As the number of IoT devices grows, ensuring the secure validation and processing of the data they generate has become critical. This challenge becomes even more pronounced for mobile nodes, such as vehicles, which must continually reconnect to new Edge Servers while in motion. This work proposes a decentralized architecture for connecting IoT devices to Edge Servers, enabling secure data delivery to applications while minimizing overhead and ensuring trustworthy handovers between Edge Servers. A fundamental concern is the integrity of Edge Servers, as they may be compromised or exhibit malicious behavior. To address this, the proposed architecture relies on an external verifier service to continuously verify the integrity of Edge Servers. To demonstrate its feasibility, two prototypes were implemented using distinct consensus technologies and evaluated under realistic conditions. The first prototype, based on BFT-SMaRt, achieved lower latency and higher throughput but required dedicated, proprietary infrastructure and lacked global auditability. In contrast, the second prototype, leveraging an existing blockchain network, provides complete auditability and decentralization without proprietary infrastructure, though at the cost of higher latency. Experimental results confirm that both approaches deliver strong security guarantees, with trade-offs between performance and transparency, validating the architecture’s suitability for dynamic IoT environments.

Mixture-of-Experts (MoE) models enhance deep learning scalability by activating only a subset of specialized subnetworks per input. However, conventional MoEs often suffer from unstable expert specialization and biased routing due to entangled gradient updates and unstructured input handling. To address these problems, this paper proposes SEAS-GMoE ( S tructured E xpert A ssignment and S upervised G ating M ixture o f E xperts), a hybrid framework that decouples expert routing from gradient noise through a double-stage feature clustering and semantic pseudo-labeling mechanism. SEAS-GMoE employs a K-means and KNN-based cluster refinement technique, followed by a Siamese Neural Network (SNN) that assigns semantic pseudo-labels to clustered features. This leads to generating interpretable and semantically coherent clusters that are explicitly mapped to dedicated experts. A supervised managing (gating) network learns these mappings through a bidirectional training process, reinforcing both expert specialization and routing reliability. In the bidirectional training process, as expert specialization improves, the gating supervision signal is simultaneously enhanced, and vice versa. During inference, soft expert routing is applied through the managing network’s probabilistic output. This enables flexible aggregation of expert(s)’ predictions on new inputs. Experimental results on GTSRB, MNIST, and CIFAR-10 show that SEAS-GMoE outperforms reimplemented V-MoE and dense baselines, achieving up to 4.1% higher accuracy, 60% lower inference latency, and 11% greater specialization stability. These results confirm SEAS-GMoE’s effectiveness for robust, interpretable, and efficient inference.

• A novel stacked ensemble model that enhances predictive accuracy for solar irradiance, load, and energy prices. • A hierarchical recommendation heuristic integrating solar, battery, and grid energy sources for real-time cost optimization. • Promotion of the development of new research on renewable energy use and technological innovation. In recent years, the integration of solar power systems has seen substantial growth within smart houses. However, the inherent intermittency of solar irradiance introduces fluctuations in the available solar power, presenting a challenge for stable energy management. Developing and implementing precise recommendation systems based on solar power, energy market price, and weather forecasting methods become imperative in strategic planning and seamless energy management systems to address this issue effectively. This paper presents the A rtificial In T elligence-driven I oT RE comme NDE r E nergy System (ATIRENDEE) designed to enhance energy management in smart houses. Leveraging deep learning algorithms, a novel stacked ensemble model is developed to optimize energy consumption. Our results support the ATIRENDEE approach’s significant reduction in energy costs for smart homes, achieving an average monthly savings of over 45 % and reaching up to 68 % in specific cases. The multi-layered recommendation approach used enhances the efficiency and adaptability of the ATIRENDEE solution. By dynamically optimizing energy allocation among solar, battery, and grid sources, ATIRENDEE consistently outperformed static grid-based approaches, promoting cost-effective and sustainable energy use in eco-friendly smart houses or buildings.

Container virtualization has recently gained popularity due to its low performance and resource allocation overhead. The rise of this technology can be attributed to the advancement of cloud computing and the adoption of micro-services architecture. These new approaches offer a more efficient and fine-grained system design through the benefits of containerization, such as isolation, portability, and improved performance. However, container-based systems have created new challenges in monitoring due to their automated flexibility, ephemerality, and the increasing number of containers in a system. So there is a practical need for effective monitoring and performance management tools. This paper analyses the key performance metrics for machine, container and application services, including CPU usage, memory usage, disk usage, and network usage. Furthermore, we review several widespread tools for collecting and monitoring these metrics and present the Monintainer tool. It is a solution designed to monitor entire container-based systems, from applications to their underlying infrastructure, allowing users to better understand their systems’ behaviour in run-time. The tool’s results can aid container-based systems’ design, implementation and optimization.

