Utilizing Deep Learning To Enhanced Security In the Internet of Medical Things Via Intrusion Detection Systems

Abstract

The incorporation of the Internet based Things (IoT) into medical applications has significantly improved healthcare operations and patient treatment. Real-time patient monitoring systems, coupled with remote diagnostics using Internet based Medical Things (IoMT) technology, empower physicians to efficiently handle more cases and potentially save lives. However, IoMT devices are susceptible to cybersecurity threats, posing risks to data security and privacy. Due to constraints in computing power and memory utilization of IoMT devices, implementing traditional security measures becomes impractical. This article introduces a groundbreaking system, ParticleSwarmNetGuard (PS-NG), which combines Element Swarm Optimization with a Deep Neural based Network structure to establish a robust intrusion detection system in IoMT. This innovative system exceeds current data-security standards and achieves an impressive 96% accuracy in detecting network intrusions by utilizing a combined dataset of network traffic and patient sensing data. Furthermore, the performance analysis conducted compares numerous Machine Learning type mechanism (ML) techniques aimed at network intrusion detection in IoMT, confirming the superior performance of DL models over ML models. © 2024 IEEE.