Hybrid Deep Learning Approach for Early Detection of Railway Track Faults to Enhance Railway Safety
DOI:
https://doi.org/10.63282/3050-9246.IJETCSIT-V7I2P138Keywords:
Railway Track Inspection, Spectrograms, Acoustic Signals, Machine Learning, Fault DiagnosisAbstract
Railway tracks are the primary infrastructure that allows trains to move and any defect in the railway track can lead to minor or major accidents. Consistent monitoring and condition-based maintenance are crucial for safety and minimizing hazards. In this paper, an automated fault detection system is proposed for railway track fault diagnosis using Hybrid RF+LSTM model based on the Railway Track Fault Detection (RTFD) dataset. Class imbalance is handled by noise removal, normalization, encoding features, and balancing with SMOTE, while the dataset's quality is enhanced. Their proposed hybrid model combines Random Forest and Long Short-Term Memory (LSTM) to model sequential dependencies, leading to improved classification accuracy. Prior ML and DL models, such as CNN, ResNet-50, and VGG16, are applied for evaluate model's performance. With scores of 98% for accuracy, 96% for precision, 98% for recall, and 97% for F1-score, the Hybrid RF+LSTM model surpasses all other models in the experiments. Research like this lends credence to the idea that the proposed method could greatly enhance railway safety and reliability by making track defects much easier to spot.
Downloads
References
[1] C. Yang, Y. Sun, C. Ladubec, and Y. Liu, “Developing Machine Learning-Based Models for Railway Inspection,” Appl. Sci., vol. 11, no. 1, p. 13, Dec. 2020, doi: 10.3390/app11010013.
[2] T. Bai, J. Gao, J. Yang, and D. Yao, “A Study on Railway Surface Defects Detection Based on Machine Vision,” Entropy, vol. 23, no. 11, p. 1437, Oct. 2021, doi: 10.3390/e23111437.
[3] M. Trigka and E. Dritsas, “A Comprehensive Survey of Deep Learning Approaches in Image Processing,” Sensors, vol. 25, no. 2, p. 531, Jan. 2025, doi: 10.3390/s25020531.
[4] M. Kari, “Deep Learning-Based Fault Prediction Models for Enhanced Network Security Monitoring,” Int. J. Adv. Res. Sci. Commun. Technol., vol. 3, no. 3, p. 492, Jun. 2023, doi: 10.48175/IJARSCT-11600I.
[5] A. K. Padhy, T. Pravinbhai Patel, V. Soni, S. Shivam, G. B. Thokala, and B. Vulugundam, “Machine Learning-Based Fault Prediction in Large- Scale Distributed Systems,” in 2026 IEEE 5th International Conference on AI in Cybersecurity (ICAIC), IEEE, Feb. 2026, pp. 1–6. doi: 10.1109/ICAIC67076.2026.11395778.
[6] A. K. Padhy, N. Seshagiri, V. Soni, A. K. Elengovan, G. Babu Thokala, and M. Kumari, “AI-Enabled Autonomous Data Preprocessing: A Scalable Architecture for Intelligent Machine Learning Pipeline Management,” in 2026 6th International Conference on Image Processing and Capsule Networks (ICIPCN), IEEE, Jan. 2026, pp. 1318–1323. doi: 10.1109/ICIPCN67432.2026.11438503.
[7] A. Malekjafarian, E. OBrien, P. Quirke, and C. Bowe, “Railway Track Monitoring Using Train Measurements: An Experimental Case Study,” Appl. Sci., vol. 9, no. 22, p. 4859, Nov. 2019, doi: 10.3390/app9224859.
[8] V. Badrinarayanan, A. Kendall, and R. Cipolla, “SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 39, no. 12, pp. 2481–2495, Dec. 2017, doi: 10.1109/TPAMI.2016.2644615.
[9] Y. Zou, R. Liu, Y. Li, Y. Ma, and G. Wang, “Signal adaptive cooperative control of two adjacent traffic intersections using a two-stage algorithm,” Expert Syst. Appl., vol. 174, p. 114746, Jul. 2021, doi:10.1016/j.eswa.2021.114746.
[10] I. Bravo, U. Alvarado, J. Nieto, and P. Ciáurriz, “On-board accelerometers in railway track condition monitoring. A systematic review,” Transp. Res. Interdiscip. Perspect., vol. 33, p. 101572, Sep. 2025, doi: 10.1016/j.trip.2025.101572.
[11] R. Ghiasi and A. Malekjafarian, “Monitoring of railway tracks maintenance needs using dynamic responses collected by an in-service train,” Railw. Eng. Sci., vol. 33, no. 4, pp. 553–580, Dec. 2025, doi: 10.1007/s40534-025-00380-w.
