A non-orthogonal multiple access (NOMA) assisted semi-grant-free (SGF) transmission framework is formulated to enable channel access for grant-free users (GFUs) by leveraging residual resources from grant-based users (GBUs). The goal of reducing age-of-information (AoI) while improving throughput is formulated by a joint optimization problem of transmission scheduling and beamforming design. In an effort to solve the pertinent problem, a hierarchical learning is proposed, which is based on deep reinforcement learning to obtain the channel state information of GBUs and the transmission status of GFUs. Specifically, a high-level policy is trained to perform beamforming from a global perspective, while a lower-level policy adapts to keep the AoI minimized. Numerical results demonstrate that the proposed approach outperforms existing adaptive and state-dependent baselines in AoI reduction, while achieving a throughput improvement of approximately 31.82%.

Obtaining data on active travel activities such as walking, jogging, and cycling are important for refining sustainable transportation systems (STS). In order to provide an accurate and privacy-preserving sensing solution, a deep learning (DL)-enhanced distributed acoustic sensing (DAS) system for recognizing active travel activities is proposed. By leveraging the ambient vibrations captured by DAS, this scheme infers motion patterns without relying on image-based or wearable devices, thereby addressing privacy concerns. We conduct real-world experiments in two geographically distinct locations and collect a comprehensive dataset to evaluate the performance of the proposed system. To address the generalization challenges posed by heterogeneous deployment environments, we propose two solutions based on network availability: 1) an Internet-of-Things (IoT) scheme based on federated learning (FL) is proposed, and it enables geographically different DAS nodes to be trained collaboratively to improve generality; 2) an off-line initialization approach enabled by meta-learning is proposed to develop high-generality initialization for DL models and to enable rapid model fine-tuning with limited data samples, facilitating generalization in newly established or isolated DAS nodes. Experimental results of the walking and cycling classification problem demonstrate the performance and generality of the DL-enhanced DAS system, paving the way for practical, large-scale DAS monitoring of active travel.

The Internet of Things (IoT) introduces diverse requirements and ubiquitous connections, necessitating efficient and affordable energy consumption as the ecosystem continues to grow. To address this challenge, we investigate a pure nonorthogonal multiple access (pure-NOMA) beamforming scheme to enhance system capacity by accommodating more IoT devices within the same spectrum. An energy efficiency (EE) maximization problem is formulated, jointly optimizing the beamforming matrix, power allocation, and device clustering. Due to the dynamic nature of the transmission channel and the coupling nonconvex mixed integer nonlinear programming (MINLP) problem, it is challenging to solve this problem by conventional mathematical methods. Additionally, the high dimensionality and coupling nonconvex MINLP problem pose significant challenges for traditional reinforcement learning (RL) methods. To overcome these issues, we propose a curiosity-driven approach that leverages intrinsic information from the base station (BS) to achieve energy efficient resource allocation. Simulation results demonstrate that pure-NOMA offers up to a 25% improvement in EE compared to hybrid-NOMA, while the curiosity-driven learning method outperforms baseline techniques, including deep RL (DRL), zero-forcing, and random methods, achieving a 14.78% reward gain over the DRL approach. The effectiveness of the proposed method is validated across various beam settings, device counts, quality-of-service requirements, and time consumption metrics, all while maintaining comparable computational complexity.

A distributed acoustic sensor (DAS)-enabled smart vehicle monitoring system is investigated in this article to detect the type and passenger occupancy of vehicles for intelligent transportation systems (ITSs). Accurate detection of the number of occupants and the vehicle type is critical for ITS to monitor the occupancy of vehicles, improve vehicle operation efficiency, and achieve intelligent traffic management. We have developed several deep learning (DL) algorithms for occupancy detection and vehicle type discrimination according to the signals provided by DAS. To be more specific, a novel type of basic neural network structure, namely sparse residual (SR) block is proposed, and several DL models are developed for DAS signals based on the basic SR block unit. The proposed DL approaches are tested using a unique dataset collected from a road experiment. The test results indicate that 1) the proposed SR network (SR-Net) and Alex SR (Alex-SR) network can achieve detection accuracy of over 90%; 2) the proposed models exhibit superior convergence, stability, and accuracy than the conventional residual network (ResNet); and 3) the proposed DL solutions have superiority in terms of complexity and model size compared to benchmarks.

