We propose iMoT, an innovative Transformer-based inertial odometry method that retrieves cross-modal information from motion and rotation modalities for accurate positional estimation. Unlike prior work, during the encoding of the motion context, we introduce Progressive Series Decoupler at the beginning of each encoder layer to stand out critical motion events inherent in acceleration and angular velocity signals. To better aggregate cross-modal interactions, we present Adaptive Positional Encoding, which dynamically modifies positional embeddings for temporal discrepancies between different modalities. During decoding, we introduce a small set of learnable query motion particles as priors to model motion uncertainties within velocity segments. Each query motion particle is intended to draw cross-modal features dedicated to a specific motion mode, all taken together allowing the model to refine its understanding of motion dynamics effectively. Lastly, we design a dynamic scoring mechanism to stabilize iMoT's optimization by considering all aligned motion particles at the final decoding step, ensuring robust and accurate velocity segment estimation. Extensive evaluations on various inertial datasets demonstrate that iMoT significantly outperforms state-of-the-art methods in delivering superior robustness and accuracy in trajectory reconstruction.

Despite remarkable progress in knowledge transfer across visual and textual domains, extending these achievements to indoor localization, particularly for learning transferable representations among Received Signal Strength (RSS) fingerprint datasets, remains a challenge. This is due to inherent discrepancies among these RSS datasets, largely including variations in building structure, the input number and disposition of WiFi anchors. Accordingly, specialized networks, which were deprived of the ability to discern transferable representations, readily incorporate environment-sensitive clues into the learning process, hence limiting their potential when applied to specific RSS datasets. In this work, we propose a plug-and-play (PnP) framework of knowledge transfer, facilitating the exploitation of transferable representations for specialized networks directly on target RSS datasets through two main phases. Initially, we design an Expert Training phase, which features multiple surrogate generative teachers, all serving as a global adapter that homogenizes the input disparities among independent source RSS datasets while preserving their unique characteristics. In a subsequent Expert Distilling phase, we continue introducing a triplet of underlying constraints that requires minimizing the differences in essential knowledge between the specialized network and surrogate teachers through refining its representation learning on the target dataset. This process implicitly fosters a representational alignment in such a way that is less sensitive to specific environmental dynamics. Extensive experiments conducted on three benchmark WiFi RSS fingerprint datasets underscore the effectiveness of the framework that significantly exerts the full potential of specialized networks in localization.

With the emergence of advanced use cases in In-dustrial, Healthcare and Smart City loT networks, applications and devices exhibit a wide range of QoS requirements. In the realm of loT, WiFi is a widely adopted technology that employs Enhanced Distributed Channel Access (EDCA) for QoS provisioning. However, EDCA's reliance on just four Access Categories (ACs) falls short in meeting the diverse range of QoS requirements of loT networks. SG-EmPOWER, a Software Defined Networking (SDN) framework, has introduced Network Slicing to WiFi networks, offering a robust solution to address the QoS diversity challenges in loT networks. SG-EmPOWER ensures slice priorities through quantum assignments to the slices, however this approach faces limitations in achieving effective slice differentiation, particularly in light of WiFi's Distributed Coordination Function (DCF). To overcome this limitation, we have introduced channel access parameter such as Contention Window (CW) into the slice configuration in addition to quantum value. This not only improves slice QoS differentiation but also brings robust and dynamic control over network QoS. Our proposed slice control mechanism has demonstrated substantial improvements, with an average increase of 47.66 % in per-slice throughput and an average reduction in jitter in high-priority slices of 23 %.

The article evaluates the effectiveness of coarsegrained channel quality prediction (CQP) for 5G-RedCap/5G NR-Light devices within industrial IoT (IIoT) networks. Finegrained predictions refine real-time communication, enhancing throughput and reducing resource utilization (RU), albeit with increased computation complexity. In contrast, coarse-grained CQP offers low computational overhead while optimizing longterm network characteristics, such as redundancy planning. Our study investigates the potential applications of coarsegrained CQP in real-time communication within the IIoT context, aiming to enhance the efficiency of simple devices (5G-RedCap) without adding the computational overhead. The varying traffic profiles and quality of service levels across diverse 5G use cases, including massive Machine Type Communication (mMTC), enhanced Mobile Broadband (eMBB), and Ultra-Reliable and Low Latency Communication (URLLC), present different challenges. For mMTC devices, coarse-grained CQP demonstrates comparable RU gains to fine-grained CQP (with up to a 50% reduction in RU), showcasing its effectiveness without added complexity. However, in the eMBB scenario, where throughput is paramount, it yields only marginal improvements. Similarly, RU gains for URLLC devices are negligible due to their stricter QoS requirements. The effectiveness of coarse-grained CQP is intricately linked to the variability in experienced channel quality across different scenarios within an indoor IIoT network. This research underscores the potential of AI applications for enhancing the performance of simple 5G-RedCap/5G NR-Light devices without compromising device complexity.

