Related Experiment Video
Updated: May 12, 2026
16:14
Trajectory Data Analyses for Pedestrian Space-time Activity Study
Published on: February 25, 2013
13.5K
4D trajectory prediction and conflict detection in terminal areas based on an improved convolutional network
Xin Ma1, Linxin Zheng2, Xikang Lu2
1Nanjing University of Aeronautics and Astronautics, Nanjing, China.
Plos One
|February 14, 2025
Summary
A new CNN-BiGRU model improves Four Dimensional (4D) Trajectory prediction accuracy and conflict detection reliability for air traffic management. This advanced trajectory prediction enhances aviation safety and efficiency in busy airport terminal areas.
Area of Science:
- Aviation safety and air traffic management.
- Artificial intelligence in aerospace engineering.
- Deep learning for trajectory prediction.
Background:
- Increasing passenger traffic and expanding airline networks necessitate safer, more scientific air traffic service modes.
- The International Civil Aviation Organization (ICAO) introduced Trajectory Based Operation (TBO) to optimize air traffic services.
- Current methods face challenges in accurately predicting four-dimensional trajectories and detecting multi-trajectory conflicts.
Purpose of the Study:
- To enhance the accuracy of Four Dimensional (4D) Trajectory prediction.
- To improve the reliability of multi-trajectory flight conflict detection under Trajectory Based Operation (TBO).
- To develop and validate an advanced trajectory prediction model for air traffic management.
Main Methods:
- Developed a trajectory prediction model using Convolutional Neural Networks-Bidirectional Gated Recurrent Unit (CNN-BiGRU).
- Implemented conflict evaluation using a trajectory distance detection function.
- Validated the model using real Automatic Dependent Surveillance-Broadcast (ADS-B) historical track data from a busy airport terminal area.
Main Results:
- The CNN-BiGRU model demonstrated superior performance compared to single Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) models across multiple evaluation metrics.
- Accurate conflict detection was achieved for future 800-second intervals between two trajectories.
- The model effectively addresses the challenges of 4D trajectory prediction and conflict detection in complex air traffic scenarios.
Conclusions:
- The CNN-BiGRU model offers a significant advancement in 4D trajectory prediction accuracy and conflict detection reliability for Trajectory Based Operation (TBO).
- This approach provides a more scientific and safer optimization for air traffic service modes.
- The findings support the integration of advanced deep learning models in future air traffic management systems.