Mobile-edge generation (MEG) is an emerging technology that allows the network to meet the challenging traffic load expectations posed by the rise of generative artificial intelligence (GAI). A novel MEG model is proposed for deploying GAI models on edge servers (ESs) and user equipment (UE) to jointly complete text-to-image generation tasks. In the generation task, the ES and UE will cooperatively generate the image according to the text prompt given by the user. To enable the MEG, a pretrained latent diffusion model (LDM) is invoked to generate the latent feature, and an edge-inferencing MEG protocol is employed for data transmission exchange between the ES and the UE. A compression coding technique is proposed for compressing the latent features to produce seeds. Based on the above seed-enabled MEG model, an image quality optimization problem with energy constraint is formulated. The transmitting power of the seed is dynamically optimized by a deep reinforcement learning (DRL) agent over the fading channel. The proposed MEG-enabled text-to-image generation system is evaluated in terms of image quality and transmission overhead. The numerical results indicate that, compared to the conventional centralized generation-and-downloading scheme, the symbol number of the transmission of MEG is materially reduced. In addition, the proposed compression coding approach can improve the quality of generated images under low signal-to-noise ratio (SNR) conditions, and the DRL-enabled dynamic power control further improves the image quality under the energy constraint compared to static transmit power control.

Mobile edge generation (MEG) is an emerging technology that allows the network to meet the challenging traffic load expectations posed by the rise of generative artificial intelligence (GAI). A novel MEG model is proposed for deploying GAI models on edge servers (ES) and user equipment (UE) to jointly complete test-to-image generation tasks. In the generaation task, the user uploads the text prompt and the ES and UE will cooperatively generate the image for the user. To enable the data transmission exchange between the ES and the UE, a seed based MEG protocol is employed, where a coded latent feature is created as a generation seed. A pre-trained latent diffusion model (LDM) is invoked to generate the latent feature, and a compression coding technique is proposed for compressing the latent features. The proposed MEG enabled text-to-image generation system is evaluated in terms of image quality and transmission overhead. The numerical results indicate that, compared to the conventional centralized generation-and-downloading scheme, the symbol number of the transmission of MEG is materially reduced. In addition, the proposed compression coding approach can improve the quality of generated images under low signal-to-noise ratio (SNR) conditions.

It is widely recognized that the health and well-being of physicians play a crucial role in ensuring the effectiveness and quality of healthcare delivery. Burnout and work-related stress have emerged as significant challenges affecting physicians' mental health worldwide. This global phenomenon has impacted the overall delivery of healthcare services and resulted in an increased risk of adverse effects for both physicians and patients. To overcome these challenges, addressing work-related stress issues is imperative for upholding healthcare standards and sustaining healthcare services. To this end, we present a new dataset that was meticulously crafted and collected through structured surveys and non-invasive finger-worn sensors at King Adbulla Medical City(KAMC). The key objective of this study was to identify the factors influencing physician well-being in high-stress clinical environments. In addition, we examined the relationships among various physiological, psychological, and work-related factors that might increase stress among medical practitioners working in different clinical settings. Our analysis provides promising insights into the intricate dynamics of stressful healthcare environments and the measured stress factors.

Active travel is an essential component in intelligent transportation systems. Cycling, as a form of active travel, shares the road space with motorised traffic which often affects the cyclists' safety and comfort and therefore peoples' propensity to uptake cycling instead of driving. This paper presents a unique dataset, collected by cyclists across London, that includes video footage, accelerometer, GPS, and gyroscope data. The dataset is then labelled by an independent group of London cyclists to rank the safety level of each frame and to identify objects in the cyclist's field of vision that might affect their experience. Furthermore, in this dataset, the quality of the road is measured by the international roughness index of the surface, which indicates the comfort of cycling on the road. The dataset11https://github.com/Chiayen0503/AllTheDocks_Dataset/tree/main will be made available for open access in the hope of motivating more research in this area to underpin the requirements for cyclists' safety and comfort and encourage more people to replace vehicle travel with cycling.