This paper explores the use of federated learning in a realistic household employing existing infrastructure to add new devices and locations by rotating the role of the transmitter among smart devices in a multi-person scenario. Current solutions employ channel state information-based sensing for health care monitoring in various ways to propagate knowledge efficiently; however, these solutions often consider (i) ideally placed devices in (ii) single-participant scenarios and (iii) do not consider the different roles of these devices in a network. Data is collected from four smart devices in a household, assuming three participants, one of which is monitored and the other two function as noise, are assigned to perform activities to replicate a realistic household scenario. Insights are provided on using federated learning in realistic at-home health care when adding a new activity location and client devices, both transmitter-only and full communication devices. Results indicate new devices and locations can quickly be adopted with less data by the federated model without intensive retraining, even in multi-person environments, when doing extensive pre-training.

Digital technologies in, on, and around bicycles and cyclists are gaining ground. Collectively called Smart Cycling Technologies (SCTs), it is important to evaluate their impact on subjective cycling experiences. Future evaluations can inform the design of SCTs, which in turn can help to realize the abundant benefits of cycling. Wearable body sensors and advanced driver assistance systems are increasingly studied in other domains, however evaluation methods integrating such sensors and systems in the field of cycling research were under-reviewed and under-conceptualized. This paper therefore presents a systematic literature review and conceptual framework to support the use of body sensors in evaluations of the impact of SCTs on perceptions, emotions, feelings, affect, and more, during outdoor bicycle rides. The literature review (n = 40) showed that there is scarce research on this specific use of body sensors. Moreover, existing research designs are typically not tailored to determine impact of SCTs on cycling experience at large scale. Most studies had small sample sizes and explored limited sensors in chest belts and wristbands for evaluating stress response. The evaluation framework helps to overcome these limitations, by synthesizing crucial factors and methods for future evaluations in four categories: (1) experiences with SCTs, (2) experience measurements, (3) causal analysis, (4) confounding variables. The framework also identifies which types of sensors fit well to which types of experiences and SCTs. The seven directions for future research include, for example, experiences of psychological flow, sensors in e-textiles, and cycling with biofeedback. Future interactions between cyclists and SCTs will likely resemble a collaboration between humans and artificial intelligence. Altogether, this paper helps to understand if future support systems for cyclists truly make cycling safer and more attractive.

This study presents a non-invasive method using thermal imaging to estimate heart and respiration rates in calves, avoiding the stress from wearables. Using Kernelised Correlation Filters (KCF) for movement tracking and advanced signal processing, we targeted one ROI for respiration and four for heart rate based on their thermal correlation. Achieving Mean Absolute Percentage Errors (MAPE) of 3.08 of thermal imaging in vital signs monitoring, offering a practical, less intrusive tool for Precision Livestock Farming (PLF), improving animal welfare and management.

In this paper, we present all-embracing Transformers (AaTs) that are capable of deftly manipulating attention mechanism for Received Signal Strength (RSS) fingerprints in order to invigorate localizing performance. Since most machine learning models applied to the RSS modality do not possess any attention mechanism, they can merely capture superficial representations. Moreover, compared to textual and visual modalities, the RSS modality is inherently notorious for its sensitivity to environmental dynamics. Such adversities inhibit their access to subtle but distinct representations that characterize the corresponding location, ultimately resulting in significant degradation in the testing phase. In contrast, a major appeal of AaTs is the ability to focus exclusively on relevant anchors in RSS sequences, allowing full rein to the exploitation of subtle and distinct representations for specific locations. This also facilitates disregarding redundant clues formed by noisy ambient conditions, thus enhancing accuracy in localization. Apart from that, explicitly resolving the representation collapse (i.e., none-informative or homogeneous features, and gradient vanishing) can further invigorate the self-attention process in transformer blocks, by which subtle but distinct representations to specific locations are radically captured with ease. For that purpose, we first enhance our proposed model with two sub-constraints, namely covariance and variance losses at the Anchor2Vec. The proposed constraints are automatically mediated with the primary task towards a novel multi-task learning manner. In an advanced manner, we present further the ultimate in design with a few simple tweaks carefully crafted for transformer encoder blocks. This effort aims to promote representation augmentation via stabilizing the inflow of gradients to these blocks. Thus, the problems of representation collapse in regular Transformers can be tackled. To evaluate our AaTs, we compare the models with the state-of-the-art (SoTA) methods on three benchmark indoor localization datasets. The experimental results confirm our hypothesis and show that our proposed models could deliver much higher and more stable accuracy.

