cahtНаучный вестник МГТУ ГАCivil Aviation High Technologies2079-06192542-0119Moscow State Technical University of Civil Aviation (MSTU CA)10.26467/2079-0619-2025-28-1-20-38caht-2498Research ArticleТРАНСПОРТНЫЕ СИСТЕМЫTRANSPORTATION SYSTEMSПрименение Seq2seq-моделей для прогнозирования развития грозовой деятельности с целью повышения уровня ситуационной осведомленности пилота в полетеApplication of Seq2Seq models for predicting the development of thunderstorm activity to enhance the pilot’s situational awareness in flightКоваленкоГ. В.KovalenkoG. V.

Коваленко Геннадий Владимирович, доктор технических наук, профессор, профессор кафедры летной эксплуатации и безопасности полетов в гражданской авиации

г. Санкт-Петербург

Gennadiy V. Kovalenko, Doctor of Technical Sciences, Professor, Professor of the Chair of Flight Operations and Flight Safety in Civil Aviation

Saint Petersburg

kgvf@inbox.ruЯдровИ. А.YadrovI. A.

Ядров Илья Александрович, аспирант кафедры летной эксплуатации и безопасности полетов в гражданской авиации

г. Санкт-Петербург

Ilya A. Yadrov, Postgraduate Student of the Chair of Flight Operations and Flight Safety in Civil Aviation

Saint Petersburg

yadrov.ilya@gmail.comСанкт-Петербургский государственный университет гражданской авиации имени Главного маршала авиации А.А. НовиковаРоссияSaint Petersburg State University of Civil Aviation named after Chief Marshal of Aviation A.A. NovikovRussian Federation2025040320252812038Copyright © Коваленко Г.В., Ядров И.А., 20252025Коваленко Г.В., Ядров И.А.Kovalenko G.V., Yadrov I.A.This work is licensed under a Creative Commons Attribution 4.0 License.https://avia.mstuca.ru/jour/article/view/2498

В статье представлены результаты применения моделей Seq2seq на основе нейронных сетей для наукастинга – прогнозирования с заблаговременностью до 2 часов грозовой активности с целью повышения ситуационной осведомленности экипажей воздушных судов. На основе данных радиолокационных метеорологических наблюдений за грозовыми очагами были созданы и обучены различные рекуррентные и сверточные рекуррентные модели. Результаты исследования показали, что сверточные рекуррентные нейронные сети (ConvRNN, ConvLSTM, ConvGRU) превосходят классические рекуррентные модели, и при этом позволяют улучшить прогноз развития грозы на 25–30 % по метрике RMSE (корень среднеквадратической погрешности) по сравнению с базовой моделью, каждый раз в качестве предсказания выбирающей последнее доступное на момент предсказания радиолокационное изображение. Тем не менее, несмотря на то что сверточные рекуррентные сети позволяют достаточно точно передать общую тенденцию изменения формы грозового облака, точность предсказания интенсивности элементов грозового очага оказывается, как правило, завышенной. Применение предложенной технологии прогнозирования грозовой активности может способствовать повышению уровня ситуационной осведомленности летного экипажа, улучшая проекцию текущей обстановки на ближайшее будущее и оптимизируя процесс принятия решений по обходу грозы за счет предоставления членам экипажа прогностической информации о развитии грозы на экране навигационного дисплея. В рамках будущих исследований предполагается дальнейшая оптимизация архитектуры моделей, а также интеграция прогностической технологии в системы поддержки принятия решений экипажем.

The paper presents the results of application of Seq2seq models based on neural networks for nowcasting-forecasting with a lead time of up to 2 hours – of thunderstorm activity in order to increase situational awareness of aircraft crews. Various recurrent and convolutional recurrent models were created and trained on the basis of radar meteorological observations of thunderstorm cells. The results showed that convolutional recurrent neural networks (ConvRNN, ConvLSTM, ConvGRU) outperform classical recurrent models and improve the thunderstorm forecast by 25–30% in terms of RMSE (root mean square error) metric compared to the baseline model, which always selects the most recent radar image available at the time of prediction. Nevertheless, despite the fact that the convolution recurrence models can accurately represent the general trend of thunderstorm cloud shape changes, the accuracy of predicting the intensity of thunderstorm cells is usually overestimated. Application of the proposed thunderstorm activity forecasting technology can enhance the situational awareness of the flight crew improving the projection of the current situation into the near future and optimizing the decision-making process for thunderstorm avoidance by providing crew members with predictive information about thunderstorm development on the navigation display screen. Future research is expected to further optimize the model architecture and integrate the predictive technology into flight crew decision support systems.

