AGENTIC AI & DIGITAL TWIN FOR AIRCARFT DISAPPEARANCE PREDICTION

The project "Agentic AI & Digital Twin for Aircraft Disappearance Prediction" has been successfully designed and developed as a robust, real-time solution for intelligent aviation investigation and search zone prediction. In an era where aviation safety and spectrum integrity are increasingly critical, this system provides a meaningful bridge between complex flight telemetry data and actionable investigative intelligence. By integrating a Monte Carlo stochastic flight simulation engine with Inmarsat BTO satellite arc constraints, the system effectively models hundreds of possible aircraft trajectories and identifies high-probability crash corridors. The Logistic Regression machine learning classifier demonstrates reliable performance in analysing telemetry parameters including fuel depletion, communication loss, route deviation, and weather risk to output accurate disappearance probability scores. The implementation of an interactive Flask dashboard equipped with Leaflet.js live maps, animated flight markers, confidence-based search zone overlays, Chart.js risk visualizations, and a 300-day ocean debris drift simulation ensures that complex aviation investigation data is accessible to both expert researchers and operational search teams. Overall, the Agentic AI & Digital Twin system stands as a scalable, open-source, and efficient tool addressing modern challenges in aviation accident investigation, maritime search optimization, and aerospace intelligence. The project successfully meets all its initial objectives, delivering a high-performance  investigative  platform  ready  for deployment in real-world aviation monitoring and search-and-rescue environments.

1.         The Search for Malaysia Airlines Flight MH370 – Journal of Navigation, Cambridge. Available: https://www.cambridge.org/core/journals/j ournal-of-navigation  2.         Ocean Surface Current Modeling and Debris Drift Prediction – Ocean Modelling, Sciencedirect. Available: https://www.sciencedirect.com/journal/oce an-modelling  3.         Scikit-Learn: Machine Learning in Python – Pedregosa et al., JMLR, 2011. Available: https://jmlr.org/papers/v12/pedregosa11a. html  4.         Monte Carlo Simulation Techniques – Environmental Modelling and Software, Sciencedirect. Available: https://www.sciencedirect.com/journal/en vironmental-modelling-and-software  5.         Stefan Brennsteiner, Towards Real-Time Machine Learning Enhanced Signal Detection in 5G-NR, University of Edinburgh, 2023. Available: https://era.ed.ac.uk/handle/1842/42067  6.         Flask Documentation. Available: https://flask.palletsprojects.com 7.         Scikit-Learn Documentation. Available: https://scikit-learn.org 8.         Leaflet.js Documentation. Available: https://leafletjs.com 9.         Pandas Documentation. Available: https://pandas.pydata.org  10.     NumPy Documentation. Available: https://numpy.org/doc