The development of intelligent Industrial Internet of Things (IIoT) systems promises to revolutionize operational and maintenance practices, driving improvements in operational efficiency. Anomaly detection within IIoT architectures plays a crucial role in preventive maintenance and spotting irregularities in industrial components. However, due to limited message and processing capacity, traditional Machine Learning (ML) faces challenges in deploying anomaly detection models in resource-constrained environments like LoRaWAN. On the other hand, Federated Learning (FL) solves this problem by enabling distributed model training, addressing privacy concerns, and minimizing data transmission. This study explores using FL for anomaly detection in industrial and civil construction machinery architectures that use IIoT prototypes with LoRaWAN communication. The process leverages an optimized autoencoder neural network structure and compares federated models with centralized ones. Despite uneven data distribution among machine clients, FL demonstrates effectiveness, with a mean F1 score (of 94.77), accuracy (of 92.30), TNR (of 90.65), and TPR (92.93), comparable to centralized models, considering airtime of Trainning messages of 52.8 min. Local model evaluations on each machine highlight adaptability. At the same time, the performed analysis identifies message requirements, minimum training hours, and optimal round/epoch configurations for FL in LoRaWAN, guiding future implementations in constrained industrial environments.

Human-in-the-loop cyber-physical Systems use data from various sources to provide valuable assistance to users, but privacy concerns arise when sensitive information is shared. Federated Learning is a promising solution that enables the processing of user data without sharing sensitive information. While this method holds great potential, its efficacy in detecting sleep problems remains an open question. In this way, using a real-world dataset application, our study meticulously evaluates and comprehends the impact of incorporating Federated Learning on sleep detection. Our study evaluates the impact of incorporating Federated Learning on sleep detection and compares it with traditional Machine Learning models. Our findings reveal that our approach delivers accurate sleep detection results over 84% on par with conventional techniques. Our results emphasize the critical importance of handling human error inputs, as this factor significantly influences the accuracy of results in both methods.

The evolution of the Internet of Things (IoT), progressing from a hyper-connected world of seamless device interaction to a paradigm that emphasizes the human factor as a central component, is driven by approaches like human-in-the-loop IoT (HITL-IoT). This shift enables the development of innovative systems in various environments. For example, in the context of a smart city, which comprises diverse solutions aimed at enhancing citizens’ quality of life and urban service efficiency, HITL-IoT approaches are employed to promote sustainability. Given the strong human involvement in these systems, challenges related to personal data protection, transparency, and participation must be addressed. Additionally, ensuring interoperability among different smart city entities is crucial for the coexistence and synergy of these solutions. This article presents a model that incorporates contemporary technologies like Blockchain and LoRa to integrate citizen-centric IoT privacy preservation solutions, fostering the vision of smart and sustainable cities. To demonstrate the feasibility of the proposed model, a prototype was designed, implemented, and deployed in a realistic scenario. Assessment results include metrics, such as latency, throughput, and communication aspects of the components.

New concepts based on the Internet of Things propose the integration of the human factor as a key component of novel interconnected ecosystems, to offer them new services and features. However, the data acquisition process, carried out by the devices that allow the deployment of these new paradigms, leads to concerns about what kind of information these entities are able to collect and/or infer. In fact, in most cases, users are not aware that their information is being gathered, nor can they properly control these data flows. Nowadays, the European Union’s General Data Protection Regulation calls for the adoption of transparent practices comprising, inter alia, adequate notification mechanisms, proper consent management schemes, and the development of models to empower the user as an active participant, with full control over the data flows, along with privacy preservation techniques. In this specific context, the aim of this manuscript is to present a review of existing user-centric privacy preservation models. This is, as far as we know, one of the few papers that addresses this particular privacy preserving dimension within an IoT context. Finally, the paper identifies and summarizes open issues, including a discussion of research opportunities in this area.

Citizens are a crucial factor in the sustainability of the cities because they are one of the pillars for balancing their environment. Smart city projects can offer citizens an opportunity to understand how they contribute to their city’s sustainability. Furthermore, an interactive smart city system can create synergies that benefit both the citizens and the city itself. This paper presents the Green Bear smart city platform, an IoT Human-in-the-Loop system that uses LoRaWAN nodes to gather information on city green spaces, bike lanes, and recycling depots. A user-centered mobile application that allows for active user participation, feedback, and incentivization is used to close the loop with humans. This system allows citizens to evaluate their participation in aspects that improve the city’s sustainability through a gamification scheme, obtaining points for different activities in the city’s public spaces and personal activities to improve their quality of life. This solution is being implemented in the city of Coimbra, Portugal. After an overview of the system, the paper describes each main system module. The Green Bear prototype was subject to functional and technological assessment, and the results are presented and discussed.