Seq2seqпрогнозированиерекуррентные нейронные сетисверточные рекуррентные нейронные сетиситуационная осведомленностьобход грозыSeq2seqpredictionrecurrent neural networksconvolutional recurrent neural networkssituational awarenessthunderstorm avoidanceReferencesBolstad C.A., Riley J.M. Using goal directed task analysis with Army brigade officer teams // Proceedings of the Human Factors and Ergonomics Society Annual Meeting. SAGE Publications. 2002. Vol. 46, no.3. Pp. 472–476. DOI:10.1177/154193120204600354Bolstad, C.A., Riley, J.M. (2002). Using goal directed task analysis with Army brigade officer teams. In: Proceedings of the Human Factors and Ergonomics Society Annual Meeting. SAGE Publications, vol. 46, no. 3, pp. 472–476. DOI: 10.1177/154193120204600354Stanton N.A., Chambers P.R.G., Piggott J. Situational awareness and safety // Safety science. 2001. Vol. 39, no. 3. Pp. 189–204. DOI: 10.1016/S0925-7535(01)00010-8Stanton, N.A., Chambers, P.R.G., Piggott, J. (2001). Situational awareness and safety. Safety science, vol. 39, no. 3, pp. 189–204. DOI: 10.1016/S0925-7535(01)00010-8Sarter N.B., Woods D.D. Situation awareness: A critical but Ill-defined phenomenon // The International Journal of Aviation Psychology. 1991. Vol. 1, no. 1. Pp. 45–57. DOI: 10.1207/s15327108ijap0101_4Sarter, N.B., Woods, D.D. (1991). Situation awareness: A critical but Ill-defined phenomenon. The International Journal of Aviation Psychology, vol. 1, no. 1, pp. 45–57. DOI: 10.1207/s15327108ijap0101_4Стрелков Ю.К. Инженерная и профессиональная психология: учеб. пособие. М.: Академия; Высшая школа, 2001. 360 с.Strelkov, Yu.K. (2001). Engineering and professional psychology: textbook. Moscow: Vysshaya shkola; Akademiya, 360 p. (in Russian)De Gooijer J.G., Hyndman R.J. 25 years of time series forecasting // International Journal of Forecasting. 2006. Vol. 22, no. 3. Pp. 443–473. DOI: 10.1016/j.ijforecast.2006.01.001De Gooijer, J.G., Hyndman, R.J. (2006). 25 years of time series forecasting. International Journal of Forecasting, vol. 22, no. 3, pp. 443–473. DOI: 10.1016/j.ijforecast.2006.01.001Stevenson S. A comparison of the forecasting ability of ARIMA models // Journal of Property Investment & Finance. 2007. Vol. 25, no. 3. Pp. 223–240. DOI: 10.1108/14635780710746902Stevenson, S. (2007). A comparison of the forecasting ability of ARIMA models. Journal of Property Investment & Finance, vol. 25, no. 3, pp. 223–240. DOI: 10.1108/14635780710746902Chatfield C. A new look at models for exponential smoothing / C. Chatfield, A.B. Koehler, J.K. Ord, R.D. Snyder // Journal of the Royal Statistical Society: Series D (The Statistician). 2001. Vol. 50, no. 2. Pp. 147–159. DOI: 10.1111/1467-9884.00267Chatfield, C., Koehler, A.B., Ord, J.K., Snyder, R.D. (2001). A new look at models for exponential smoothing. Journal of the Royal Statistical Society: Series D (The Statistician), vol. 50, no. 2, pp. 147–159. DOI: 10.1111/1467-9884.00267Bentéjac C., Csörgő A., Martínez-Muñoz G. A comparative analysis of gradient boosting algorithms // Artificial Intelligence Review. 2021. Vol. 54. Pp. 1937–1967. DOI: 10.1007/s10462-020-09896-5Bentéjac, C., Csörgő, A., Martínez-Muñoz, G. (2021). A comparative analysis of gradient boosting algorithms. Artificial Intelligence Review, vol. 54, pp. 1937–1967. DOI: 10.1007/s10462-020-09896-5Faloutsos C. Classical and contemporary approaches to big time series forecasting / C. Faloutsos, J. Gasthaus, T. Januschowski, Y. Wang // SIGMOD '19: Proceedings of the 2019 International Conference on Management of Data, 2019. Pp. 2042–2047. DOI: 10.1145/3299869.3314033Faloutsos, C., Gasthaus, J., Januschowski, T., Wang, Y. (2019). Classical and contemporary approaches to big time series forecasting. In: SIGMOD '19: Proceedings of the 2019 International Conference on Management of Data, pp. 2042–2047. DOI: 10.1145/3299 869.3314033Makridakis S., Spiliotis E., Assimakopoulos V. The M4 Competition: Results, findings, conclusion and way forward // International Journal of Forecasting. 