Unlocking air traffic flow prediction through microscopic aircraft-state modeling
Unlocking air traffic flow prediction through microscopic aircraft-state modeling
Authors: Bin Wang, Anqi Liu, Jiangtao Zhao, Yanyong Huang, Peilan He, Guiyuan Jiang, Feng Hong, Yanwei Yu, Tianrui Li Date: 2026-05-11 Paper ID: openalex:2605.10083
Summary
AeroSense is a novel modeling framework for air traffic flow prediction that replaces traditional aggregated time-series analysis with a state-to-flow mapping approach. By representing terminal airspace as a dynamic set of microscopic aircraft states derived from ADS-B trajectories, the framework preserves fine-grained kinematic and intent-based information. This approach allows the model to predict future traffic density directly from current instantaneous conditions without needing historical look-back windows, outperforming existing forecasting baselines.
Key Contributions
- Introduces AeroSense, a framework that maps microscopic aircraft state sets directly to future traffic flow.
- Demonstrates superior performance over aggregation-based time-series forecasting methods, especially in high-density traffic scenarios.
- Shows that instantaneous airspace snapshots can replace traditional historical look-back window requirements for accurate traffic prediction.
Open Questions & Future Work
Key Concepts
- state-to-flow-modeling: A modeling paradigm that predicts aggregate traffic flow by processing instantaneous, microscopic state representations of individual agents rather than historical aggregated time series.
Archivist Review
I approved the ‘state-to-flow modeling’ concept as it represents a significant methodological shift from traditional aggregated time-series forecasting. I also approved two open questions concerning horizon extension and multimodal data integration, as these address fundamental limitations in current state-based forecasting paradigms. I maintained a strict selection policy by only including high-level methodological shifts and research bottlenecks.
Approved Concepts
- State-to-flow modeling: Moves away from traditional aggregated time-series forecasting by directly mapping microscopic individual states to macroscopic flow.
Approved Open Questions
- Extended Air Traffic Forecasting Horizons: The current evaluation is limited to a 15-minute horizon; confirming performance at longer horizons is essential for strategic air traffic flow management.
- Multi-modal Air Traffic Prediction: Incorporating external variables is critical for robustness in volatile operational environments where state snapshots are insufficient.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2605.10083
- paper_id
- 2605.10083
- paper_source
- openalex
- domain
- time-series
- tags
- forecastingtime-series
- architectures
-
- datasets
-
- concept_slugs
- state-to-flow-modeling
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-05-14T06:10:14Z