机场附近频繁变换场景下的6D飞行轨迹预测

6D flight trajectory prediction in scenarios with frequent scene changes near the airport

日益增多的航空活动对航空交通管制提出了挑战,轨迹预测技术在保障空中交通的安全和有序中发挥着重要作用.机场附近更加密集的航班给轨迹预测带来了困难.基于广播式自动相关监视系统,提出了一种基于卷积注意力(Attention-CNNs)的双向长短时记忆网络(BiLSTM)和极端梯度提升(XGBoost)的混合神经网络模型,能够对飞行轨迹的6D信息(时间、经度、纬度、高度、速度、航向角)进行预测.由包含空间位置和时间戳的时空信息和飞行动态信息组成的轨迹集用于证明该方法的效率.定量分析表明,所提出的模型在评价指标上的表现优于对比模型,为机场环境下航空管理系统的安全运行提供了有效方法.

The increasing volume of aviation activities poses challenges to air traffic control,with trajectory prediction technology playing a pivotal role in ensuring the safety and orderliness of airspace traffic.The heightened density of flights near airports presents difficulties for trajectory prediction.A hybrid neural network model based on Attention-CNNs,bidirectional long short-term memory(LSTM),and XGBoost is proposed using data from the automatic dependent surveillance-broadcast(ADS-B)system.This model is designed to forecast 6D information pertaining to flight trajectories,including time,longitude,latitude,altitude,velocity,and heading angle.A trajectory dataset composed of spatial-temporal information and flight dynamics data is used to validate the efficiency of our approach.Quantitative analysis reveals that the proposed model outperforms comparative models in terms of evaluation metrics.This method offers an effective solution for ensuring the safe operation of aviation management systems in airport environments.