Recognition of Scenarios for Civil Aviation Radiotelephony Based on Events

Civil aviation radiotelephony is the primary method for ensuring effective communication between air traffic and ground control. During various phases of flight, pilots and controllers engage in communication following specific procedures that are both specialized and context-sensitive. By analyzing the characteristics of phrases used at different stages, we can extract structured event information from unstructured instructions. This information, combined with advanced multi-label text classification techniques like Bidirectional Long Short-Term Memory (BiLSTM) and Graph Attention Network (GAT), helps identify the patterns and scenarios associated with civil aviation radiotelephony. This analysis aids in determining the dialogue modes during different flight phases, which is crucial for enhancing semantic understanding of subsequent communications and providing early warnings. Furthermore, the module assesses ATC instruction standardization and verifies pilot readback consistency.