2018. Vol. 34, iss. 4. Pp. 802–808. DOI: 10.1016/j.ijforecast.2018.06.001Makridakis, S., Spiliotis, E., Assimakopoulos, V. (2018). The M4 Competition: Results, findings, conclusion and way forward. International Journal of Forecasting, vol. 34, issue 4, pp. 802–808. DOI: 10.1016/j.ijforecast.2018.06.001Taieb S.B., Sorjamaa A., Bontempi G. Multiple-output modeling for multi-step-ahead time series forecasting // Neurocomputing. 2010. Vol. 73, iss. 10–12. Pp. 1950–1957. DOI: 10.1016/j.neucom.2009.11.030Taieb, S.B., Sorjamaa, A., Bontempi, G. (2010). Multiple-output modeling for multi-step-ahead time series forecasting. Neurocomputing, vol. 73, issue 10–12, pp. 1950–1957. DOI: 10.1016/j.neucom.2009.11.030Sutskever I., Vinyals O., Quoc V.L. Sequence to sequence learning with neural networks // Proceedings of the 27th International Conference on Neural Information Processing Systems. 2014. No. 2. Pp. 3104–3112. DOI: 10.48550/arXiv.1409.3215Sutskever, I., Vinyals, O., Quoc, V.L. (2014). Sequence to sequence learning with neural networks. In: Proceedings of the 27th International Conference on Neural Information Processing Systems, no. 2, pp. 3104–3112. DOI: 10.48550/arXiv.1409.3215Caterini A.L. Recurrent neural networks / A.L. Caterini, D.E. Chang, A.L. Caterini, D.E. Chang. In book: Deep Neural Networks in a Mathematical Framework. Springer Briefs in Computer Science. Springer, Cham, 2018. Pp. 59–79. DOI: 10.1007/978-3-319-75304-1_5Caterini, A.L., Chang, D.E., Caterini, A.L., Chang, D.E. (2018). Recurrent neural networks. In book: Deep Neural Networks in a Mathematical Framework, Springer Briefs in Computer Science. Springer, Cham, pp. 59–79. DOI: 10.1007/978-3-319-75304-1_5Rumelhart D.E., Hinton G.E., Williams R.J. Learning representations by backpropagating errors // Nature. 1986. No. 323. Pp. 533–536. DOI: 10.1038/323533a0Rumelhart, D.E., Hinton, G.E., Williams, R.J. (1986). Learning representations by back-propagating errors. Nature, no. 323, pp. 533–536. DOI: 10.1038/323533a0Rasamoelina A.D., Adjailia F., Sinčák P. A review of activation function for artificial neural network // 2020 IEEE 18th World Symposium on Applied Machine Intelligence and Informatics (SAMI). Slovakia, Herlany, 2020. Pp. 281–286. DOI: 10.1109/SAMI48414.2020.9108717Rasamoelina, A.D., Adjailia, F., Sinčák, P. (2020). A review of activation function for artificial neural network. In: 2020 IEEE 18th World Symposium on Applied Machine Intelligence and Informatics (SAMI), Herlany, Slovakia, pp. 281–286. DOI: 10.1109/SAMI48414.2020.9108717Toharudin T. Employing long shortterm memory and Facebook prophet model in air temperature forecasting / T. Toharudin, R.S. Pontoh, R.E. Caraka, S. Zahroh, Y. Lee Y, R.C. Chen // Communications in Statistics-Simulation and Computation. 2023. Vol. 52, iss. 2. Pp. 279–290. DOI: 10.1080/03610918.2020.1854302Toharudin, T., Pontoh, R.S., Caraka, R.E., Zahroh, S., Lee, Y., Chen, R.C. (2023). Employing long short-term memory and Facebook prophet model in air temperature forecasting. Communications in Statistics-Simulation and Computation, vol. 52, issue 2, pp. 279–290. DOI: 10.1080/03610918.2020.1854302Schmidhuber J., Hochreiter S. Long short-term memory // Neural Computation. 