Due to the increasing demand for Internet of Things (IoT) applications with sensitive data, the use of encryption on constrained devices is crucial for protecting and ensuring privacy and security. The devices used in such applications have limited computational power, memory, and energy resources, making them vulnerable to attacks that exploit these limitations. Lightweight cryptography has been growing fast in recent years aiming to circumvent the lack of security in these types of devices. We propose a modular solution that capitalizes on the strengths of lightweight cryptography to offer an approach that can be easily adapted to meet the needs of private wireless sensor networks. The solution is designed to work in multi-hop communication networks, where Nodes out of range of the Gateway can be part of the network, offering the same security level that a Node in the communication range of the Gateway has. To further enhance the security and lifetime of the network, our solution incorporates a key renewal mechanism and supports the use of signature schemes for integrity. Additionally, it is designed to be distributed to achieve high levels of scalability. Experimental results show that using our solution on top of standard communication protocols does not introduce significant overhead in terms of performance.

The VEDLIoT project targets the development of energy-efficient Deep Learning for distributed AIoT applications. A holistic approach is used to optimize algorithms while also dealing with safety and security challenges. The approach is based on a modular and scalable cognitive IoT hardware platform. Using modular microserver technology enables the user to configure the hardware to satisfy a wide range of applications. VEDLIoT offers a complete design flow for Next-Generation IoT devices required for collaboratively solving complex Deep Learning applications across distributed systems. The methods are tested on various use-cases ranging from Smart Home to Automotive and Industrial IoT appliances. VEDLIoT is an H2020 EU project which started in November 2020. It is currently in an intermediate stage with the first results available.

The Internet of Things has been applied in different application contexts, ranging from industrial, healthcare or simple home life applications. It allows adding a new level of automation to objects, making them appealing to everyone. Since several applications involve data, and in some cases, it is sensitive, those applications are more vulnerable to be attacked. As a result, researchers are constantly exploring secure mechanisms to keep the data and its owners safe. For the resource-constrained nature of these IoT applications, the design of cryptography solutions becomes a challenging task because IoT nodes should be cheap and require low power, which means lower computational performance. Lightweight cryptography algorithms are attractive solutions to reduce computation complexity, keeping the desired level of security. This work provides an experimental performance analysis of several cryptography solutions targeted for embedded platforms to help choose better algorithms for IoT systems in terms of efficiency. Execution time, code efficiency and communication influences are evaluated for a set of cryptography algorithms. The results of this work intend to help IoT developers to choose a suitable cryptography algorithm for their IoT applications.

Collecting real-time information on rivers and bays is essential to identify, predict, and manage critical situations, such as floods or chemical detection, which may damage the environment. The LoRa technology has emerged as a solution for low power, long-range IoT applications. Its capabilities provide either a long-range at the cost of a lower data rate or higher throughput at a shorter range. In this paper, we study the capabilities of using LoRa in deployments of private networks for IoT applications. First, we consider the problems that may arise when LoRa nodes are deployed near large water bodies, with antennas at low heights, where the communication link are affected by the presence of the water and nodes maybe not in the range of a gateway. In these cases, the standard LoRaWAN protocol can’t be used because it is not designed for node-to-node communication. We propose an improved LoRa reliable protocol for remote monitoring applications (AQUAMesh) that can replace LoRaWAN. It uses the LoRa physical layer, which provides long-range communications and adds the capability to deal with point-to-point communications using s custom time-division multiple access (TDMA). This proposed protocol overcomes the shortcomings of the existing LoRaWAN and can be effectively deployed in private networks of various applications.

Clinical laboratories use a myriad of medical devices (e.g., automated analyzers) to determine the results of its tests. Modern laboratories have evolved towards almost complete automation, by making use of laboratory information systems (LIS) that interact with medical devices in an automatic way using real-time communication protocols. However, laboratories have been increasing in terms of size and complexity mostly due to economic reasons: large groups of geographically distributed laboratories have been replacing small laboratories. This trend poses challenges for traditional LIS architectures based on a centralized system, given that the distributed nature of modern laboratories has a negative impact on the performance of the communication between medical devices and LIS, and also increases the risk of temporary network disconnection. In this paper we address these challenges by proposing a novel architecture that seamlessly integrates distributed medical devices using mechanisms that provide real-time operation and ensure fault-tolerance. The proposed system is expected to continuously provide trustworthy services even in critical scenarios.