While the traditional Internet of Things (IoT) relies on electronic sensors/actuators, today's IoT involves a variety of sensors, comprising not only physical, electronic-based devices but also virtual and even human social sensors/actuators. In this context, how can we efficiently and effectively manage different types of IoT data sources? In this article, we propose a solution to manage different types of sensors/actuators and analyze which management protocols are best in terms of performance. The proposal is based on open and broadly adopted technologies in IoT, with emphasis on the FIWARE middleware. We showed that the management of the heterogeneity of the sensing/actuating IoT devices is feasible, by presenting functionalities related to real use cases. We took advantage of the implemented prototype to compare the performances of lightweight machine to machine (LwM2M) and ultralight device management services in FIWARE. In addition to demonstrating the viability of the proposed approach, the obtained results point to mixed advantages/disadvantages of one protocol over the other.

The last decade has witnessed a paradigm change in Industrial Control Systems (ICS), from closed, isolated systems to fully-connected, Internet-capable systems. At the same time, new wireless technologies, mainly coming from the Internet of Things domain, emerged. IEEE 802.15.4 wireless-based standards like WirelessHART, ISA100.11a, ZigBee, and WIA-PA are increasingly used for monitoring industrial processes. Nevertheless, along with this paradigm change, some new threats appeared that menace current industrial infrastructures and economies (e.g., Stuxnet, Mariposa, Slammer). Additionally, in order to keep up with the requirements of new Industry 4.0 applications, sensor nodes software and hardware are becoming more complex and, thus, more prone to faults. In this paper, using a monitoring architecture proposed in our previous work, we injected and subsequently proceeded to detect representative firmware and hardware anomalies (namely, buffer overflow attacks, SPI faults, under-voltage, and high temperature faults) that can be used by attackers to cause major losses or even damage industrial control systems. We evaluated the performance of several machine learning techniques commonly used to detect anomalies (i.e. OCSVM, kNN, AutoEnconder), in order to determine if they could be useful to detect such faults. The obtained results demonstrate that simple and broad scope classifiers, using features that consume little resources, can be developed to detect such faults.

Industrial Control Systems are now exploring the use of Internet of Things technologies not only to make them fitter to their job but also to explore the advantages that come from connecting them to the Internet. Nevertheless, with this paradigm shift, new threats appear, of which the stuxnet worm is just an example, and Intrusion Detection Systems architectures and solutions were and still are being considered. However, most existing projects concentrate on high level system aspects and thus neglect security aspects at wireless communication standards level, such as WirelessHART (the standard with largest market share), choosing not to address security solutions to common, known attacks identified by the community. In this paper, using a monitoring architecture proposed in our previous work, we will monitor a WirelessHART testbed, and at the same time, conduct network attacks from an outsider perspective. As main contributions this paper presents a new exhaustion attack for WirelessHART that, until now, to the best of our knowledge, has not been yet described. Additionally, the presented work proves that using classifiers like One Class Support Vector Machines (OSVM), and our monitoring architecture we are capable to detect the new exhaustion attack and more common attacks like jamming and collision.

Use of wireless sensor networks (WSNs) in conjunction with mobile devices is growing on a daily basis. Mobile phones are not only personal devices that provide communication services, but also versatile and resourceful elements in the Internet of Things (IoT). Nevertheless, despite the referred growth, one key aspect for real-world IoT deployments is still largely overlooked: reliability. In this respect, there is a clear gap between theoretical models and real implementations, as no practical reliability mechanisms are in place for IoT. Proposing, studying, and providing IoT reliability solutions is, thus, fundamental for the massive deployment of IoT technology across all sectors of society. In this context, the main purpose of this paper is to propose, explore, and assess reliability mechanisms for IoT scenarios, based on the use of redundant routes and devices that take advantage of mobile devices, as key elements of IoT. These mechanisms were evaluated in a variety of failure scenarios, both theoretically and through implementations constructed with real, off-the-shelf equipment, showing their good capability to maintain effective communication even in the presence of link failures.