The detection of electricity theft is an important concern for both developed and developing economies to avoid financial loss and improve power distribution stability. Therefore, many automated solutions are proposed in terms of machine learning (ML) and deep learning (DL) for electricity theft detection (ETD) based on smart energy IoT data. However, these solutions do not consider the privacy of electricity consumers during ETD. Moreover, another overarching challenge in ETD is the predominant composition of honest users; therefore, models are often biased towards representing honest features. Therefore, a privacy-aware federated learning (FL) based solution is proposed in this paper for ETD. The proposed counter-bias FL-based solution employs a convolutional neural network (CNN), we call as FL-based CNN (FL-CNN), to detect the electricity theft consumers while using the Iot sourced electricity consumption (EC) data. The performance evaluation results state that the proposed FL-CNN outperforms state-of-the-art theft detection in terms of bandwidth requirement and reduces false theft attribution.

Recent sustainability drives place energy-consumption metrics in centre-stage for the design of future radio access networks (RAN). At the same time, optimising the trade-off between performance and system energy usage by machine-learning (ML) is an approach that requires large amounts of granular RAN data to train models, and to adapt in near realtime. In this paper, we present extensions to the system-level discrete-event AIMM (AI-enabled Massive MIMO) Simulator, generating realistic figures for throughput and energy efficiency (EE) towards digital twin network modelling. We further investigate the trade-off between maximising either EE or spectrum efficiency (SE). To this end, we have run extensive simulations of a typical macrocell network deployment under various transmit power-reduction scenarios with a range of difference of 43 dBm. Our results demonstrate that the EE and SE objectives often require different power settings in different scenarios. Importantly, low mean user CPU execution times of 2.17 ± 0.05 seconds (2 s.d.) demonstrate that the AIMM Simulator is a powerful tool for quick prototyping of scalable system models which can interface with ML frameworks, and thus support future research in energy-efficient next generation networks.

The Fifth-Generation of Mobile Communications (5G) is intended to meet users' growing needs for high-quality services at any time and from any location. The unique features of Low Earth Orbit (LEO) satellites in terms of higher coverage, reliability, and availability, can help expand the reach of 5G and beyond technologies to support those needs. However, because of their high speeds, a single LEO satellite is unable to provide continuous service to multiple User Equipments (UEs) spread over a large (potentially worldwide) area, resulting in the need for LEO satellite constellations with a high number of satellites and a consequent high amount of satellite handovers (HOs). Moreover, UEs can only acquire partial information about the satellite system and compete for the limited available communication resources of the satellites, requiring the implementation of a decentralized satellite HO strategy to avoid network congestion. In this paper, we propose a decentralized Load Balancing Satellite HO (LBSH) strategy based on multi-agent reinforcement Q-learning, implemented within the software Network Simulator 3 (NS-3). LBSH aims to reduce the total number of HOs and the blocking rate while balancing the load distribution among satellites. Our results show that the proposed LBSH method outperforms the state-of-the-art methods in terms of a 95% drop in the average number of HOs per user and an 84% reduction in blocking rate.

With the advent of modern smart grid networks, advanced metering infrastructure provides real-time information from smart meters (SM) and sensors to energy companies and consumers. The smart grid is indeed a paradigm that is enabled by the Internet of Things (IoT) and in which the SM acts as an IoT device that collects and transmits data over the Internet to enable intelligent applications. However, IoT data communicated over the smart grid could however be maliciously altered, resulting in energy theft due to unbilled energy consumption. Machine learning (ML) techniques for energy theft detection (ETD) based on IoT data are promising but are nonetheless constrained by the poor quality of data and particularly its imbalanced nature (which emerges from the dominant representation of honest users and poor representation of the rare theft cases). Leading ML-based ETD methods employ synthetic data generation to balance the training the dataset. However, these are trained to maximise average correct detection instead of ETD. In this work, we formulate an energy-aware evaluation framework that guides the model training to maximise ETD and minimise the revenue loss due to mis-classification. We propose a convolution neural network with positive bias (CNN-B) and another with focal loss CNN (CNN-FL) to mitigate the data imbalance impact. These outperform the state of the art and the CNN-B achieves the highest ETD and the minimum revenue loss with a loss reduction of 30.4% compared to the highest loss incurred by these methods.

