Secure IOT Framework: Blockchain Authentication and Intrusion Detection

Blockchain security leverages decentralized, immutable ledger technology for identity verification and data integrity but faces challenges like high computational overhead, scalability issues, and privacy-explainability trade-offs. To overcome these, this research proposes a framework integrating the Variable Elliptical Anonymous Candidate protocol (VEAQS) with blockchain for secure authentication, ensuring transparency and efficiency. The Quantum Kernel Convoluted Gazelle Self-attention model (QKCGS) enhances intrusion detection by leveraging quantum kernels and self-attention mechanisms. It efficiently handles sequential and contextual data, improving both short- and long-term dependency learning. The model uses quantum gates to regulate past knowledge retention and irrelevant data filtering, boosting detection accuracy. By integrating a deep learning model with SHAP, the framework enhances the interpretability and explains the ability of intrusion detection deployed in the Internet of Things, improving the detection of complex attack patterns while ensuring transparent decision-making. The model further optimizes authentication processes by minimizing computational overhead and latency through Single Candidate Optimization (SCO), streamlining the key generation process. The Gazelle Optimization Algorithm (GOA) is used to fine-tune the intrusion detection systems, making it more responsive to real-time attacks while maintaining the integrity and security of the authentication process. This combination of deep learning, blockchain security, and explainable Artificial Intelligence (AI) creates a robust, efficient, and transparent security solution for modern Internet of Things systems, enabling secure communication, tamper-proof logging, and real-time attack detection. The proposed methods achieve 99.50% of accuracy, 99.52% precision, and 99.60% F1-score, demonstrating significant improvements over existing methods.