Cloud computing has been pointed out as one of the pillars of the Internet of Things, providing this emerging infrastructure with unlimited resources and easing its deployment. However, cloud computing relies on remote, mainly centralized resource provisioning, which poses substantial problems in the support of time-critical and location-aware IoT applications. Fog computing, on the other hand, recently emerged as an intermediate solution to solve the mentioned cloud computing limitations, but its embryonic status still prevents it from being used in real developments. In this paper we present a fog computing characterization and assess the feasibility of its use in a worst-case scenario, which we name opportunistic fog network. Subsequently, as a first step towards its effective use and deployment in an IoT scenario, we propose a fog computing network architecture, constituted by virtual clusters running on peer-to-peer overlays, capable of abstracting the complexity of lower layers. Opportunistic fog computing can benefit from thousands of end-devices moving in urban areas everyday, providing connectivity and processing resources to every crowded spot, and promoting green computing by allowing a more efficient local resource usage and by decreasing remote communications to the cloud. Furthermore, opportunistic fog computing overcomes cloud computing drawbacks whenever a set of independent and heterogeneous end-devices agree to share their own resources.

The specific capabilities of Wireless Sensor Networks (WSNs), such as fast deployment and flexibility, together with the low cost solutions that can be achieved, bring new opportunities for an all-new range of applications, far from the typical scenarios of low requirements and high redundancy. Some of these new scenarios involve the assurance of predefined performance goals. To guarantee that these goals are achieved, a careful constant monitoring must exist. However, this monitoring depletes nodes energy (especially in middle nodes that have to retransmit the information to the sink), increases interference between nodes and causes an overhead in the used bandwidth, a foreseeable result if measuring performance of a large multilevel WSN. In the protocol to be presented, data fusion is used to enhance the benefits of continuous monitoring while minimizing its implied overhead, using the specific characteristics of WSNs in its favor.

Welcome to Coimbra—home to the oldest university in Portugal and one of the oldest in the world, dating back from 1290—and welcome to the IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM), the intellectual hub and melting pot of ideas from researchers and practitioners interested in all aspects of wireless, mobile, and multimedia pervasive communications. Sponsored by the IEEE Computer Society and its Technical Committee on Computer Communications (TCCC), and co-sponsored by Missouri University of Science and Technology and the University of Coimbra, the 17th WoWMoM is held in Coimbra, Portugal, in the iconic Quinta das Lágrimas Hotel. During its history, the reputation of WoWMoM has rapidly grown and it has become one of the premier conferences in the field of wireless, mobile, and multimedia networking.

Over the last years, Wireless Sensor Networks (WSN) went from being a promising technology for countless industrial applications to a de facto technology used in todays' applications. WSNs have been gaining momentum over costly wired technologies, offering low installation costs, self-organization, and added functionality. As a consequence of their enormous potential, WSNs were subject to standardization and some industrial standards and open source solutions like WirelessHART, Zigbee, ISA100, IEEE802.15.4 and OpenWSN were announced. However, despite considerable efforts to provide mechanisms that increase the availability, reliability, security and maintainability of this type of networks, WSNs have kept one of their main characteristics: fault-proneness. As a result, the offer of post-deployment diagnostic tools has been increasing in the last decade in order to diagnose WSN failures as soon as possible. Nevertheless, current WSN diagnostic tools still have many limitations and cannot be considered “ready to use” in real-world scenarios. In this paper we present an autonomous diagnostic tool that addresses these limitations in a real industrial Internet of Things (IIoT) scenario. Our tool is based on simple metrics, a logging tool, a data-mining algorithm, and available network metrics, and it monitors the condition of the sensor nodes firmware, hardware and the network itself. The proposed demonstration was tested and validated using the WirelessHART IIoT standard.

Smartphones have spawned a new generation of people-centric sensing applications, where their sensors gather information from users to achieve a wide range of objectives, from fitness monitoring to the urban-wide management of traffic. The connections between devices are of utmost importance for these systems, in fact, switching between the multitude of network interfaces available to smartphones (3G/4G, WiFi) can be beneficial to improve connectivity, the distribution of network traffic and save battery power. Current solutions in the literature for the management of network interfaces are limited in the sense that they only consider either application requirements or the system's status. In this paper, we propose a new model that considers not only these aspects but also the human-context, that is, the user's position and activity to intelligently manage which interfaces should be used. To support the switching between different interfaces without interrupting existing connections, we also discuss several handoff techniques. Finally, we present an experimental evaluation of our model based on Multipath TCP.