Distributed Control Systems in industrial settings need to be carefully dimensioned to provide correct operations timings. In heterogeneous settings, with wired and wireless components, there are many factors that have implications regarding timings and other issues such as security and fault-tolerance. In this paper we review the most important existing protocols and strategies for wired, wireless and hybrid industrial distributed control systems, focus on the problem of planning in both Fieldbus and heterogeneous industrial control systems, describe and test a solution and finally we indicate a set of open research issues.

One of the major driving forces behind technology development is the establishment of equality among everyone. It is intended that the strategic use of technology complements competencies or diminished capacities, improving self confidence and autonomy. Blind and partially sighted people often deal with this inequality in the performance of tasks or daily activities. The presented work reviews some of the proposed technologies to assist visually impaired people in common activities, specifically navigation and going shopping.

Deployment of embedded systems in industrial environments requires preconfiguration for operation, and, in some contexts, easy reconfiguration capabilities are also desirable. It is therefore useful to define a mechanism for embedded devices that will operate in sensor and actuator networks to be remotely (re)configured and to have flexible computation capabilities. We propose such a configuration, reconfiguration, and processing mechanism in the form of a software architecture. A node component should be deployed in any embedded device and implements application programming interface (API), configuration, processing, and communication. The resulting system provides remote configuration and processing of data in any node in a most flexible way, since every node has the same uniform API, processing, and access functionalities. The experimental section shows a working deployment of this concept in an industrial refinery setting, as part of the EU FP7 project Ginseng.

There have been both research and commercial advances on applying Wireless Sensor and Actuator Networks (WSN) in industrial premises. These have cost advantages related to avoiding some cabled deployments. A possible architecture involves a Networked Control System (NCS) with many small WSN subnetworks, cabled nodes and computer servers (e.g., servers, control stations). In those systems individual sensor nodes can be programmed, as opposed to cabled analog systems. We investigate approaches for networked-wide configuration, where all nodes—cabled or WSN sensors—can be configured with simplicity from a single interface, instead of hand-coding or complex configurations of individual nodes. We propose an architecture and approach for configuration and operation. Previous related proposals on middleware involving WSNs suffer from two major limitations: they either program within an individual WSN or configure operation outside WSNs, wrapping data coming from WSN. They do not allow configuring WSN and non-WSN nodes for operation from a single interface. We discuss the architecture and propose the NCSWSN configuration and operation approach. We are applying this system in an industrial testbed, therefore we test the approach and also show user interfaces and results from the deployment.

Today's industrial facilities, such as oil refineries, chemical plants, and factories, rely on wired sensor systems to monitor and control the production processes. The deployment and maintenance of such cabled systems is expensive and inflexible. It is, therefore, desirable to replace or augment these systems using wireless technology, which requires us to overcome significant technical challenges. Process automation and control applications are mission-critical and require timely and reliable data delivery, which is difficult to provide in industrial environments with harsh radio environments. In this article, we present the GINSENG system which implements performance control to allow us to use wireless sensor networks for mission-critical applications in industrial environments. GINSENG is a complete system solution that comprises on-node system software, network protocols, and back-end systems with sophisticated data processing capability. GINSENG assumes that a deployment can be carefully planned. A TDMA-based MAC protocol, tailored to the deployment environment, is employed to provide reliable and timely data delivery. Performance debugging components are used to unintrusively monitor the system performance and identify problems as they occur. The article reports on a real-world deployment of GINSENG in an especially challenging environment of an operational oil refinery in Sines, Portugal. We provide experimental results from this deployment and share the experiences gained. These results demonstate the use of GINSENG for sensing and actuation and allow an assessment of its ability to operate within the required performance bounds. We also identify shortcomings that manifested during the evaluation phase, thus giving a useful perspective on the challenges that have to be overcome in these harsh application settings.

When wireless sensor networks are introduced in industrial settings, they will be part of a much larger heterogeneous sensor and actuation network that includes those sub-networks together with Ethernet cabled Programmable Logic Controllers (PLCs) and control stations. Performance issues arise in such systems. Particularly, actuation latencies are considered crucial in those scenarios. In this paper, we present an approach for planning and evaluating a heterogeneous supervisory control and data acquisition (SCADA) system with wireless sensor networks. We propose closed-loop schemas and mechanisms for measuring and verifying the latency bounds in the whole system. As a case study, we deploy and evaluate our approach on an oil refinery, where several heterogeneous networks are used to monitor and control a water treatment process. Results demonstrated the effectiveness of our method in planning, debugging and providing feedback about the network (in)ability to provide timely actuation guarantees.

