Research on fine-grained aircraft taxiing time prediction and conflict avoidance strategies

• A novel strategy for aircraft taxiing trajectory extraction is proposed, leveraging the ADS-B dataset to accurately reconstruct aircraft ground movement paths. • A novel idea is proposed that integrates taxi time prediction with taxi route planning, providing a unified approach to improve airport surface operation efficiency. • The concept of key point delay along taxi routes is introduced, and a prediction model, Enhanced PatchTST for Node Arrival Time Prediction (EPNATP) is designed to estimate arrival times at critical nodes. • A conflict avoidance algorithm based on the Attention-guided Taxiing Conflict Optimization Model(ATCOM) is proposed to detect and avoid potential conflicts during the taxiing process. Taxiway conflict avoidance is a critical challenge for enhancing airport operational efficiency and safety, particularly with the rapid growth in air traffic and limited airport infrastructure. Although there are many aircraft taxi path planning methods, they often overlook a detail factor that is important to the results, the time required to reach each taxiway node. This study proposes an Optimization Model for Aircraft Taxiing Timing Prediction and Conflict Avoidance, which integrates three key components: Aircraft Taxiway Trajectory Mining, Enhanced PatchTST for Node Arrival Time Prediction Model (EPNATP), and the Attention-guided Taxiing Conflict Optimization Model (ATCOM). The trajectory mining component extracts key taxiway paths from historical Automatic Dependent Surveillance-Broadcast (ADS-B) data, providing valuable insights into aircraft movement patterns under various operational conditions. The EPNATP model then uses this information to predict taxi times at a fine-grained level, and ATCOM minimizes waiting times while eliminating potential conflict hotspots. Experiments conducted using real-world ADS-B data from Shanghai Pudong International Airport, China (January April 2024) show that the EPNATP model reduces mean absolute error by 12%-28.5% compared to existing methods. The ATCOM successfully resolves all conflicts, reducing taxi times by 20.74%-24.74% compared to traditional algorithms such as the improved ant colony algorithm. This research provides an effective solution for taxiway conflict avoidance and operational scheduling, significantly improving airport efficiency and ensuring safe taxiing operations.