پژوهشنامه حمل و نقل

پژوهشنامه حمل و نقل

طراحی مدلی هوشمند جهت پیش‌بینی ریسک ایمنی پرواز- فاز اپروچ با استفاده از الگوریتم ‌BI.M-LSTM

نوع مقاله : مقاله پژوهشی

نویسندگان
منصور یحیوی 1
عباس طلوعی اشلقی 2
محمدعلی افشار کاظمی 2
رضا رادفر 3
1 دانشجوی دکتری، دانشکده مدیریت، واحد تهران مرکزی، دانشگاه آزاد اسلامی، تهران، ایران
2 استاد، دانشکده مدیریت و اقتصاد، دانشگاه علوم و تحقیقات‌ دانشگاه آزاد اسلامی، تهران، ایران
3 استاد، گروه مدیریت تکنولوژی، دانشگاه آزاد اسلامی، واحد علوم تحقیقات تهران، تهران، ایران
10.22034/tri.2025.466427.3263
چکیده
مقاله حاضر، مدلی نوآورانه (BI.M-LSTM) ترکیبی از الگوریتم (BI)‌ و شبکه عصبی بازگشتی (LSTM) جهت پیش‌بینی ریسک ایمنی پرواز فاز اپروچ ارایه می‌دهد. فاز اپروچ با سهم 3 درصد از فرایند کل پرواز به عنوان خطرناک‌ترین مرحله هر پرواز است. روش پیشنهادی شامل آموزش شبکه‌های عصبی نظارت شده برای برآورد پارامترهای هدف است. بدین منظور از دیتاست استاندارد مربوط به سال ۲۰۱۹ تا ۲۰۲۰ پس از خلاصه‌سازی، پاکسازی، نرمال‌سازی تعداد ۲۸۸۱۳ رکورد مربوط به پارامترهای ریسک ایمنی، مانند شرایط آب و هوایی، پیکربندی هواپیما، اطلاعات پرواز، سرعت، ارتفاع و ترافیک هوایی انتخاب شد. به علت وابستگی داده‌های پرواز به ورودی‌های ما قبل خود و نیاز به نوعی حافظه، آموزش توسط الگوریتم (LSTM) در محیط پایتون انجام گرفت. پس از یادگیری، میانگین خطای مجذور انحرافات حدود 38/6درصد بدست آمد. نتیجه نشان داد، درصد خطا قابل اغماض و مدل پیشنهادی نسبت به مدل‌های مشابه از اعتبار بالایی برخوردار است. این مدل به دلیل برخورداری از ابزارهای پیشرفته از جمله ETL، متادیتا و مانیتوریگ لحظه‌ای مشکل اکتشاف و پاکسازی انبوه داده‌های پرواز را حل کرد و توانست مهم‌ترین عامل ریسک ایمنی فاز اپروچ یعنی کنترل سرعت و ارتفاع لندینگ را با دقت بالا پیش‌بینی‌ کند. این الگو با راهبردی قابل اعتماد به خدمه پرواز در راستای کنترل پارامترهای مهم ریسک ایمنی از جمله، از دست رفتن کنترل پرواز ، سرعت هواپیما، موقعیت تاچ داون و کنترل جلوگیری از خروج هواپیما از باند کمک می‌کند.
کلیدواژه‌ها
ریسک ایمنی پرواز
حمل نقل هوایی
لندینگ
مدل‌
یادگیری عمیق
BI.M-LSTM

موضوعات

عنوان مقاله English

Design of an Intelligent Model for Predicting Flight Safety Risk in the Approach Phase Using the BI.M-LSTM Algorithm

نویسندگان English

Mansour Yahyavi 1
Abbas Toloie Ashlaghi 2
Mohammad Ali Afshar Kazemi 2
Reza Radfar 3
1 Ph. D., Student, Department of Information Technology Management, Central Tehran Branch, Islamic Azad University, Tehran, Iran‌.
2 Professor, Department of Industrial Management, Science and Research Branch, Azad Islamic University, Tehran, Iran.
3 Professor, Department of Technology Management, Faculty of Management and Economic, Science and research branch, Islamic Azad University, Tehran, Iran.
چکیده English

