The accelerating growth of global air traffic is widening the gap between traditional ATM tools and the realtime, data-intensive decisions modern control towers must make. We introduce AEROGRAM11AEROGRAM is available as open-source software here.22A Demo video can be found here., an open-source artefact that merges a combination of LSTM and GMM capacity predictor with a dashboard-driven MAPE-K adaption loop. Developed and calibrated for Amsterdam Schiphol Airport, AEROGRAM continuously ingests live ADS-B, A-SMGCS and METAR feeds, evaluates three interchangeable strategies (rule-based baseline, pattern-based GMM, deep LSTM) and surfaces rerouting advice, delay forecasts and uncertainty thresholds in an interactive GUI. Experimental results on Schiphol traffic scenarios show that the LSTM based adaptive strategy cuts average delay by 33 % and sustains 85-90% efficiency during peak hours, while the GMM alternative delivers moderate gains with half the compute footprint and the baseline remains lightweight but least effective.
Susanu, C., Raibulet, C., Gerostathopoulos, I. (2025). AEROGRAM: Adaptive Environment & Rerouting Optimiser with GMM-Augmented LSTM Airspace Model. In 2025 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C) (pp.239-244). Institute of Electrical and Electronics Engineers Inc. [10.1109/ACSOS-C66519.2025.00068].
AEROGRAM: Adaptive Environment & Rerouting Optimiser with GMM-Augmented LSTM Airspace Model
Raibulet C.
Co-primo
;
2025
Abstract
The accelerating growth of global air traffic is widening the gap between traditional ATM tools and the realtime, data-intensive decisions modern control towers must make. We introduce AEROGRAM11AEROGRAM is available as open-source software here.22A Demo video can be found here., an open-source artefact that merges a combination of LSTM and GMM capacity predictor with a dashboard-driven MAPE-K adaption loop. Developed and calibrated for Amsterdam Schiphol Airport, AEROGRAM continuously ingests live ADS-B, A-SMGCS and METAR feeds, evaluates three interchangeable strategies (rule-based baseline, pattern-based GMM, deep LSTM) and surfaces rerouting advice, delay forecasts and uncertainty thresholds in an interactive GUI. Experimental results on Schiphol traffic scenarios show that the LSTM based adaptive strategy cuts average delay by 33 % and sustains 85-90% efficiency during peak hours, while the GMM alternative delivers moderate gains with half the compute footprint and the baseline remains lightweight but least effective.| File | Dimensione | Formato | |
|---|---|---|---|
|
Susanu-2025-ACSOS 2025-AAM.pdf
accesso aperto
Tipologia di allegato:
Author’s Accepted Manuscript, AAM (Post-print)
Licenza:
Licenza open access specifica dell’editore
Dimensione
2.56 MB
Formato
Adobe PDF
|
2.56 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