Parallel Shared-Nothing architectures are frequently used to handle large star-schema Data Warehouses (DW). The continuous increase in data volume and the star-schema storage organization introduce severe limitations to scalability due to the well-known parallel join issues and the resulting need to use solutions such as on-the fly repartitioning of data or intermediate results, or massive replication of large data sets that still need to be joined locally, constraining their ability to deliver fast results. Parallelism may improve query performance, however some business decisions may require that query results be timely available which, even with additional parallelism and significant upgrade costs (both monetary and due to disturbance of normal operations), cannot be guaranteed. We propose a Timely-aware Execution Parallel Architecture (TEEPA) which balances data load and query processing among an elastic set of non-dedicated heterogeneous nodes in order to provide scale-out performance and timely query results. Data is allocated using adaptable storage models to minimize join costs (the major uncertainty factor) which best fit the nodes' capabilities, while preserving a consistent logical view of the star-schema. We present experimental evaluation of TEEPA and demonstrate its ability to provide timely results.

The advancements in communications and embedded systems have led to the proliferation of wireless sensor and actuator networks (WSANs) in a wide variety of application domains. One important key of many such WSAN applications is how to configure/program functionalities after deployment. Some application domains even require that sensor nodes be deployed in harsh environments (e.g., refineries), where they need to be configured over-the-air. Over-the-air programming takes more than 4 second to transmit an image code to a node where, typically, a single image has 40kB. With the proposed model, small code blocks (states and transitions) are generated (about 4kB) to upload to nodes. This methodology allows reducing the programming time to 5 times less, release bandwidth and keep battery on nodes. To support our approach we propose an architecture based on state machines model which provides an easy user-readable high-level representation of states and transitions. The communication time is reduced through the reduction of the code image size that it is needed to send to each node. We develop an architecture that allows users to define and program applications based on Markov techniques, to simultaneously facilitate the application design and take into account important requirements such as reliability. The experimental section shows a working deployment of this concept in an industrial refinery setting.

In this paper we describe an approach to provide over-the-air operation configuration capabilities and end-to-end operation timing monitorization and control in networked control systems (NCS) with wireless sensor network (WSN) components. The idea is to deploy wireless sensor networks to cover regions of the factory premises, plugging each into a NCS through a gateway, then configuring nodes for operation using simple commands, as opposed to programming and uploading every node. The system should also provide deterministic timing control. In this paper we describe the configuration approach and timing planning, then we show the configuration user interface and operation timing monitoring results.

Applications for Sensor and Actuator Networks (SANs) are being spread to areas in which easy configuration by non-programmers will enhance acceptance, and also contexts where reconfiguration is needed during the application lifespan. Wireless SAN (WSAN) may include wired and/or wireless devices, PCs and control stations arranged in a heterogeneous distributed system. Instead of assuming that embedded device nodes (e.g. a TelosB mote) and gateways (e.g. PC running Linux) are disparate entities with their own programming and processing model, it should be viewed as a single heterogeneous distributed system, offering more uniformity, simplicity and flexibility. In this paper we propose MidSN - an approach to hide heterogeneity and offer a single common configuration and processing component for all nodes of that heterogeneous system. The main contribution of this paper is how to design a middleware architecture with a single uniform component to work with such heterogeneous underlying parts as a WSAN. We have successfully applied an embodiment of this concept as middleware component in an industrial refinery setting and in our lab testbed within EU project Ginseng.

Star schema model, has been widely used as the facto DW storage organization on RDBMS. Business measures are stored in a central fact table along with a set of foreign keys referencing dimension tables. While this storage organization offers a good trade-off between storage size and performance for a single node, it doesn't scale in a predictable manner in shared-nothing parallel architectures. Although fact tables can be linearly partitioned among nodes, the same doesn't apply to dimensions, which unbalances (increases) the dimensions/fact_table size ratio, and consequently introduces limits to the number of parallel nodes. In this paper we propose and evaluate a parallel DW storage model, that overcomes these limitations and deliver optimal speed-up and scale-up capabilities with top efficiency. We use the TPC-H benchmark to evaluate the scalability and efficiency of the proposed model.

Wireless sensor networks (WSNs) are deployed to sense, monitor and act on the environment. Deployments in scenarios such as industrial sense and react environments require a set of functionality, and for both ease and reconfiguration capabilities it is important to offer an appropriate framework. We propose a framework for interaction with real-world devices by abstracting proprietary protocols, allowing the interaction between client applications and heterogeneous sensor network platforms and protocols. The framework allows the (re)configuration of alarms, actions or closed-loop techniques, offering flexibility and the possibility to modify for providing performance control guarantees. It allows users to configure and apply various operations, including complex closed-loop techniques that monitor and act over any actuator in the WSN, independently of the underlying WSN infrastructure. The framework is being deployed in a real scenario in the context of European FP7 GINSENG project (wireless networks with performance control guarantees).