An Ensemble-Based Machine Learning-Envisioned Intrusion Detection in Industry 5.0-Driven Healthcare Applications

With the emergence of advanced technology and biotechnology, medical institutions like hospitals are increasingly relying on smart devices to create an efficient ecosystem. The Industry 5.0-driven healthcare system has started its focus on personalizing products/services having unique and special needs for patients with various diseases. In the current scenario of smart healthcare, we need to consider a human-centric solution that induces the Internet of Things (IoT), Internet of Medical Things (IoMT) and Artificial Intelligence (AI). The fifth industrial revolution, which is being pushed by Industry 5.0, is noted for its ability to meet the customized needs of both patients and healthcare providers. It offers a product to patients and medical professionals in accordance with their unique needs. The Industry 5.0-driven healthcare system has a vast variety of applications, such as remote consultation, routine health monitoring and support, critical care and alert, etc. However, it also suffers from various security and privacy-related issues as various cyber attacks can be launched on the Industry 5.0-driven healthcare system. Therefore, we need a robust security mechanism to protect sensitive healthcare data and other resources. The machine learning (ML)/deep learning (DL) model can be effective to some extent, ensemble-based models have emerged as a promising approach for addressing the potential security threats. In this article, we propose a novel ensemble-based ML-envisioned scheme to detect different types of intrusions for the Industry 5.0-driven healthcare system (in short EIDS-HS). We validate the proposed EIDS-HS on a standard data set and evaluate its performance using key performance parameters, such as accuracy, precision, recall, F1-score, and computational complexity. The security analysis of the proposed EIDS-HS proves its security against various possible attacks. Furthermore, EIDS-HS performs better than existing intrusion detection schemes in terms of important performance parameters.