The AUTHORIZED BLOCK MINING-BASED INTRUSION DETECTION SYSTEM IN BLOCK-CHAIN ENABLED IOT DEVICES USING HOMOMORPHIC SIGNATURES AND GRU (GATED RECURRENT UNITS) WITH CNN HYBRID (GRU-CNN)

BLOCK-CHAIN ENABLED IOT DEVICES

Abstract

Background Information: IoT devices have increased connection, but because of their decentralized architecture and constrained processing capability, they have also increased cybersecurity vulnerabilities. By addressing vulnerabilities to both known and unknown cyberattacks, integrating blockchain technology with intrusion detection systems (IDS) improves data security, privacy, and trustworthiness in Internet of Things networks.

Methods: GRU-CNN hybrid models are used to detect IoT abnormalities, XGBoost is used for feature selection, and PoA (Proof-of-Authority) is used for trusted block mining in the proposed Authorized Block Mining-Based Intrusion Detection System (ABM-IDS). Enhancing the security and scalability of the system, homomorphic signatures guarantee cryptographic verification of messages without disclosing data.

Objectives: In order to increase intrusion detection accuracy, boost block mining trust, and decrease detection latency for IoT devices, this project intends to construct a secure and effective IDS for IoT. It will do this by utilizing homomorphic signatures, GRU-CNN models, and PoA consensus.

Results: ABM-IDS outperformed traditional machine learning methods such as ECDEA, achieving 99.1% accuracy, 98.95% precision, and an extremely low detection latency of 7.5 seconds.

Conclusion: Through the use of cutting-edge cryptographic algorithms and deep learning models, the ABM-IDS provides a more precise, economical, and scalable intrusion detection solution, enhancing the safety, effectiveness, and trustworthiness of IoT networks.