基于数字地形的机场终端区离场航迹预测

Digital Terrain- based Trajectory Prediction for Terminal Area Departures

高效准确的航迹预测能够掌握飞机的运行轨迹,是空管自动化与智能化领域的关键要素,旨在提高空中交通的运行能力和可预测性。针对高原机场终端区内离场航空器运行,通过挖掘大量历史数据的特征关系,提出基于长短时记忆(Long Short-Term Memory,LSTM)神经网络的航迹预测模型。此外,考虑高原机场复杂的地形环境使得离场航空器安全运行条件更为严苛,引入支持向量回归(Support Vector Regression,SVR)算法建立数字地形模型,得到离场航迹上经纬度所对应地形高度的剖面图。采用拉萨贡嘎国际机场的真实离场航迹与地形数据进行实例验证,结合地形条件对所预测离场航迹的安全性进行评估。实验结果表明:基于LSTM神经网络建立的离场航迹预测模型具有较高的精度,且离场航空器能够实现满足最小超障余度的安全运行。

Efficient and accurate trajectory prediction is a key element in the field of ATC automation and intelligence.It′s aimed at improving the operational capability and predictability of air traffic.A trajectory prediction model based on LSTM neural networks is proposed for the scenario of departing flight operations in highland airports.In addition,considering the complex terrain conditions of highland airports which make the safe operation of departing aircraft more demanding,the SVR algorithm was introduced to build a digital terrain model to obtain a profile of the terrain height.This is followed by an example validation using real departure trajectory and terrain data from Lhasa Kongga International Airport to assess the safety of the predicted departure trajectory in relation to the terrain conditions.The experimental results show that the prediction model built by LSTM has a good accuracy and the departing aircraft is guaranteed to meet the minimum overrun margin for safe operation.