Pervasive e-healthcare services have evolved rapidly in the recent years with the surge of the Internet of Medical Things (IoMT). Intelligent stress monitoring assistant is such an example that uses affective computing to detect stress levels based on biological signals. Existing works use different forms of machine learning and deep learning methods to successfully detect stress in pre-defined individuals. However, such models fail to recognise stress in unseen people, a problem that is harder to solve but of critical importance when three quarters of people in our society feel unable to cope due to stress. In this work, we propose a deep learning method that leverages self-attention to compute representations of the network's input and output layers and yields a generic stress detection model. We present a comparative analysis in which we examine the efficacy of Random Forest with handcrafted features, CNN-based deep learning, and transformer network with multi-modal data. Our results show that the proposed transformer solution outperforms the state of the art in all scenarios with accuracy of 96% and F1 score of 97%. More importantly, the model is validated using a leave-one-subject-out (LOSO) approach, hence is exportable and can successfully detect the stress condition of any unseen person.

As an advanced version of the Internet of Things (IoT), the Internet of Everything (IoE) network enables massive access for users and machine-type devices. A surge in the number of connected devices results in a stringent requirement for multiple access. To meet this challenge, a beamforming scheme in a downlink multiple-input and single-output (MISO) transmission is investigated, where an overloaded base station (BS) is committed to serving every user. Due to the deficiency of the number of antennas, the overload user equipment (OUE) needs to ride on an existing beam to connect to the BS. The optimization goal in this work is to maximize the energy efficiency (EE) of the system by joint optimizing for the selection of the shared beam and power allocation, subject to a constraint of the total power allocated to each beam. To tackle the formulated problem, a proximal policy optimization (PPO) based solution is proposed for jointly determining the beam choice and power allocation. Simulation results revealed that an average EE gain of 27.75 % can be achieved by the proposed solution compared with a spectral efficiency (SE) based method. Furthermore, the simulation with an increasing number of overload users highlights the performance of the joint optimization.

A novel approach is presented in this work for context-aware connectivity and processing optimization of Internet of Things (IoT) networks. Different from the state-of-the-art approaches, the proposed approach simultaneously selects the best connectivity and processing unit (e.g., device, fog, and cloud) along with the percentage of data to be offloaded by jointly optimizing energy consumption, response time, security, and monetary cost. The proposed scheme employs a reinforcement learning algorithm and manages to achieve significant gains compared to deterministic solutions. In particular, the requirements of IoT devices in terms of response time and security are taken as inputs along with the remaining battery level of the devices, and the developed algorithm returns an optimized policy. The results obtained show that only our method is able to meet the holistic multiobjective optimization criteria, albeit, the benchmark approaches may achieve better results on a particular metric at the cost of failing to reach the other targets. Thus, the proposed approach is a device-centric and context-aware solution that accounts for the monetary and battery constraints.

Subject identification has several applications. In transportation companies, knowing who is driving their vehicles might prevent theft or other ill-intended actions. On the other hand, privacy concerns, and the respective legislation, hinder the applicability of several traditional recognition techniques based on invasive technologies, such as video cameras. Moreover, in order to keep the driver's distractions to a minimum, this technologies must be non-disruptive, that is, they must be able to identify the subject seamlessly without them taking any action. In this context, we propose using deep learning applied to smart watch data for recognizing the person driving a vehicle based on a training set. Our proposal relies on the possibility of using transfer learning to avoid long training sessions for new drivers and to deliver a solution which can be deployed in practice. In this paper, we describe the convolutional neural network used in the solution and present results according to a real data-set collected by us, achieving accuracies ranging from 75 to 94%.

