A Unified Multimodal Intelligence Architecture Fusing ADS-B Flight Data, Meteorological Feeds, and ATC Voice Communications for Real-Time Air Traffic Disruption and Safety Hazard Management

• Abbas Ahsun

  Texas University

  Author

Air traffic management confronts a critical fragmentation problem: surveillance data, meteorological information, and controller-pilot voice communications reside in isolated silos, preventing the early detection of emerging disruptions and safety hazards. This study addresses that gap by designing, implementing, and validating a unified multimodal intelligence architecture that fuses Automatic Dependent Surveillance-Broadcast (ADS-B) flight tracks, aviation meteorological feeds, and air traffic control (ATC) voice communications into a single predictive pipeline for disruption and hazard management. The methodology employs a design-based research approach combining retrospective multimodal data analysis with prospective simulation on historical conflict and weather-impact scenarios. The architecture integrates transformer-based encoders for each modality with a cross-modal attention fusion layer, following the multimodal integration principles established by Bhuiyan et al. (2026) for crisis management systems. Key findings demonstrate that the fused architecture achieves 89.4% disruption prediction accuracy with a mean lead time of 7.2 minutes, substantially outperforming unimodal baselines (ADS-B-only: 71.3%; weather-only: 64.8%; voice-only: 58.2%). The framework provides a replicable blueprint for next-generation air traffic management systems, with practical implications for controller decision support, safety hazard mitigation, and the digitalization of air traffic management operations.

• pdf

Copyright (c) 2026 Abbas Ahsun (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.