数据对模型进行了验证。结果表明:滑出时间影响因素相关性大小排序为:起飞队列长度、同时段起飞航空器数量、30 min平均滑出时间、同时段落地航空器数量、起飞使用跑道、滑出距离。机器学习方法能实现对航空器滑出时间的有效预测,分类器的优劣排序为支持向量机(support vector machine,SVM)、BP(back propagation)神经网络、随机森林(random forest,RF)。引入弱相关的影响因素后,滑出时间预测精度会有一定程度的降低。

Accurately predicting the taxi out time of departure flights is very important to improve the efficiency and safety of airport surface operation.Based on the analysis of airport surface operation situation,the space-time distribution characteristics of departure flights and inbound flights were obtained,so as to accurately define the number of departure flights,the number of inbound flights and the length of departure queue.Based on the conclusion of correlation analysis of influencing factors,a prediction model of taxi-out time based on machine learning was constructed,and the model was verified by using the two-week actual operation data of a hub airport in central south China.The results show that the order of the correlation of the factors affecting the taxi out time is:the length of the takeoff queue,the number of flights taking off at the same time,the average taxi out time in half an hour,the number of flights at the same time,the runway used,and the taxi out distance.The machine learning method can effectively predict the taxi-out time.The advantages of the classifier are support vector machine(SVM),back propagation(BP)neural network and random forest(RF).The prediction accuracy of taxi-out time will be reduced to a certain extent after introducing weak correlation factors.