Intrusion detection system for piezoelectric umbilical cords data using hybrid ML-DL model with blockchain model

To secure communication on piezoelectric umbilical cables of vital infrastructure, this paper proposes a new and potent Intrusion Detection System (IDS). The proposed solution will be based on a Decentralized Blockchain Architecture and a Hybrid ML-DL model. CNN-XGBoost-PSO model, which lies at the core of the detection engine, approaches the particular issue of classifying time-series sensors data correctly, where data is recorded at high frequencies. To train by learning complex features of time and space, the Convolutional Neural Network (CNN) automatically detects and isolates anomalies in the raw sensor responses very fast and with great accuracy. These characteristics are then inputted in the highly efficient XGBoost classifier. Particle Swarm Optimization (PSO) is used to fine-tune the critical parameters of the combined approach that provide more effective performance metrics than traditional ones, including a low False Positive Rate (FPR) and high accuracy. Combining the Blockchain Model further guarantees that any security events may be audited, non-repudiable and be stored in an immutable withdrawn registry, which is critical to avoid intrusion notifications. This combined architecture provides a solid defense to the modern digital and physical threats because it is demonstrated experimentally to significantly enhance the privacy and dependability of the electromechanical umbilical cord data.