IoTHS-based framework for developing a healthcare security system

The Internet of Things (IoT) has greatly revolutionized the electronic healthcare sector by allowing the centralization of patient care, enabling more efficient connections, and providing real-time monitoring. The Internet of Medical Things (IoMT) is a network of wirelessly connected sensors and devices that may be affixed to a person’s body to automatically control medication administration and track health data in real time. IoMT has garnered considerable attention in recent months as a potential approach to delivering healthcare services via remote access. Significant advancements and innovative methods have resulted from the Internet of Things Health Systems’ (IoTHS) quick growth and convergence in the healthcare sector, but they have also brought to light serious weaknesses, especially in cybersecurity. By enabling medical equipment to transmit patient health information online, IoTHS introduces the risk of data breaches. To make the healthcare system more confidentiality of patients’ sensitive health information is crucial. To build an intrusion detection system (IDS) that protects patient health data and lowers hazards to the Internet of Things (IoTHS), this study advocates for the employment of deep learning techniques, including Visual Geometry Group (VGG) 16, long short-term memory (LSTM), and fully connected layer multilayer perception (FCLMLP). The model was adjusted to meet the intricate security requirements of the IoTHS using a standard dataset comprising three types of instances: attacks, patient monitoring, and environmental monitoring. The model is also evaluated thoroughly. To improve our understanding, we identified which elements of the decision-making process most contributed to our overall knowledge using the Shapley additive explanations (SHAP) technique. The results show that the FCLMLP model achieves an accuracy of 99.79%. The suggested model’s mean squared error (MSE) and Matthews correlation coefficient (MCC) was 0.10 and 0.99, respectively, according to error analysis. It has been shown that modern deep learning algorithms can successfully detect threats and provide sufficient security for IoT healthcare systems.