The necessity to improve modern medical networks such as internet of medical things (IoMT) to monitor patients and their health condition has raised due to the effects of population ageing, increasing number of patients, deficiency of treatment facilities, and spread of widespread diseases. However, resisting cyber-attacks is a challenging concern for researchers. This chapter proposes an explainable learning machines-based security framework for detecting cyber-attacks against IoMT networks in real-time. The proposed model is based on the phenomenon of extreme learning machines to detect multiple kinds of cyber-space attacks carried against IoMT systems. The authors also explain the detection decisions to expand trust management in the employed machine learning algorithm and facilitate security professionals to comprehend the undiscovered data evidence and causal inference. Experiments show the effectiveness of the proposed approach signifying its utility as a workable security framework in contemporary networks of IoMT-based healthcare systems.
TopIntroduction
The speedy development in communication technologies and sensing tools make the Internet of Things (IoT) networks capable to link numerous physical entities (Idrees, Alhussaini, and Salman, 2020). This progression led to the origination of various applications of IoT, such as intelligent transportation systems, smart homes, ambient assisted living, remote healthcare monitoring, etc. (Idrees and Al-Qurabat, 2021). The medical field has progressed so much by applying IoT tools and applications to design a bright term named the Internet of Medical Things (IoMT) (Khan et al., 2022). According to a report (Salman et al., 2020), the IoT sector will be instituting of more than 5.8 billion items by the end of year 2021, out of which the 40% of the devices will be related to IoMT. More significantly, this report also mentioned that the IoMT industry is projected to become $136.8 billion industry worth, thus concluding that the reliance of medical field on IoMT-based devices can save up to $300 billion.
The progression of IoMT networks motivated by the enlarged amount of connected medical devices that are capable to produce, gather, evaluate, fuse, and transmit the sensed healthcare data to Cloud or any receiver. The IoMT networks are composed of data collected from several biosensor and healthcare devices and applications (Nehra et al., 2021). These IoMT devices are employed to gather clinical data, monitor the real-time health of the patient, and communicate it to health professionals through the remote data centers. Which is why the core objective of IoMT is to enhance the quality of treatment and connected medical systems. Because such medical systems and healthcare applications necessitate speedy response and decision system to support the emergency cases, through the usage of high bandwidth IoMT network for transferring the sensed data. These requirements signify great challenges in the implementation process of an IoMT network. However, a key problem is to defend these networks against evolving cyber-attacks. The healthcare data from these networks needs to be kept safe at various levels (creation, collection, communication, and storing levels) to protect the privacy and security of both data and patients. According to the writers of (CyberMDX, 2020), nearly 50% of the IoMT-centric healthcare devices and tools are exposed to cyber-attacks and exploits. Since the IoMT networks involves identities of patients, the healthcare data costs on average 50 times more as related to other categories of data, subsequently making it enormously treasured on the dark-web and black-market (Maddox, 2019).
Despite the fact that linked IoMT devices bring numerous benefits, they also increase the issues like security against cyber-attacks since these systems process sensitive and life-critical data (CyberMDX, 2020). The attacks against IoMT systems could have substantial physical and lethal harm to patients. E.g., an attack on connected insulin healthcare pump or pacemaker device could actually lead to patients’ death if the levels of insulin or pacemaker device are manipulated. Several researchers have pointed out that attacks like eavesdropping, DoS, identity theft, false data injection etc., can actually disturb the security parameters thus threatening the availability of these life-critical devices in IoMT networks (Yaqoob, Abbas, & Atiquzzaman, 2019), or in worst cases it can push the IoMT ecosystem into a complete chaos (CyberMDX, 2020).