Wireless backhauling is recognised as a main contender for connecting small cells to the core network during the rise of 5G, especially in the absence of last mile fibre optic links. However, wireless backhauls present serious competition for the finite radio resources traditionally employed for radio access and are increasingly in demand due to exponential data growth. This work proposes a reinforcement learning approach that dynamically adjusts the sharing of radio resources between backhaul and radio access depending on the subtleties of the users' requirements and the network conditions. The metrics governing the reinforcement learning techniques are both usercentric and network-centric. They aim to maximise the network throughput while satisfying the differing requirements of users and corresponding applications. Our results indicate significant gains on the network performance (up to 23% throughput improvement) and the users' satisfaction (up to 15% improvement with respect to latency) as compared to static spectrum sharing methods for wireless backhaul.

The cell range extension (CRE) has been successfully implemented to bias the user to base station (BS) association policy in a way that achieves load balancing and increases the capacity of heterogeneous networks. The user-centric backhaul (UCB) scheme is a CRE evolution that is both backhaul-aware and user-context-aware -two constraints that are shaping the 5G network development. In this work, we formulate and solve the multi-objective optimisation problem of the UCB user-BS association. We derive analytical expressions of the ergodic throughput resulting from the UCB and, accordingly, identify the optimum association policy. The study demonstrates the gain margins that can be realised with pertinent user-cell association which is aware of the end-to-end network limitations and users requirements.

In the advent of ultra-dense networks with unprecedented complex and heterogeneous infrastructure, the role of automation in network optimization becomes vital for sustaining the target performance. In this work, we address the challenge of identifying and classifying sub-par performing nodes in near-real time through a machine-learning inspection of streaming performance indicators from multiple probe points. We present a novel K-means-based solution for classifying node performance over a sliding time segment and further categorizing the type of failure. The K-means solution first identifies the performance instances of interest. These are then inspected in a second clustering round for automated performance labeling. Next, the labeled data-set is employed to train a Support Vector Machine based classifier that is continuously classifying incoming performance instances from the network. The method is tested using a real network data set comprising call detail records. The results advocate the potential of our method for effectively and accurately identifying and classifying performance degradation in any node in the network.

With the advent of network densification and the development of other radio interface technologies, the major bottleneck of future cellular networks is shifting from the radio access network to the backhaul. The future networks are expected to handle a wide range of applications and users with different requirements. In order to tackle the problem of downlink user-cell association, and allocate users to the best cell, an intelligent solution based on reinforcement learning is proposed. A distributed solution based on Q-Learning is developed in order to determine the best cell range extension offsets (CREOs) for each small cell (SC) and the best weights of each user requirement to efficiently allocate users to the most appropriate SC, based on both backhaul constraints and user demands. By optimizing both CREOs and user weights, a user-specific allocation can be achieved, resulting in a better overall quality of service. The results show that the proposed algorithm outperforms current solutions by achieving better user satisfaction, mitigating the total number of users in outage, and minimizing user dissatisfaction when satisfaction is not possible.

Wireless technology is the strongest contender for catering for the 5G backhaul (BH) stipulated performance, where optical fiber is unavailable. In the presence of ultra-dense networks, such occurrences are exponentially increasing, and different wireless technologies are investigated for this application. We present the first BH-specific wireless link performance modeling that considers its inherent line-of-sight nature, together with an appropriate representation of the network topology using stochastic geometry. To this end, novel tractable models are obtained to capture the performance of wireless BH links. These are integrated into a multi-hop hybrid BH performance modeling framework and are applied in the analysis of a BH-aware user association optimization problem.

