Generation of flight samples and data sets for aircraft landing studies using machine learning techniques

Аuthors

Petrov I. A.^1*, Proshkina N. A.^2**, Strelkov V. V.^1***

1. Central Aerohydrodynamic Institute named after N.E. Zhukovsky (TsAGI), Zhukovsky, Moscow region, Russia
2. Moscow Institute of Physics and Technology (National Research University), 9, Institutskiy per., Dolgoprudny, Moscow region, 141701, Russia

*e-mail: peredreyhus1@mail.ru
**e-mail: proshkina.na@phystech.edu
***e-mail: v_strelkov@tsagi.ru

Abstract

Potentially available volumes of flight data and the trend of their multiple increase in the coming years indicate the relevance and importance of the development of new approaches to the study of flight safety problems, which are based on the analysis of information using machine learning (ML) methods. The reliability of the results of such analysis is largely determined by the quality of preprocessing of primary flight data and prepared arrays (datasets) for the direct application of ML methods. Data arrays are formed for a specific applied problem to be solved.
This paper deals with the formation of flight samples and data arrays for their subsequent analysis using ML methods as applied to the task of predicting the coordinate of an aircraft touchdown point on a runway.
The whole process of data preparation is divided into the following separate stages: collection, transcription and annotation of flight parametric recorder data; formation of flight sample; calculation of non-registered parameters; alignment of records of different flights; validation of recorded flight data; data consolidation and synchronization; preliminary (exploratory) data analysis.
Each of the steps is discussed in detail and illustrated by the results of real flight data processing.
The paper uses records of passenger aircraft flight parameters under operating conditions on the route network of one of the airlines, METAR archives, reference data on landing airfields and nominal technical characteristics of the aircraft.
It is shown that the work on the formation of a sample of flights and data arrays for the study of aircraft landing using machine learning methods is a very labor-intensive, but necessary part of the implementation of a promising approach to the study of flight safety problems.

Keywords:

flight safety, machine learning, prediction, flight data, flight sampling, dataset landing, touchdown point coordinate

References

1. G. van Es. A Study of Runway Excursions from a European Perspective. NLR-CR-2010-259. URL: https://skybrary.aero/sites/default/files/bookshelf/2069.pdf
2. Jenkins M., Aaron R.F. Reducing Runway Landing Overruns. Boeing Aero Magazine. URL: https://www.boeing.com/commercial/aeromagazine/articles/2012_q3/3/
3. Sharov V.D. Prognozirovanie i predotvrashchenie vykatyvanii samoleta za predely vzletno-posadochnoi polosy (Forecasting and Prevention of Aircraft Rolling Out of the Runway), Germaniya, Lambert Academic Publishing, 2013, pp. 112.
4. Rybin A.V. Trudy MAI, 2015, no. 81. URL: ttps://trudymai.ru/eng/published.php?ID=57761
5. Eremin A.I., Sel'vesyuk N.I. Trudy MAI, 2018, no. 100. URL: https://trudymai.ru/eng/published.php?ID=93451
6. Semakov S.L. Trudy MAI. 2019. № 104. URL: https://trudymai.ru/eng/published.php?ID=102188
7. Barry D., Greeves M., Solis T., Angel M. Advanced methods for analysis flight data for runway excursions risk factors. Cranfield University, Airbus Defence&Space. Future Sky Safety. Project: Solutions for Runway Excursions, 2017. URL: https://www.futuresky-safety.eu/wp-content/uploads/2018/03/FSS_P3_CU_D3.5_v2.0.pdf/
8. Kobzar' A.I. Prikladnaya matematicheskaya statistika (Applied mathematical statistics), Moscow, Fizmatlit, 2006, 813 p.
9. Onlain-uchebnik po mashinnomu obucheniyu ot ShAD. URL: https://education.yandex.ru
10. Lishuai Li R., John Hansman, Rafael Palacios, Roy Welsch. Anomaly detection via a Gaussian Mixture Model for flight operation and safety monitoring, Transportation Research Part C Emerging Technologies, 2016, vol. 64, pp, 45–57. DOI: 10.1016/j.trc.2016.01.007
11. Janakiraman V., Matthews B., Oza N. Discovery of Precursors to Adverse Events Using Time Series Data, SIAM International Conference on Data Mining, May 5-7, 2016, Miami, FL. DOI: 10.1137/1.9781611974348.72
12. Melnyk I., Banerjee A., Matthews B., Oza N. Semi-Markov Switching Vector Autoregressive Model-Based Anomaly Detection in Aviation Systems, ACM International Conference on Knowledge Discovery and Data Mining (KDD), 2016. DOI: 10.1145/2939672.2939789
13. Raphael Monstein, Benoit Figuet, Timothé Krauth, Manuel Waltert and Marcel Dettling. Large Landing Trajectory Dataset for Go-Around Analysis, Engineering Proceedings, 2022, vol. 28 (2). DOI: 10.3390/engproc2022028002
14. Burov Ya.V., Strelkov V.V. Svidetel'stvo o registratsii programmy dlya EVM № 2017613196, 13.03.2017.
15. Alberto Rizzoli. Training Data Quality: Why It Matters in Machine Learning, An Introductory Guide to Quality Training Data for Machine Learning. URL: https://www.v7labs.com/blog/quality-training-data-for-machine-learning-guide
16. Pobedinskii G.G. Sistemy koordinat v geodezii i ikh svyazi (Coordinate systems in geodesy and their relationships), Nizhnii Novgorod, Nizhegorodskii gosudarstvennyi arkhitekturno-stroitel'nyi universitet, 2023, 144 p.
17. Geraud de Rivals. Landing Trajectory Computation, 3rd conference of the European Operators Fligh Data Monitoring forum (EOFDM), Cologne, 06 February 2014. URL: https://www.pressreleasepoint.com/european-operators-flight-data-monitoring-eofdm
18. Emel'yanov D.S., Strelkov V.V., Gorskii E.B. XXX Nauchno-tekhnicheskaya konferentsiya po aerodinamike: tezisy dokladov. Zhukovskii, TsAGI, 2019, pp. 111-112.
19. Emel'yanov D.S. Programma vosstanovleniya traektorii dvizheniya samoleta na zaklyuchitel'nom etape snizheniya po glissade, vyravnivanii i dvizhenii po vzletno-posadochnoi polose (TrajectoryScan). Svidetel'stvo № 2018664260, 2018.
20. G. van Es, P. der Geest, Cheng A., Hackler L., Dillard A.E. A Study of Normal Operational Landing Performance on Subsonic, Narrow-Body Jet Aircraft during Instrument Landing System Approaches. Report No. DOT/FAA/AR-07/72007.

Download