The document presents a system for real-time tracking of aircraft in crowdsourced air traffic networks. The system utilizes data from the OpenSky Network to track aircraft locations reported by onboard GPS sensors. It develops models to predict aircraft trajectories and times of arrival based on historical data and weather impacts. The system is built on Apache Spark for real-time and historical data analysis. It includes APIs for flight tracking and predicting missing location data using simple localization estimates. The goals are to enable cleaner, safer and more efficient air traffic control.
Real-Time Tracking of Aircrafts in Crowdsourced Air Traffic Networks with Simple Localization Estimates
1. Company
LOGO
Real-Time Tracking of Aircrafts in
Crowdsourced Air Traffic Networks
with Simple Localization Estimates
Abderazek Chakka Bannour -LaTICE Lab.
Rim Moussa -LaTICE Lab.
Tarek Bejaoui -MEDIATRON Lab. SupCom
7th
IEEE International Symposium on
Networks, Computers and Communications
(ISNCC'20)
Montreal, Canada
2. 2
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Outline
● Context
● Goals
● OpenSky Network
● Architecture
● Excerpts from ground sensors data
● Excerpts from flights logs
● A Big Data Architecture for Air Traffic Control
● Flight Tracking API
● Prediction of missing geo-data API
● Related Work Survey
● Conclusion and Future Work
3. 3
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Context
● (2019) The aviation industry supports
● 87.7 million jobs around the world. Some of these roles are
within the industry itself, at airports, for airlines, and in civil
aerospace and air navigation services.
● $3.5 trillion (4.1%) of the world's gross domestic product.
● Multiple groups promote the transformation of aviation into cleaner,
safer, more efficient and predictable system, such as
● High Level Group on Aviation Research Europe Commission in
2011: European Aviation Vision 2050
● SESAR (Europe) in 2017 DART project (abrev. Data-driven
AiRcraft Trajectory prediction research)
● Next Generation Air Transportation (2012-2025) USA
4. 4
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Goals:
Cleaner, Safer and more Efficient Air Traffic Control
● Track in real-time aircrafts and know where an aircraft is at
any given time,
● Learn aircrafts’ trajectories and impacts of weather data
(winds, rains, fogs, storms, thunders et cetera) on aircrafts’
trips,
● Plan aircrafts’ routes and trips schedules,
● Predict times of arrivals of flights accurately,
5. 5
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
OpenSky Network
– ADS-B
● https://opensky-network.org/
● A crowd-sourced network
● Aircrafts simply report their
exact locations
(determined with on-board
GPS sensors) to ground
stations periodically.
8. 8
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
A Big Data Architecture for Air Traffic Control
● We use Apache Spark (PySpark) for both
●Historical data analysis
●Real-time data analysis
9. 9
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Flights Tracking API
● Allows to track flights in real time
● Calculate flights
patterns from historical data
10. 10
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Simple Prediction Model for Missing Coordinates
API
● Given
● B geo-coordinates are,
11. 11
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Bearing Calculus
● Given
● Bearing is computed as follows,
12. 12
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Aircraft Localization Estimates
--derived data
13. 13
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Aircraft Localization Estimates
--Map
Real aircraft in blue
Predicted position of the aircraft is in red
14. 14
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
zooming
Real aircraft in blue
Predicted position of the aircraft is in red
15. 15
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Error Information over 150 missing geo-data
attributes
Real coordinates Predicted coordinates
16. 16
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Related Work Survey
● Doshi and Palacios (2005) chronicle the design, training,
performance, and analysis of a position prediction neural network,
and present the resulting optimal neural network for the Airport
Movement Area Safety System in cooperation with the Federal
Aviation Administration on airport ground.
● Ayhan and Samet (2016) describe a stochastic trajectory prediction
approach for Air Traffic Management.
● They consider the airspace as a 3D grid network, where each
grid point is a location of a weather observation (temperature,
wind speed, and wind direction). Each cube is defined by its
centroid (latitude, longitude, altitude).
● Raw trajectories are aligned to a set of cube centroids which are
basically fixed 3D positions independent of trajectory data.
● They propose a Hidden Markov Model (HMM), to predict
trajectories taking environmental uncertainties. Experimental
study focuses on a ground-based tactical trajectory prediction
system.
17. 17
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Related Work ctnd.1
● Hamed et al. (2016) propose to predict intervals rather than precise
aircraft positions.
● They implement a standard point-mass model and statistical
regression, neural networks and Loess methods to predict the
altitude of aircrafts at take-off.
● They find out that the prediction intervals obtained by the point-
mass model are less reliable than statistical regression, neural
networks and Loess methods.
● Strohmeier et al. (2018) compare multilateration to grid-based
localization approach based on k-NN approach.
● Authors show that the grid-based k-NN approach is precise and
can increase the effective air traffic surveillance coverage
compared to multilateration by a factor of up to 2.5.
18. 18
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Related Work ctnd. 2
● Liu and Hansen (2018), propose a tree-based matching algorithm
to construct image-like feature maps from meteorological datasets.
● They model the track points on trajectories as conditional
Gaussian mixtures with parameters learned from an end-to-end
convolutional recurrent neural network.