Agitation is a commonly found behavioral condition in persons with advanced dementia. It requires continuous monitoring to gain insights into agitation levels to assist caregivers in delivering adequate care. The available monitoring techniques use cameras and wearables which are distressful and intrusive and are thus often rejected by older adults. To enable continuous monitoring in older adult care, unobtrusive Wi-Fi channel state information (CSI) can be leveraged to monitor physical activities related to agitation. However, to the best of our knowledge, there are no realistic CSI datasets available for facilitating the classification of physical activities demonstrated during agitation scenarios such as disturbed walking, repetitive sitting–getting up, tapping on a surface, hand wringing, rubbing on a surface, flipping objects, and kicking. Therefore, in this paper, we present a public dataset named Wi-Gitation. For Wi-Gitation, the Wi-Fi CSI data were collected with twenty-three healthy participants depicting the aforementioned agitation-related physical activities at two different locations in a one-bedroom apartment with multiple receivers placed at different distances (0.5–8 m) from the participants. The validation results on the Wi-Gitation dataset indicate higher accuracies (F1-Scores ≥0.95) when employing mixed-data analysis, where the training and testing data share the same distribution. Conversely, in scenarios where the training and testing data differ in distribution (i.e., leave-one-out), the accuracies experienced a notable decline (F1-Scores ≤0.21). This dataset can be used for fundamental research on CSI signals and in the evaluation of advanced algorithms developed for tackling domain invariance in CSI-based human activity recognition.

Detection of fatigue helps prevent injuries and optimize the performance of horses. Previous studies tried to determine fatigue using physiological parameters. However, measuring the physiological parameters, e.g., plasma lactate, is invasive and can be affected by different factors. In addition, the measurement cannot be done automatically and requires a veterinarian for sample collection. This study investigated the possibility of detecting fatigue non-invasively using a minimum number of body-mounted inertial sensors. Using the inertial sensors, sixty sport horses were measured during walk and trot before and after high and low-intensity exercises. Then, biomechanical features were extracted from the output signals. A number of features were assigned as important fatigue indicators using neighborhood component analysis. Based on the fatigue indicators, machine learning models were developed for classifying strides to non-fatigue and fatigue. As an outcome, this study confirmed that biomechanical features can indicate fatigue in horses, such as stance duration, swing duration, and limb range of motion. The fatigue classification model resulted in high accuracy during both walk and trot. In conclusion, fatigue can be detected during exercise by using the output of body-mounted inertial sensors.

WiFi, a widely used technology in IoT, provides QoS through IEEE 802.11e EDCA Access Categories (AC) which have fixed configuration with strict priorities. This makes WiFi QoS unsuitable to the rising QoS diversity, changing wireless conditions and network dynamics. QoS slicing, where network resources are divided into chinks for diverse QoS requirements, is a potent technology that can provide more flexible, adaptable and highly configurable QoS in $W$ iFi based IoT networks. However, resource management of QoS slices with diverse IoT applications, limited network capacity and varying channel conditions is a complex and challenging task. Traditional queuing theoretic and optimization models become intractable to solve such complex and dynamic problem. Therefore, we have developed a Deep Reinforcement Learning (DRL) based slice resource management scheme to meet IoT QoS requirements in a real world 5GEmpower controlled SDN network. Our proposed scheme outperforms Airtime Excess Round Robin (ATERR) scheme and no slicing-based scheme in terms of slice throughput (QoS) satisfaction. Moreover, our proposed scheme is tested in real world environment and can adapt to the changing slice requirements in an IoT network.

