A Hybrid Deep Learning-Based Intrusion Detection System for IOT Devices

F.C Uzoezie; I.C Emeto; A.A. Galadima; I.H. Ezeh; A.C Okoloegbo; S. Kwaghbee

doi:10.56201/ijcsmt.vol.12.no1.2026.pg155.163

📄 Download Paper

A Hybrid Deep Learning-Based Intrusion Detection System for IOT Devices

F.C Uzoezie, I.C Emeto, A.A. Galadima, I.H. Ezeh, A.C Okoloegbo, S. Kwaghbee

Abstract

The rapid growth of the Internet of Things has created a highly interconnected digital environment, but this expansion has also introduced complex security risks, especially at the application layer where most user-level operations occur. Many existing intrusion detection systems are unable to cope with the fast-evolving and heterogeneous nature of IoT traffic. In this study, a hybrid deep learning intrusion detection framework is presented to address these challenges. The system integrates Convolutional Neural Networks to extract spatial features from log data and Recurrent Neural Networks to learn temporal patterns associated with malicious behaviour. The combined architecture enables the model to recognize subtle and sophisticated attack signatures in real time. Experiments conducted using both synthetic traffic and real-world event datasets show that the proposed IDS achieves higher accuracy, precision, recall, and F1-score when compared with traditional machine learning approaches. To demonstrate its practical relevance, the model was deployed in a Flask-based web application that performs live monitoring and analysis of IoT traffic. Tests carried out in a smart-city surveillance context confirm the system’s effectiveness in identifying application-layer threats as they occur.

Keywords

IoT securityDeep learningIntrusion Detection SystemsCNNRNNApplication layerSmart environmentsCyber-physical systems.

References

[1] R. Roman, J. Zhou, and J. Lopez, “Challenges and solutions for securing the internet of things: A survey,” IEEE Internet Things J., vol. 7, no. 4, pp. 3367–3382, Oct. 2020, doi: 10.1109/JIOT.2020.2987889. [2] S. Asharf, A. Shakil, M. Hassan, and M. Habib, “Security and privacy in IoT: Architecture, challenges, and countermeasures,” J. Comput. Netw. Commun., vol. 2020, Art. no. 8838654, 2020, doi: 10.1155/2020/8838654. [3] P. Chopra, S. Gupta, and R. Singh, “Security challenges in IoT: A comprehensive survey,” J. Cyber Secur. Mobil., vol. 9, no. 2, pp. 203–220, 2020, doi: 10.13052/jcsm2245- 1439.922. [4] M. Alghamdi, “Lightweight cryptographic algorithms for IoT security: A survey,” J. Netw. Comput., vol. 5, no. 6, p. 5, 2022. [5] P. Gupta and S. Mukherjee, “Anomaly-based intrusion detection in IoT using machine learning,” Comput. Secur., vol. 107, p. 102303, 2021, doi: 10.1016/j.cose.2021.102303. [6] M. A. Al-Garadi, A. Mohamed, K.-K. R. Choo, and F. S. Al-Anzi, “A survey of machine and deep learning methods for internet of things security,” IEEE Commun. Surv. Tutor., vol. 22, no. 3, pp. 1646–1685, 2020. [7] A. Williams, J. Brown, and K. Wilson, “IoT security: A comparative study of security challenges and countermeasures,” Int. J. Secur. Netw., vol. 15, no. 2, pp. 129–141, 2020, doi: 10.1504/IJSN.2020.108839. [8] S. Hassan, F. Umer, and M. A. Pasha, “Secure IoT architecture: A comprehensive review,” IEEE Internet Things J., vol. 8, no. 3, pp. 1970–1983, 2021, doi: 10.1109/JIOT.2020.3019273. [9] P. Hilton, “The 2016 Dyn cyberattack: How IoT devices were used in a massive DDoS attack,” J. Cybersecur., vol. 4, no. 1, pp. 21–28, 2020, doi: 10.1093/cybsec/tyz018. [10] L. Pycroft and T. Z. Aziz, “Cybersecurity in healthcare: IoT devices and patient safety,” J. Med. Syst., vol. 44, no. 3, pp. 52–62, 2020, doi: 10.1007/s10916-020-15395. [11] M. Singh and S. Sharma, “Intrusion detection in IoT networks: A review,” J. Inf. Secur. Appl., vol. 65, p. 102978, 2022, doi: 10.1016/j.jisa.2022.102978. [12] A. Almomani, M. Desouki, and M. Ayyash, “Lightweight IDS for IoT networks using random forests,” Comput. Secur., vol. 90, p. 101683, 2020, doi: 10.1016/j.cose.2019.101683. [13] T. Smith, R. Patel, and Y. Li, “Convolutional neural networks for smart home anomaly detection,” Neurocomputing, vol. 441, pp. 205–217, 2021, doi: 10.1016/j.neucom.2020.08.045. [14] J. Kim, S. Lee, and Y. Park, “Hybrid intrusion detection system for industrial IoT,” IEEE Internet Things J., vol. 9, no. 1, pp. 125–136, 2022, doi: 10.1109/JIOT.2021.3095586. [15] S. Sharma, R. Gupta, and V. Kumar, “Federated learning for intrusion detection in IoT: Challenges and opportunities,” J. Inf. Secur. Appl., vol. 60, p. 102894, 2021, doi: 10.1016/j.jisa.2021.102894. [16] H. Zhang, Y. Wang, and Z. Liu, “Blockchain-based decentralized IDS for IoT networks,” J. Commun. Netw., vol. 23, no. 4, pp. 215–228, 2021, doi: 10.1109/JCN.2021.00025. [17] G. Singh and A. Sharma, “IoT-specific attack detection using hybrid models: A deep learning approach,” Int. J. Netw. Secur., vol. 22, no. 5, pp. 511–521, 2020. [18] T. Chen, X. Liu, and W. Zhang, “Distributed IDS for smart grids using collaborative anomaly detection,” J. Electr. Comput. Eng., vol. 2021, Art. no. 4562123, 2021, doi: 10.1155/2021/4562123. [19] M. Ahmed, S. Khan, and M. Ali, “Deep learning-based intrusion detection system for healthcare IoT,” J. Netw. Comput. Appl., vol. 209, p. 102589, 2023. [20] Y. Zhang, J. Lin, and L. Zhao, “Hybrid federated learning for IoT security,” IEEE Trans. Ind. Informat., vol. 18, no. 6, pp. 3765–3774, 2022, doi: 10.1109/TII.2021.3085415.

A Hybrid Deep Learning-Based Intrusion Detection System for IOT Devices

Abstract

Keywords

References

More Articles from INTERNATIONAL JOURNAL OF COMPUTER SCIENCE AND MATHEMATICAL THEORY