The present article introduces an innovative model (BI.M-LSTM), which combines the BI algorithm and the Long Short-Term Memory (LSTM) neural network to predict flight safety risks during the approach phase. The approach phase, accounting for 3% of the total flight process, is considered the most dangerous stage of any flight. The proposed method involves training supervised neural networks to estimate target parameters. For this purpose, a standardized dataset from the years 2019 to 2020 was used. After summarization, cleaning, and normalization, a total of 28,813 records related to safety risk parameters, such as weather conditions, aircraft configuration, flight information, speed, altitude, and air traffic, were selected. Due to the dependency of flight data on previous inputs and the need for some form of memory, training was performed using the LSTM algorithm in the Python environment. After learning, the mean squared error of deviations was approximately 6.38%. The result showed that the error rate is negligible, and the proposed model has high credibility compared to similar models. This model, equipped with advanced tools including ETL (Extract, Transform, Load), metadata, and real-time monitoring, addressed the challenges of exploring and cleaning large-scale flight data. It successfully predicted the most critical safety factor during the approach phase: speed and altitude control during landing. This reliable approach assists flight crews in controlling important safety parameters, including avoiding loss of control, aircraft speed, touch-down position, and preventing runway excursions.

کلیدواژه‌ها English

Flight Safety Risk
Air Transportation
Landing
Model
Deep Learning
BI.M-LSTM
-Aleksandrs Bitiņs, Ruta Bogdane, Vladimir Shestakov, Anastasija Stepanova (2022). The oretical and Methodological approaches to The Information Base for an Airlines Flight Safety System, Transactions on Aerospace Research eISSN. Vol. 266, No. 1/2022, 75-83.doi: 10.2478/tar-2022-0006
-Aysha S. Hameed­, Bindu G．R, Srianish Vutukuri  (2024). Approach and landing guidance using constrained model predictive static programming Aerospace Science and Technology 144 (2024) 108732. 
-Baars, H. & Kemper, H.-G. (2008). Management Support with Structured and Unstructured Data: An Integrated Business Intelligence Framework, Information Systems Management, 25(2). 132-148.
-Balachandran. S and Ella. M. Atkins (2015). Flight Safety Assessment and anagement for Takeoff Using Deterministic Moore Machines, Journal of Aerospace Information Systems, Vol. 12, No. 9.
-Borst, C., Grootendorst, F. H., Brouwer, D. I. K., Bedoya, C., Mulder, M., and van Paassen, M. M., (2013). Design and Evaluation of a Safety Augmentation System for Aircraft. Journal of Aircraft, Vol. 51, No. 12–22.
 doi:10.2514/1.C031500
-Chongfeng Li­, Ruishan Sun , Xing Pan (2023).  Takeoff runway overrun risk assessment in aviation safety based on human pilot behavioral characteristics from real flight data Safety Science 158, 105992.
-D. P. Kingma and J. Ba, (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
-Eduardo Gallo(2021) Quasi Static Atmospheric Model for Aircraft Trajectory Prediction and Flight Simulation Systems and Control (eess.SY) arXiv:2101.10744v1 [eess.SY] 26 Jan 2021
-F. A. Gers, J. Schmidhuber, and F. Cummins, (1999). Learning to forget: Continual prediction with lstm.
-Gabriel Jarry, Daniel Delahaye, Eric Féron.(2020)  Approach and landing aircraft on-board parameters estimation with LSTM networks. AIDA-AT 2020, 1st conference on Artificial Intelligence and Data Analytics in Air Transportation, Feb 2020, Singapore, Singapore. ISBN: 978-1-7281-5381-0.
-Govindarajan, N., De Visser, C., Van Kampen, E., Krishnakumar, K., Barlow, J., and Stepanyan, V., (2015). Optimal Control Framework for Estimating Autopilot Safety Margins. Journal of Guidance, Control, and Dynamics, Vol. 38, No. 7, 1197–1207. doi:10.2514/1.G000271