● Convolutional layers are integrated into the pipeline to learn
representations from the high-dimension weather features.
● During inference process, beam search, adaptive Kalman filter,
and Rauch-Tung-Striebel smoother algorithms are used to prune
the variance of generated trajectories.
19. 19
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Conclusion and Future Work
● In this research work, we succeed to:
● Propose and design an architecture for Big Data system dealing
with flights logs and flights streaming data based on PySpark
● implement multiple prospective analytics business queries, such
as airport congestion,flight frequency . . .
● implement business queries solutions for tracking aircrafts, as
well as for predicting an aircraft missing geo-locations, in real-
time using a sound model navigation on estimate
● Future work will address the Investigation of other machine learning
algorithms such as
● Logistic regression
● Hidden Markov Chains
● Naive Bayes
● Deep Learning -Recurrent Neural Networks
as well as Large scale experiments
20. 20
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
References
● M. Dareck, C. Edelstenne, T. Enders, E. Fernandez, J.-P. Herteman, M.
Kerkloh, I. King, P. Ky, M. Mathieu, G. Orsi, G. Schotman, C. Smith, and
J.-D. Worner, “FlightPath 2050: Europe’s Vision for Aviation -Maintaining
Global Leadership and Serving Society’s Needs,”http://www.sesarju.eu/,
2010, online; accessed 10 August 2020.
● SESAR, “SESAR 2020,” http://www.sesarju.eu/, online; accessed 10
August 2020.
● ——, “Final Project Results Report - DART,” https://sesarju.eu/node/3179,
2019, online; accessed 10 August 2020.
● European Union and EuroControl and SESAR, “The DART Project: Data-
Driven Aircraft Trajectory Prediction Research,” http://dart-research.eu/,
online; accessed 10 August 2020.
● M. Schafer, M. Strohmeier, V. Lenders, I. Martinovic, and M. Wilhelm,
“Bringing up opensky: a large-scale ADS-B sensor network for research,”
in IPSN’14, Proceedings of the 13th International Symposium on
Information Processing in Sensor Networks (part of CPS Week), April 15-
17, 2014, Berlin, Germany, 2014, pp. 83–94.
21. 21
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
References
● US NextGen, “Modernization of United States Airspace,”
https://www.faa.gov/nextgen/, 2019, online; accessed 10 August 2020.
● Mattias Schaffer and Vincent Lenders and Ivan Martinovis, “OpenSky
Network: Open Air Traffic Data for Research,” https://opensky-
network.org/, online; accessed 10 August 2020.
● M. Schafer, M. Strohmeier, V. Lenders, I. Martinovic, and M.
Wilhelm,“Demonstration abstract: Opensky: a large-scale ADS-B sensor
network for research,” in IPSN’14, Proceedings of the 13th International
Symposium on Information Processing in Sensor Networks (part of CPS
Week), April 15-17, 2014, Berlin, Germany, 2014, pp. 313–314.
● A. Doshi, “Aircraft position prediction using neural networks,” Ph.D.
dissertation, Massachusetts Institute of Technology. Dept. of Electrical
Engineering and Computer Science, Newark, may 2005.
● S. Ayhan and H. Samet, “Aircraft trajectory prediction made easy with
predictive analytics,” in Proceedings of the 22nd ACM SIGKDD
International Conference on Knowledge Discovery and Data Mining, San
Francisco, CA, USA, August 13-17, 2016, 2016, pp. 21–30.
22. 22
7Th
IEEE Intl. Symposium on Networks, Computers and Com. @ Montreal, Canada 2020
Conclusion and Future Work
● M. G. Hamed, R. Alligier, and D. Gianazza, “High confidence intervals
applied to aircraft altitude prediction,” IEEE Trans. Intelligent
Transportation Systems, vol. 17, no. 9, pp. 2515–2527, 2016.
● M. Strohmeier, I. Martinovic, and V. Lenders, “A k-NN-based localization
approach for crowdsourced air traffic communication networks,” IEEE
Trans. Aerospace and Electronic Systems, vol. 54, no. 3, pp. 1519–1529,
2018.
● Y. Liu and M. Hansen, “Predicting aircraft trajectories: A deep generative
convolutional recurrent neural networks approach,” CoRR, vol.
abs/1812.11670, 2018. [Online]. Available: http://arxiv.org/abs/1812.11670
● Mattias Schaffer and Martin Strohmeier, “OpenSky Workshops ,”
https://workshop.opensky-network.org/, online; accessed 10 August 2020.
● R. Moussa, “Scalable maritime traffic map inference and real-time
prediction of vessels’ future locations on apache spark,” in Proceedings of
the 12th ACM International Conference on Distributed and Event-based
Systems, DEBS 2018, Hamilton, New Zealand, June 25-29, 2018, 2018,
pp. 213–216.
23. Company
LOGO Thank you for your Attention
Q & A
Real-Time Tracking of Aircrafts in Crowdsourced Air
Traffic Networks with Simple Localization Estimates
A. Chakka Bannour, R. Moussa and T. Bejaoui
7th
IEEE International Symposium on
Networks, Computers and Communications
(ISNCC'20)
Montreal, Canada