Decision support systems are becoming increasingly sophisticated (e.g., being machine learning-based), attempting to automate decisions as much as possible. However, it remains challenging to extract meaningful value from large quantities of data while also maintaining transparency in seeking justification for the choices made. Instead of creating methods for increasing the interpretability of black box models, one way forward is to design models that are inherently interpretable in the first place. Rule-based methods can automate decisions with great transparency and accuracy, helping to ensure compliance with regulations and adherence to organizational guidelines. In this paper, we propose an approach that uses a decision tree machine learning classification technique for extracting business rules from IoT-generated data to predict the asset status of Smart Returnable Transport Items (SRTIs). We report on an industrial case study that uses two years of historical data, obtained from an SRTI provider in the Netherlands, to predict the status of smart pallets. We compare the performance with the results obtained by using a support-vector machine (SVM) technique. Our experiments show that our solution is both accurate and flexible in terms of business rule elicitation. The obtained decision trees are human-interpretable, can easily be combined with other decision-making techniques, and provide a prediction accuracy marginally higher than an SVM technique.

“Smart cycling technologies” (SCTs) that support cyclists while cycling outdoors are on the rise, and it is important to evaluate their impact on cycling experiences. This research therefore aims to develop a framework that guides evaluations of the impact of SCTs on cycling experience. A key question is if and how wearable body sensors can deliver valuable input for such evaluations. Research methods are a systematic literature review, research through design, and empirical studies. Results so far are a conceptualization of the framework, a study about measuring and supporting shared flow in cyclists, and a study about predicting pleasantness ratings from data about brain oxygen consumption. Next steps are to develop a mixed method sensor system to measure experiences with an SCT, and to develop a data analysis method that moves beyond understanding correlations towards cause-effect relationships in experiences with SCTs.

In this paper, we propose Anchor-agnostic Transformers (AaTs) that can exploit the attention mechanism for Received Signal Strength (RSS) based fingerprinting localization. In real-world applications, the RSS modality is inherently well-known for its extreme sensitivity to dynamic environments. Since most machine learning algorithms applied to the RSS modality do not possess any attention mechanism, they can only capture superficial representations, yet subtle but distinct ones characterizing specific locations, thereby leading to significant degradation in the testing phase. In contrast, AaTs are enabled to focus exclusively on relevant anchors at every Received Signal Strength (RSS) sequence for these subtle but distinct representations. This also facilitates the model to neglect redundant clues formed by noisy ambient conditions, thus achieving better accuracy in fingerprinting localization. Moreover, explicitly resolving collapse problems at the feature level (i.e., none-informative or homogeneous features) can further invigorate the self-attention process, by which subtle but distinct representations to specific locations are radically captured with ease. To this end, we enhance our proposed model with two sub-constraints, namely covariance and variance losses that are mediated with the main task within the representation learning stage towards a novel multi-task learning manner. To evaluate our AaTs, we compare the models with the state-of-the-art (SoTA) methods on three benchmark indoor localization datasets. The experimental results confirm our hypothesis and show that our proposed models could provide much higher accuracy.

The proliferation of Internet of Things (IoT) applications is rapidly expanding, generating increased interest in the incorporation of blockchain technology within the IoT ecosystem. IoT applications enhance the efficiency of our daily lives, and when blockchain is integrated into the IoT ecosystem (commonly referred to as a blockchain-IoT system), it introduces crucial elements, like security, transparency, trust, and privacy, into IoT applications. Notably, potential domains where blockchain can empower IoT applications include smart logistics, smart health, and smart cities. However, a significant obstacle hindering the widespread adoption of blockchain-IoT systems in mainstream applications is the absence of a dedicated governance framework. In the absence of proper regulations and due to the inherently cryptic nature of blockchain technology, it can be exploited for nefarious purposes, such as ransomware, money laundering, fraud, and more. Furthermore, both blockchain and the IoT are relatively new technologies, and the absence of well-defined governance structures can erode confidence in their use. Consequently, to fully harness the potential of integrating blockchain-IoT systems and ensure responsible utilization, governance plays a pivotal role. The implementation of appropriate regulations and standardization is imperative to leverage the innovative features of blockchain-IoT systems and prevent misuse for malicious activities. This research focuses on elucidating the significance of blockchain within governance mechanisms, explores governance tailored to blockchain, and proposes a robust governance framework for the blockchain-enabled IoT ecosystem. Additionally, the practical application of our governance framework is showcased through a case study in the realm of smart logistics. We anticipate that our proposed governance framework will not only facilitate but also promote the integration of blockchain and the IoT in various application domains, fostering a more secure and trustworthy IoT landscape.

