Scalable Hybrid Blockchain with PoA-PBFT Consensus for Intrusion Detection and Security in Smart City IoT Ecosystems

Abstract

IoT networks in smart city infrastructure include smart devices that use open channel internet to gather and process data. Centralism, safety, confidentiality, transparency, scalability, verification, and managing the quick adaption of smart cities are some of the issues that have arisen with the current data transport technologies. This study introduces a multi-level architecture that combines deep learning and blockchain technology to provide a safe and private intrusion detection system for smart city settings. At Level 1, a PoA-PBFT-based privacy approach is used to safeguard smart city data, with smart contracts and blockchain ensuring safe data collection, validation, and on-chain storage. This decentralized system improves transparency, data integrity, and tamper resistance. With feature mapping and Binary Tree Optimization (BTO) for effective feature selection, Level 2 emphasizes privacy protection and cycle-consistent optimal data processing. This step preserves important information for precise detection while reducing dimensionality. Using a suggested hybrid CNN-LSTM algorithm based on IChOA, Level 3 deploys an enhanced intrusion detection model that efficiently captures both temporal and spatial attack patterns. Experiments on the proposed method shown that the CNN-LSTM optimized framework based on IChOA is reliable and useful for identifying network intrusions for attack classification, with outstanding recognition accuracy, precision and recall.