1997. Vol. 9, iss. 8. Pp. 1735–1780. DOI: 10.1162/neco.1997.9.8.1735Schmidhuber, J., Hochreiter, S. (1997). Long short-term memory. Neural Computation, vol. 9, issue 8, pp. 1735–1780. DOI: 10.1162/neco.1997.9.8.1735De Mulder W., Bethard S., Moens M.F. A survey on the application of recurrent neural networks to statistical language modeling / Computer Speech & Language. 2015. Vol. 30, iss. 1. Pp. 61–98. DOI: 10.1016/j.csl.2014.09.005De Mulder, W., Bethard, S., Moens, M. F. (2015). A survey on the application of recurrent neural networks to statistical language modeling. Computer Speech & Language, vol. 30, issue 1, pp. 61–98. DOI: 10.1016/j.csl.2014.09.005Shi X. Convolutional LSTM network: A machine learning approach for precipitation nowcasting / X. Shi, Z. Chen, H. Wang, D.Y. Yeung, W.K. Wong, W.C. Woo // NIPS'15: Proceedings of the 28th International Conference on Neural Information Processing Systems. 2015. Vol. 1. Pp. 802–810. DOI: 10.48550/arXiv.1506.04214Shi, X., Chen, Z., Wang, H., Yeung, D.Y., Wong, W.K., Woo, W.C. (2015). Convolutional LSTM network: A machine learning approach for precipitation nowcasting. In: NIPS'15: Proceedings of the 28th International Conference on Neural Information Processing Systems, vol. 1, pp. 802–810. DOI: 10.48550/arXiv.1506.04214Li Z. A survey of convolutional neural networks: analysis, applications, and prospects / Z. Li, F. Liu, W. Yang, S. Peng, J. Zhou // IEEE transactions on neural networks and learning systems. 2021. Vol. 33, no. 12. Pp. 6999–7019. DOI: 10.1109/TNNLS.2021.3084827Li, Z., Liu, F., Yang, W., Peng, S., Zhou, J. (2021). A survey of convolutional neural networks: analysis, applications, and prospects. In: IEEE transactions on neural networks and learning systems, vol. 33, no. 12, pp. 6999–7019. DOI: 10.1109/TNNLS.2021.3084827Ballas N. Delving deeper into convolutional networks for learning video representations / N. Ballas, L. Yao, C.J. Pal, A. Courville [Электронный ресурс] // 4th International Conference on Learning Representations (ICLR 2016), 2016. 2 p. DOI: 10.48550/arXiv.1511.06432 (дата обращения: 08.10.2024).Ballas, N., Yao, L., Pal, C.J., Courville, A. (2016). Delving deeper into convolutional networks for learning video representations. In: 4th International Conference on Learning Representations, 2 p. DOI: 10.48550/arXiv.1511.06432 (accessed: 08.10.2024).Mahafza B.R. Radar systems analysis and design using MATLAB. 2nd ed. Chapman and Hall, CRC, 2005. 638 p. DOI: 10.1201/9781420057072Mahafza, B.R. (2005). Radar systems analysis and design using MATLAB, 2nd ed. Chapman and Hall, CRC, 638 p. DOI: 10.1201/9781420057072Van Dyk D.A., Meng X.L. The art of data augmentation // Journal of Computational and Graphical Statistics. 2001. Vol. 10, no. 1. Pp. 1–50. DOI: 10.1198/10618600152418584Van Dyk, D.A., Meng, X.L. (2001). The art of data augmentation. Journal of Computational and Graphical Statistics, vol. 10, no. 1, pp. 1–50. DOI: 10.1198/10618600152418584Masters D., Luschi C. Revisiting small batch training for deep neural networks [Электронный ресурс] // Computer Science and Machine Learning. 