Mobile sensing in Cyber-Physical Systems has been evolving proportionally with smartphones. In fact, we are witnessing a tremendous increase in systems that sense various facets of human beings and their surrounding environments. In particular, the detection of human emotions can lead to emotionally-aware applications that use this information to benefit people's daily lives. This work presents the implementation of a Human-inthe- loop emotionally-aware Cyber-Physical System that attempts to positively impact its user's mood through moderate walking exercise. Data from smartphone sensors, a smartshirt's electrocardiogram and weather information from a web API are processed through a machine learning algorithm to infer emotional states. When negative emotions are detected, the application timely suggests walking exercises, while providing real-time information regarding nearby points of interest. This information includes events, background music, attendance, agitation and general mood. In addition, the system also dynamically adapts privacy and networking configurations based on emotions. The sharing of the user's location on social networks and the device's networking interfaces are configured according to user-defined rules in order to reduce frustration and provide a better Quality of Experience.

We are living in a new communication age, which will radically transform the way we live in our society. A world where anything will be connected to Internet is being created, generating an entirely new dynamic network - The Internet of Things (IoT) - enabling new means of communication between people, things and environment. A suitable architecture for Internet of Things (IoT) demands the implementation of several and distinct technologies that range from computing (e.g. Cloud Computing), communications (e.g. 6LowPAN, 3/4G) to semantic (e.g. data mining). Thus, it is necessary to understand very well all these technologies in order to know which of them is most suitable for a given scenario. Therefore, this paper proposes an IoT architecture for disabled people and intends to identify and describe the most relevant IoT technologies and international standards for the stack of the proposed architecture. In particular, the paper discusses the enabling IoT technologies and its feasibility for people with disabilities. At the end, it presents two use cases that are currently being deployed for this population.

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.

The growth of mobile computing has increased the demand for wireless communication, causing a higher demand for the wireless medium as well as spectrum pollution. Smart radios, also called cognitive radios, monitor the network to identify the best available channel, in order to avoid interference. This paper proposes a framework for the development and testing of protocols for wireless sensor networks that employ cognitive radios (CRSN). We also developed two spectrum decision protocols for CRSN, which provide distributed mechanisms to select the best wireless channel based on the application's QoS requirements. Simulations of low, medium and high noise scenarios have shown that the protocols improve the delivery rate by up to 69%, while keeping the delay and energy consumption unaltered.

The availability of inexpensive sensing devices capable of wireless communications enables the design of exciting and new distributed sensing applications. The integration of such applications with the Internet will contribute to materialize a vision of a future Web that we nowadays describe as the Web of Things (WoT). Many security-related aspects are still open on how to address security for communications on the WoT, and that must be targeted before such applications can realistically be considered ready for deployment. We discuss the experimental evaluation of mechanisms proposed to secure end-to-end web communications with IPv6-capable sensing applications and devices, as such mechanisms may provide an important contribution towards the WoT. Our experimental evaluation study considers the impact of security on real sensing devices and requirements from sensing applications, allowing the identification of the limitations of current sensing platforms in supporting the proposed mechanisms.

A new generation of Wireless Sensor Networks (WSNs) is expanding to performance critical environments. In these new scenarios, performance must be monitored and effectively measured, in order to guarantee that the network fulfills initial expectations. However, measuring performance in WSN has not the same requirements as measuring performance in traditional networks. While new specific WSN metrics have been proposed to cope with the new specific needs of these networks, a global performance framework for evaluating WSN does not yet exist. In this paper, WSN specific communication performance needs are analyzed and, to measure them, an initial approach to a global framework of WSN performance metrics is proposed. These metrics are intended to evaluate the real-time performance of a WSN, providing the necessary measurements to indicate if the QoS expectations are being met.

This paper presents Hermes, a new modular platform that can be used in a variety of scenarios, including the role of body area network coordinator, enabling temperature, movement and heartbeat sensing, local and remote communication, and indoor and outdoor localization. Hermes is based on two main modules that can operate as a single system in a coordinated fashion or individually, depending on the application scenario. One of the modules uses a Telos-inspired architecture, with new processing, communication, sensing, storage and energy subsystems, and executing TinyOS. The other module is designed around a PIC 24F MCU, also supporting communication, sensing, storage, and executing a custom operating system. Both modules have expansion capabilities. The main innovative aspects of the implemented platform are its modularity and its capability to provide device abstraction, which considerably eases application development. The paper addresses the platform motivation, requirements, hardware architecture, implementation details, and performance evaluation.

Recent critical application sectors of sensor networks like military, health care, and industry require the use of mobile sensor nodes, something that poses unique challenges in aspects like handoff delay, packet loss, and reliability. In this paper we propose a novel mobility model that handles those challenges effectively by providing on-time mobility detection and handoff triggering. In that way soft handoffs and controlled disconnections are assured. The proposed solution uses cross-layer information from the MAC and Network layers. Our solution was implemented and evaluated in an experimental testbed, in the context of the European FP7 GINSENG project.