In this paper, we propose VarOLLA, a novel approach aimed at maximizing the throughput of 5G-RedCap devices in Industrial Internet of Things (IIoT) environments. VarOLLA addresses the sub-optimal spectral efficiency of Outer Loop Link Adaptation (OLLA) by introducing a ‘Throughput Factor’ that incentivizes adaptive decision-making based on throughput considerations. Through extensive simulations, we demonstrate significant improvements in throughput, with gains of up to 35% compared to traditional OLLA techniques, particularly in scenarios with high channel outdatedness. Moreover, VarOLLA effectively reduces consecutive transmission failures (rBLER) and achieves substantial reductions in control message overhead, up to 87.5%. Our findings highlight the strong potential of VarOLLA in IIoT networks and its significant contribution to the realization of high-performance applications during the Fourth Industrial Revolution (4IR) in the manufacturing sector.

The increasing demand for providing an animal-friendly environment that ensures animal well-being and economic efficiency in animal farming leads to more advanced and efficient animal farming methods. Precision livestock farming (PLF) can simultaneously remove time-consuming tasks and extra labor for a farmer and improve animal well-being and welfare. In this review, we describe the PLF concept with a focus on recently born pigs (piglets) which require extra care and attention due to their vulnerability and sensibility. To do so, we identify the top five mortality causes in newborn piglets: low vitality, hypothermia, starvation, crushing, and hypoxia. Furthermore, the correlation and interconnections among these mortality causes are explored. Then, the most relevant health indicators (HIs), which can be early indicators of an animal issues, are also introduced. Each mortality cause can be recognised through some symptoms, which can then be used as indicators. Considering these indicators, we can improve the animals' well-being and avoid unforeseen losses. Novel sensing and communication technologies in PLF can be utilized to monitor these HIs to give an early warning to the farmer in case something is wrong. Different novel sensing and communication technologies, whether tested on animals or not, are investigated in this review to give the researcher a better idea of available technologies that can be used in PLF implementation.

The diversity of QoS requirements in modern IoT networks is expanding thus requiring more flexible and autonomous solutions. Network Slicing is an efficient technology that can enable efficient and flexible management of QoS in IoT networks by dividing network resources into small chinks and assigning them to the slices. However, proposed solutions in literature require predefined slices which require prior knowledge of network traffic thus limiting application of these solutions. To address this limitation, we have developed a dynamic and autonomous network slicing solution in a WiFi based IoT network using 5GEmpower SDN controller that creates slices depending on user traffic flows and priorities. Moreover, it autonomously estimates slice throughput requirements to subsequently allocate network resources to meet their QoS requirements. Our system is able to differentiate between multiple QoS requirements and provides framework for flexible prioritization policies for better QoS in IoT networks.

New technologies are gaining ground in various disciplines, and road safety is not an exception. The objective of this paper is twofold: (1) to review the state-of-the-art technologies implemented in bicycles to improve cyclists’ safety, and (2) to propose a classification for the levels of smartness of emerging “smart bikes”. This paper defines six levels of smartness for bicycles based on their functionality and evaluates the Technology Readiness Levels of bicycle technologies. Furthermore, areas for future research were identified and discussed. To achieve these, we conducted a literature review which employed two academic databases –Scopus and Web of Science– and the Google Scholar search engine, following the framework of the systematic literature review methodology for the search and selection process. A total of 36 studies that met the inclusion criteria were investigated. The majority of these studies focus on warning systems aiming to forestall an imminent collision, mostly by using accelerometers/gyroscopes, LIDAR, sensors and networking communication. These systems, despite their preliminary state, demonstrate a positive effect on cyclists’ safety. The review concludes that there is a need for further deployment and testing of such systems with field trials to gain concrete evidence regarding their impact on cyclists’ safety. It also highlights that advanced technologies are scarcely implemented in bicycles and that most smart bicycle systems are based on smartphones. Thus, the question is: what lies in the future of smart bicycles from today’s perspective?

The Long Range (LoRa) network has been widely acknowledged for its efficiency and reliability in terrestrial sensing applications. However, building a robust LoRa network in the subsoil environment, which presents challenges for radio communication, remains challenging. This study evaluates the impact of antenna polarization and LoRa modulation parameters, such as bandwidth and spreading factor, on subsoil communication ranges. Based on the results of our experiments, we propose practical LoRa network configurations for the seamless transmission of subsoil sensory data to the surface.