Integration of Cybersecurity, AI, and IoT Technologies

A LoRaWAN application architecture comprises three functional components: (i) nodes, which convert and wirelessly transmit data as LoRaWAN messages; (ii) gateways, which receive and forward these transmissions; and (iii) network servers, which process the received data for application delivery. The nodes convert data into LoRaWAN messages and transmit them wirelessly with the hope that one or more LoRaWAN gateway will receive the messages successfully. Then, the gateways pass on the received messages to a distant network server, where various processing steps occur before the messages are forwarded to the end application. If none of the gateways can receive the messages, then they will be lost. Although this default behaviour is suitable for some applications, there are others where ensuring messages are successfully delivered at a higher rate would be helpful. One such scenario is the application in this paper: monitoring maritime vessels and fishing equipment in offshore environments characterised by intermittent or absent shore connectivity. To address this challenge, the Custodian project was initiated to develop a maritime monitoring solution with enhanced connectivity capabilities. Two additional features are especially welcome in this scenario. The most important feature is the transmission of messages created in offshore areas to end users who are offshore, regardless of the unavailability of the ground network server. An example would be fishermen who are offshore and wish to position their fishing equipment, also offshore, based on location data transmitted from nodes via LoRaWAN, even when both entities are far away from the mainland. The second aspect concerns the potential use of gateway-to-gateway communications, through gateways on various ships, to transmit messages to the coast. This setup enables fishing gear and fishing vessels to be monitored from the coast, even in the absence of a direct connection. The functional constraints of conventional commercial gateways necessitated the conceptualisation and implementation of C-Mesh, a novel relay architecture that extends LoRaWAN functionality beyond standard protocol implementations. The C-Mesh integrates with the Custodian ecosystem, alongside C-Beacon and C-Point devices, while maintaining transparent compatibility with standard LoRaWAN infrastructure components through protocol-compliant gateway emulation. Thus, compatibility with both commercially available nodes and gateways and those already in deployment is guaranteed. We provide a comprehensive description of C-Mesh, describing its hardware architecture (communications, power, and self-monitoring abilities) and data processing ability (filtering duplicate messages, security, and encryption). Sea trials carried out on board a commercial fishing vessel in Sesimbra, Portugal, proved C-Mesh to be effective. Location messages derived from fishing gear left at sea were received by an end user aboard the fishing vessel, independently of the network server on land. Additionally, field tests demonstrated that a single C-Mesh deployment functioning as a signal repeater on a vessel with an antenna elevation of $15m$ above sea level achieved a quantifiable coverage extension of 13 km (representing a 20% increase in effective transmission range), demonstrating the capacity of C-Mesh to increase LoRaWAN’s coverage. Full article

The size of data transmitted through various communication systems has recently increased due to technological advancements in the Artificial Intelligence of Things (AIoT) and the industrial Internet of Things (IoT). IoT communications rely on intrusion detection systems (IDS) to ensure secure and reliable data transmission, as traditional security mechanisms, such as firewalls and encryption, remain susceptible to attacks. An effective IDS is crucial as evolving threats continue to expose new security vulnerabilities. This study proposes an integrated approach combining feature selection methods and principal component analysis (PCA) with advanced deep learning (DL) models for real-time intrusion detection, significantly improving both computational efficiency and accuracy compared to previous methods. Specifically, five feature selection methods (correlation-based feature subset selection (CFS), Pearson analysis, gain ratio (GR), information gain (IG) and symmetrical uncertainty (SU)) were integrated with PCA to optimise feature dimensionality and enhance predictive performance. Three classifiers—artificial neural networks (ANNs), deep neural networks (DNNs), and TabNet–were evaluated on the RT-IoT2022 dataset. The ANN classifier combined with Pearson analysis and PCA achieved the highest intrusion detection accuracy of 99.7%, demonstrating substantial performance improvements over ANN alone (92%) and TabNet (94%) without feature selection. Key features identified by Pearson analysis included id.resp_p, service, fwd_init_window_size and flow_SYN_flag_count, which significantly contributed to the performance gains. These results indicate that combining Pearson analysis with PCA consistently improves classification performance across multiple models. Furthermore, the deployment of classifiers directly on the original dataset decreased the accuracy, emphasising the importance of feature selection in enhancing AIoT and IoT security. This predictive model strengthens IDS capabilities, enabling early threat detection and proactive mitigation strategies against cyberattacks in real-time AIoT environments. Full article

The research focuses on the limitations of traditional systems in optimizing information flow in the healthcare domain. It focuses on integrating knowledge graphs (KGs) and utilizing AI-powered applications, specifically conversational agents (CAs), particularly during peak operational hours in emergency departments (EDs). Leveraging the Cross Industry Standard Process for Data Mining (CRISP-DM) framework, the authors tailored a customized methodology, CRISP-knowledge graph (CRISP-KG), designed to harness KGs for constructing an intelligent knowledge base (KB) for CAs. This KG augmentation empowers CAs with advanced reasoning, knowledge management, and context awareness abilities. We utilized a hybrid method integrating a participatory design collaborative methodology (CM) and Methontology to construct a domain-centric robust formal ontological model depicting and mapping information flow during peak hours in EDs. The ultimate objective is to empower CAs with intelligent KBs, enabling seamless interaction with end users and enhancing the quality of care within EDs. The authors leveraged semantic web rule language (SWRL) to enhance inferencing capabilities within the KG framework further, facilitating efficient information management for assisting healthcare practitioners and patients. This innovative assistive solution helps efficiently manage information flow and information provision during peak hours. It also leads to better care outcomes and streamlined workflows within EDs. Full article