[12] R. Ghiasi, N. Lestoille, C. Diaine, and A. Malekjafarian, “Unsupervised domain adaptation for drive-by condition monitoring of multiple railway tracks,” Eng. Appl. Artif. Intell., vol. 139, p. 109516, Jan. 2025, doi: 10.1016/j.engappai.2024.109516.
[13] A. Sansiñena, B. Rodríguez-Arana, and S. Arrizabalaga, “A systematic review of acceleration-based estimation of railway track quality,” Veh. Syst. Dyn., pp. 1–28, Mar. 2025, doi: 10.1080/00423114.2025.2483972.
[14] B. Baasch, J. C. Groos, J. Heusel, and M.-C. Noll, “Data-driven sparse coding for onboard condition monitoring of railway tracks,” Mech. Syst. Signal Process., vol. 241, p. 113542, Dec. 2025, doi: 10.1016/j.ymssp.2025.113542.
[15] A. S, K. S, and I. Cephas, “A Real Time Multi Sensor Railway Track Monitoring System with Machine Learning Based Fault Detection and Predictive Maintenance,” in 2026 International Conference on Machine Learning and Autonomous Systems (ICMLAS), IEEE, Mar. 2026, pp. 1664–1670. doi: 10.1109/ICMLAS67792.2026.11483961.
[16] S. S, J. G, K. L. D, M. R J, and P. S, “Multi-modal Based Railway Accident Classification and Track Fault Prediction for optimized healthcare,” in 2025 2nd International Conference on Artificial Intelligence for Innovations in Healthcare Industries (ICAIIHI), IEEE, Dec. 2025, pp. 1–6. doi: 10.1109/ICAIIHI67124.2025.11403425.
[17] S. Sali, D. Pandhare, R. Tiwari, and S. Khiani, “Real-Time Detection of Railway Faults Using Edge-Based Machine Learning,” in 2025 2nd International Conference on Recent Trends in Electrical, Electronics and Computing Technologies (ICRTEECT), IEEE, Oct. 2025, pp. 1–6. doi: 10.1109/ICRTEECT67512.2025.11448965.
[18] R. Thinakaran, S. Jalari, V. Neerugatti, M. R. Nalluri, S. Chukka, and R. R. Cholla, “Enhancing Railway Safety with an IoT Based System for Real Time Fault Detection and Crack Monitoring,” in 2024 9th International Conference on Information Technology and Digital Applications (ICITDA), IEEE, Nov. 2024, pp. 01–04. doi: 10.1109/ICITDA64560.2024.10809610.
[19] G. Kaur, N. Sharma, S. Malhotra, S. Devliyal, and R. Gupta, “Automated Railway Track Fault Detection using ResNet50V2 Deep Learning Architecture,” in 2024 Asia Pacific Conference on Innovation in Technology (APCIT), IEEE, Jul. 2024, pp. 1–6. doi: 10.1109/APCIT62007.2024.10673489.
[20] N. Mukhtiar, K. Khuwaja, D. P. Tocut, and I. Tarca, “An Experimental Study on the Railway Track Surface Fault Detection with Automation,” in 2023 17th International Conference on Engineering of Modern Electric Systems (EMES), IEEE, Jun. 2023, pp. 1–4. doi: 10.1109/EMES58375.2023.10171771.
[21] M. Yilmazer and M. Karakose, “Yolov5 Based Fault Detection Approach in Railway Components,” in 2023 27th International Conference on Information Technology (IT), IEEE, Feb. 2023, pp. 1–4. doi: 10.1109/IT57431.2023.10078564.
[22] S. I. Eunus, S. Hossain, S. Hossain, and A. E. M. Ridwan, “Railway Track Fault Detection,” Kaggle. [Online]. Available: https://www.kaggle.com/datasets/salmaneunus/railway-track-fault-detection
[23] R. Shafique et al., “A Novel Approach to Railway Track Faults Detection Using Acoustic Analysis,” Sensors, vol. 21, no. 18, p. 6221, Sep. 2021, doi: 10.3390/s21186221.
[24] R. S. Rampriya et al., “Fault Detection and Semantic Segmentation on Railway Track Using Deep Fusion Model,” IEEE Access, vol. 12, no. August, pp. 136183–136201, 2024, doi: 10.1109/ACCESS.2024.3462822.
[25] Anita, “Railroad Track Defect Detections using Deep Learning,” Int.
[26] J. Eng. Manag. Res., vol. 15, pp. 89–96, 2025, doi: 10.5281/zenodo.15356939.