The Internet of Things (IoT) is spreading much faster than the speed at which the supporting technology is maturing. Today, there are tens of wireless technologies competing for IoT and a myriad of IoT devices with disparate capabilities and constraints. Moreover, each of many verticals employing IoT networks dictates distinctive and differential network qualities. In this work, we present a context‐aware framework that jointly optimises the connectivity and computational speed of the IoT network to deliver the qualities required by each vertical. Based on a smart port application, we identify energy efficiency, security, and response time as essential quality features and consider a wireless realisation of IoT connectivity using short range and long‐range technologies. We propose a reinforcement learning technique and demonstrate significant reduction in energy consumption while meeting the quality requirements of all related applications.

Heterogeneous networks are a key solution to serving the exponential surge in data volume and higher quality expectations. Nonetheless, such networks require the ubiquitous presence of fiber-to-the-cell to address the performance demands of 5G and fast-spreading small cells. To this end, innovative ways of optimizing the usage of realistic backhaul links are being investigated. In this work, we propose a fuzzy Q-learning-based user-centric backhaul-aware user cell association scheme. The proposed scheme aims at optimizing the user-cell association process in a context-aware and backhaul-aware manner. Complementing the scheme with fuzzy-logic requires 33.3% additional storage memory. On the other hand, it increases the computational efficiency by 60% and improves the users' performance by 12%.

The fifth generation of mobile networks (5G) is maturing fast and the target year 2020 is around the corner. However, the realistic backhaul network may not be ready for 5G arrival as it is likely to converge to 5G requirements at a slower pace than the radio counterpart. In this work, we develop a method that identifies pertinent backhaul upgrade stages that are ranked based on their associated cost. First, the User-centric- backhaul (UCB) scheme is employed to reveal the bottlenecks of the incumbent backhaul network, as perceived by users and holistic network. A multi- hop hybrid backhaul modelling framework is then employed to quantify possible rectifications that would deliver the highest improvement at the lowest cost. These are implemented and the results are verified following another usage of UCB. A case study is presented that demonstrates the strength of this method in enabling an effective and cost efficient evolution road map towards the 5G backhaul.

In the future, all devices that benefit from an Internet connection will be connected. Internet of Things technologies are key enablers of this vision by moving beyond basic connectivity machine-to-machine (M2M) communications brings to more intelligent interconnection of physical things on a massive scale. This anticipated growth is expected to challenge the planning and operation of cellular networks due to new diverse traffic models and high signaling loads. In this paper, we conduct a detailed experimental study using state-of-the-art drive testing equipment in order to measure, quantify, and analyze the signaling overhead of two classes of M2M services that resemble smart metering and vehicular applications. Two practical signaling reduction techniques are proposed and analyzed, with focus on aggregation as an efficient approach to overcome the resulting surge in signaling load. We complement the experimental results with an analytical evaluation to quantify the tradeoffs between M2M data transmission delay and the level of aggregation. Moreover, we present a novel case study to assess the potential negative impact of M2M signaling traffic on network planning and operation in 4G cellular networks.

The backhaul network is a critical challenge towards the success of 5G and corresponding difficulties are many-fold, such as network coverage expansion, very high bandwidth, ultra- low latency and energy consumption, at a minimum cost. No single backhaul solution can address all these requirements but, on the other hand, not all of the backhaul links require the same set of stringent requirements. To this end, we propose a novel scheme that capitalises on the diversity in both performance requirements and backhaul capabilities to maximise the system-centric as well as user-centric performance indicators. The user-centric backhaul provisioning scheme uses multiple attribute decision making (MADM) for the user-cell-backhaul association criteria in a way that intelligently associates users with available cells based on corresponding dynamic radio and backhaul conditions while abiding by users' requirements. Radio cells broadcast multiple bias factors, each reflecting a dynamic performance indicator of the end-to-end network performance such as capacity, latency, resilience, energy consumption, etc. A given user would employ these factors to derive a user-centric cell ranking that motivates it to select the cell with radio and backhaul capabilities that conform to the user requirements. Reinforcement learning is used by the radio cell to optimise the bias factors for each performance indicator in a way that maximises the system performance and users' end-to-end quality of experience (QoE). Preliminary results based on a case study show considerable improvement in users QoE when compared to state-of-the-art user-cell association schemes.