Auto Healing Intrusion Detection System with Reinforcement Learning for Proactive Defence in Autonomous IIoT Environment

This paper presented an Auto-Healing Intrusion Detection System (AH-IDS) designed for proactive cybersecurity defence in autonomous network environments. The proposed framework integrates CNN-LSTM-based intrusion detection with Deep Reinforcement Learning-driven autonomous mitigation to provide intelligent, adaptive, and self-healing cybersecurity capabilities. The hybrid deep learning architecture effectively captures both spatial and temporal network traffic features, while the reinforcement learning engine dynamically learns optimal mitigation strategies. Fog-assisted deployment improves low-latency processing and supports real-time attack response. Experimental evaluation using benchmark datasets demonstrated that the proposed framework achieves high detection accuracy, reduced false positives, and significantly lower response latency. Unlike traditional IDS frameworks that focus solely on attack detection, the proposed AH-IDS introduces autonomous healing capabilities capable of isolating compromised nodes, rerouting traffic, and restoring network stability with minimal human intervention. The integration of explainable AI further enhances system transparency and operational trustworthiness. Future work will focus on integrating federated learning mechanisms to support distributed collaborative learning across multiple edge devices while preserving data privacy. Additional improvements may include transformer-based architectures, blockchain-enabled trust management, and energy-efficient deployment strategies for large-scale IIoT and smart city environments. The proposed AH-IDS framework represents a significant step toward intelligent autonomous cybersecurity systems capable of self-adaptation, resilience, and proactive defence in next-generation network infrastructures.

[1]             Alabdan, R., Alabduallah, B., Alruwais, N., Arasi, M. A., Asklany, S. A., Alghushairy, O., ... & Alshareef, A. (2025). Blockchain-assisted improved interval type-2 fuzzy deep learning-based attack detection on internet of things driven consumer electronics. Alexandria Engineering Journal, 110, 153-167. [2]             Alhayan, F., Alruwais, N., Alamgeer, M., Alashjaee, A. M., Abdullah, M., Khadidos, A. O., ... & Alshareef, A. (2025). Design of advanced intrusion detection in cybersecurity using ensemble of deep learning models with an improved beluga whale optimization algorithm. Alexandria Engineering Journal, 121, 90-102. [3]             Al-Shurbaji, T., Anbar, M., Manickam, S., Al-Amiedy, T. A., Mukhaini, G. A., Hashim, H., ... & Atlam, E. S. (2025). BoT-EnsIDS: Approach for detecting IoT Botnet attacks leveraging bio-inspired based ensemble feature selection and hybrid deep learning model. Alexandria Engineering Journal, 129, 744-767. [4]             Bhukya, R., Moeed, S. A., Medavaka, A., Khadidos, A. O., Khadidos, A. O., & Selvarajan, S. (2025). SPARK and SAD: Leading-edge deep learning frameworks for robust and effective intrusion detection in SCADA systems. International Journal of Critical Infrastructure Protection, 49, 100759. [5]             Khadidos, A. O., Khadidos, A. O., Selvarajan, S., Al-Shehari, T., Alsadhan, N. A., & Singh, S. (2025). CyberSentry: Enhancing SCADA Security through Advanced Deep Learning and Optimization Strategies. International Journal of Critical Infrastructure Protection, 100782. [6]             Ramesh, R. B., Thangaraj, S. J. J., Sagayee, G. M. A., & Saravanan, K. (2025). Detection and prevention in WSN security framework using deep learning against black hole and wormhole attacks. Ain Shams Engineering Journal, 16(10), 103624. [7]             Kilincer, I. F. (2025). Explainable AI supported hybrid deep learnig method for layer 2 intrusion detection. Egyptian Informatics Journal, 30, 100669. [8]             Alkhonaini, M. A., Alohali, M. A., Aljebreen, M., Eltahir, M. M., Alanazi, M. H., Yafoz, A., ... & Khadidos, A. O. (2025). Sandpiper optimization with hybrid deep learning model for blockchain-assisted intrusion detection in iot environment. Alexandria Engineering Journal, 112, 49-62. [9]             Li, X., & Zhang, Y. (2025). Security application of intrusion detection model based on deep learning in english online education. Alexandria Engineering Journal, 124, 582-590. [10]          Elsayed, S., Mohamed, K., & Madkour, M. A. (2024). A comparative study of using deep learning algorithms in network intrusion detection. IEEE Access, 12, 58851-58870.