-Guo Y, Sun Y (2020) Flight safety assessment based on an integrated human reliability quantification approach. PLoS ONE 15(4): e0231391. doi.org/10.1371/journal. pone.0231391
-Habler, Edan, Bitton, Ron & Shabtai, Asef (2021). Evaluating the Security of Aircraft Systems arXiv: 2209.04028v1 [cs.CR] 8 Sep 2022.
-Hong Sun, Fangquan Yang, Peiwen Zhang, Yang Jiao and Yunxiang Zhao (2023). An Innovative Deep Architecture for Flight Safety Risk Assessment Based on Time Series Data Computer Modeling in Engineering & Sciences.
doi: 10.32604/cmes.2023.030131
-Hsueh-Yi Lai (2023). Breakdowns in team resilience during aircraft landing due to mental model disconnects as identified through machine Reliability Engineering and System Safety 237,  109356.
-Jammal, P., Fischer, O. P., Mavris, D. N., & Wagner, G. (2025). Advancing Aviation Safety through Predictive Maintenance: A Machine Learning Approach for Carbon Brake Wear Severity Classification. Aerospace, 12(7), 602.
-Yahyavi.M­, toloie.A, Afsharkazemi.M, Radfar,R (2024) designing 30495/JIK.2024.23254an intelligent model to optimize to safety risk of the takeoff flight using B.I.M-LSTM.
-ICAO, Doc 9859, (2022). Safety management manual, 4th ed. Montréal, Quebec, Canada: International Civil Aviation Organization (ICAO).
-Jing Lu, Longfei Pan, Jingli Deng, Hongjun Chai1, Zhou Ren1 and Yu Shi, (2022). Deep learning for Flight Maneuver Recognition: A survey ERA, 31(1): 75–102. doi: 10.3934/era.2023005
-JuanFang, QiangangZheng, ChangpengCai, HaoyinChen, HaiboZhang (2023).  Deep reinforcement learning method for turbofan engine acceleration optimization.
-Lai H-Y lee B. (2017). Unstable approach: intervention and prevention. In: Transdisciplinary engineering: a paradigm shift: proceedings of the 24th ISPE Inc. International Conference on Transdisciplinary Engineering, July 10–14, IOS Press.
-Mickael Rey, Daniel Aloise , François Soumis ,Romanic Pieugueu (2021) A data-driven model for safety risk identification from flight data analysis. Transportation Engineering 5, 100087.
-Ng Iris, Sarasvathi Nagalingham (2023). Implementation of Business Intelligence Solution for United Airlines. (IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 14, No. 1.
-PeiyaoWang, Mingxin Yu1 , Guang Yan1, Jiabin Xia, Jiawei Liu1and Lianqing Zhu (2023).
A deep learning-based method for calculating aircraft wing loads Measurement and Control 1–13.  The Author(s) 2023 Article reuse guidelines. sagepub.com/journals-permissions.
doi: 177/00202940221145971journals.sagepub.com/home/mac
-Singh, G.; Singh, J. Prabha, C. (2022). Data visualization and its key fundamentals:
A comprehensive survey. In Proceedings of the 2022th International Conference on Communication and Electronics Systems (ICCES), Coimbatore, India, 22–24 June.
-Tahsin Sejat Saniat, Tahiat Goni, Shaikat M. Galib (2020). lstm recurrent neural network assisted aircraft stall prediction for enhanced situational awareness arxiv:2012.04876v1 [cs.lg] 9 dec.
-Tejas Puranik, Evan Harrison, Sanggyu Min,Hernando Jimenez, and Dimitri Mavris (2023). General Aviation Approach and Landing Analysis using Flight Data Records Downloaded by Ecole Technologie Superieure (ETS) on December 31.
-Yi Lin , Linjie Deng, Zhengmao Chen, Xiping Wu, Jianwei Zhang, and Bo Yang (2020). A Real-Time ATC Safety Monitoring nramework using a Deep Learning Approach ieeVol. 21, No. 11, Nov.
-Zhi lu et al (2018). An Architecture of system of system (SoS) for Commerical flight Security in 5G 5th internatonal conferarancr on system and informatics (ICSAL 2018).
-Zhu, D.; Wang, Y.; Zhang, F. (2022). Energy Price Prediction Integrated with Singular Spectrum Analysis and Long Short-Term Memory Network against the Background of Carbon Neutrality. Energies, 15, 8128.