2018. Pp. 1–18. DOI: 10.48550/arXiv.1804.07612 (дата обращения: 08.10.2024).Masters, D., Luschi, C. (2018). Revisiting small batch training for deep neural networks. Computer Science and Machine Learning, pp. 1–18. DOI: 10.48550/arXiv.1804.07612 (accessed: 08.10.2024).Werbos P.J. Backpropagation through time: what it does and how to do it // Proceedings of the IEEE. 1990. Vol.78, no. 10. Pp. 1550–1560. DOI: 10.1109/5.58337Werbos, P.J. (1990). Backpropagation through time: what it does and how to do it. In: Proceedings of the IEEE, vol. 78, no. 10, pp. 1550–1560. DOI: 10.1109/5.58337Llugsi R. Comparison between Adam, AdaMax and Adam W optimizers to implement a weather forecast based on neural networks for the Andean city of Quito / R. Llugsi, S.E. Yacoubi, A. Fontaine, P. Lupera // 2021 IEEE Fifth Ecuador Technical Chapters Meeting (ETCM), 2021. Pp. 1–6. DOI: 10.1109/ETCM53643.2021.9590681Llugsi, R., Yacoubi, S.E., Fontaine, A., Lupera, P. (2021). Comparison between Adam, AdaMax and Adam W optimizers to implement a weather forecast based on neural networks for the Andean city of Quito. In: 2021 IEEE Fifth Ecuador Technical Chapters Meeting (ETCM), pp. 1–6. DOI: 10.1109/ETCM53643.2021.9590681Bejani M.M., Ghatee M. A systematic review on overfitting control in shallow and deep neural networks // Artificial Intelligence Review. 2021. Vol. 54. Pp. 6391–6438. DOI: 10.1007/s10462-021-09975-1Bejani, M.M., Ghatee, M. (2021). A systematic review on overfitting control in shallow and deep neural networks. Artificial Intelligence Review, vol. 54. Pp. 6391–6438. DOI: 10.1007/s10462-021-09975-1Passos D., Mishra P. A tutorial on automatic hyperparameter tuning of deep spectral modelling for regression and classification tasks [Электронный ресурс] // Chemometrics and Intelligent Laboratory Systems. 2022. Vol. 223. ID: 104520. DOI: 10.1016/j.chemolab.2022.10 4520 (дата обращения: 08.10.2024).Passos, D., Mishra, P. (2022). A tutorial on automatic hyperparameter tuning of deep spectral modelling for regression and classification tasks. Chemometrics and Intelligent Laboratory Systems, no. 223. ID: 104520. DOI: 10.1016/j.chemolab.2022.104520 (accessed: 08.10.2024).Koloskova A., Hendrikx H., Stich S.U. Revisiting gradient clipping: Stochastic bias and tight convergence guarantees // International Conference on Machine Learning. 2023. Pp. 17343–17363. DOI: 10.48550/arXiv.2305. 01588Koloskova, A., Hendrikx, H., Stich, S.U. (2023). Revisiting gradient clipping: Stochastic bias and tight convergence guarantees. International Conference on Machine Learning, pp. 17343–17363. DOI: 10.48550/arXiv.2305.01588Endsley M.R. Toward a theory of situation awareness in dynamic systems // Human factors. 1995. Vol. 37, no. 1. Pp. 32–64. DOI: 10.1518/001872095779049543Endsley, M.R. (1995). Toward a theory of situation awareness in dynamic systems. Human factors, vol. 37, no. 1, pp. 32–64. DOI: 10.1518/001872095779049543Kaikkonen L. Bayesian networks in environmental risk assessment: A review / L. Kaikkonen, T. Parviainen, M. Rahikainen, L. Uusitalo, A. Lehikoinen // Integrated environmental assessment and management. 2021. Vol. 17, no. 1. Pp. 62–78. DOI: 10.1002/ieam.4332Kaikkonen, L., Parviainen, T., Rahikainen, M., Uusitalo, L., Lehikoinen, A. (2021). Bayesian networks in environmental risk assessment: A review. Integrated environmental assessment and management, vol. 17, no. 1, pp. 62–78. DOI: 10.1002/ieam.4332Kovalenko G.V., Yadrov I.A., Kuts K.A. Intelligent adaptive flight crew decision support system for thunderstorm avoidance // Russian Aeronautics. 2023. Vol. 66. Pp. 552–559. DOI: 10.3103/S1068799823030170Kovalenko, G.V., Yadrov, I.A., Kuts, K.A. (2023). Intelligent adaptive flight crew decision support system for thunderstorm avoidance. Russian Aeronautics, vol. 66. Pp. 552–559. DOI: 10.3103/S1068799823030170

The authors declare that there are no conflicts of interest present.