Horses played a key role in the history of mankind, having been instrumental in the rise and fall of great empires. In modern days, they are still very important assets, from economical, operational or even sentimental points of view. Their importance is, thus, the main motivation for developing the iHorse equine monitoring system presented in this paper. iHorse is a deployed, fully-operational, commercially-available, wireless-sensor-network-based system and application, developed and tested over a three-year period. The aim of this paper is to provide information on the main features of the system, including its architecture, implementation details and performance data. The paper also discusses the lessons learnt throughout the system development and deployment phases.

Security plays a vital role in enabling future integration of Wireless Sensor Networks (WSNs) with the Internet. Although security in WSNs is a mature field of research and many proposals currently exist that can address particular security threats, how to guarantee end-to-end security at the network layer for IPv6-enabled WSNs remains an open research issue. Security should be addressed at the network layer, as sensor nodes will evolve to become IPv6 aware and new applications are adapted or designed that depend on IPv6 connectivity to sensor nodes. We propose and evaluate new compressed AH and ESP 6lowpan security headers that allow the protection of IPv6 communications on IEEE 802.15.4 WSNs. As the security headers are designed to work side-by-side with other extension headers of the 6lowpan adaptation layer, they allow also the establishment of end-to-end secure communications between Internet hosts and wireless sensor nodes. The usage of IPv6 on WSNs has the potential of enabling the integration of sensing applications with the Internet and, if end to end security is properly addressed, can represent a major contribution towards the "internet of things".

The goal of the GINSENG project is a performance-controlled sensor network suitable for applications with specific network performance requirements, such as plant automation and health monitoring. Any such sensor network must include a deterministic MAC protocol that can meet the application's targets in areas like delivery delays and reliability. We present a description and preliminary evaluation results of GinMAC, designed to be such a MAC protocol. Also included are details of a deployment to the GALP oil refinery in Sines, Portugal.

Wireless sensor networks are low cost networks constituted by modest devices with limited resources, whose main function is monitoring. Based on the low price of these devices, it will be cheap to deploy a large amount of nodes to monitor a large area. However, to provide an efficient ad hoc network using these limited devices, new and optimized algorithms should be proposed. Most of the current work about WSNs are based on simulation studies and do not take in consideration engineering processes. This paper presents a multi-sink node alternative to multi-hop solutions. The proposed solution also provides a new system for the discovery of devices and services over IPv6, allowing nodes to be automatically incorporated in the nearest WSN. This paper also presents a paradigm to efficiently provide mobility, granting a fast handover of nodes between different WSNs, without loosing the connection.

Ubiquitous broadband Internet access is an important requirement to satisfy user demands and support a new set of real time services and applications. IEEE 802.16, as a broadband wireless access technology, can significantly contribute in this area. It is able to provide high throughput over large distances, with intrinsic QoS support, in line-of-Sight and Non- line-of-Sight environments. This paper discusses a network architecture, based on IEEE 802.16 d compliant equipments, which is able to support real-time services with QoS integration through dynamic configuration of the 802.16 channel, using the NSIS framework. Tests have been carried out to evaluate the QoS performance of the defined architecture, demonstrating that it is perfectly capable of handling and supporting real time services, using dynamic configuration of the 802.16 channel, without introducing an excessive cost in the network.

Many issues still remain to be addressed in order to achieve acceptable security in wireless sensor networks (WSNs). This necessity, together with the adoption of IPv6 on WSNs being defined at the 6lowpan working group of the IETF, motivates our investigation on the feasibility of the application of IPSec on sensor nodes. IPSec is already part of IPv6, which makes it a natural candidate to be directly employed or adapted for WSNs. We discuss results obtained from practical experiments on the usage of cryptographic algorithms typically employed within IPSec in real sensor nodes. We analyze the security and performance tradeoffs involved when employing cryptography measures in WSNs, also in the context of their usage side-by-side with IPv6. The results obtained show that the adoption of a security architecture such as IPSec is viable, and also point towards the successful design and deployment of a security architecture for WSNs.

Mobility is one of the most important issues in next generation networks. As wireless sensor networks are becoming the next elements of the future Internet, it is crucial to study new models that also support mobility of these nodes. This paper presents and studies three paradigms to support mobility in sensor nodes. This study is supported by prototyping. As a result a set of proposals are proposed and discussed. The work presented in this paper proposes the use of IP in wireless sensor networks as a unifying protocol and it is in compliance with the IETF 6LowPAN group.