Use of Parallel Explanatory Models to Enhance Transparency of Neural Network Configurations for Cell Degradation Detection.

In a previous paper, we have shown that a recurrent neural network (RNN) can be used to detect cellular network radio signal degradations accurately. We unexpectedly found, though, that accuracy gains diminished as we added layers to the RNN. To investigate this, in this article, we build a parallel model to illuminate and understand the internal operation of neural networks (NNs), such as the RNN, which store their internal state to process sequential inputs. This model is widely applicable in that it can be used with any input domain where the inputs can be represented by a Gaussian mixture. By looking at RNN processing from a probability density function (pdf) perspective, we are able to show how each layer of the RNN transforms the input distributions to increase detection accuracy. At the same time we also discover a side effect acting to limit the improvement in accuracy. To demonstrate the fidelity of the model, we validate it against each stage of RNN processing and output predictions. As a result, we have been able to explain the reasons for RNN performance limits with useful insights for future designs for RNNs and similar types of NN.

RIS-assisted near-field localization using practical phase shift model.

Our research focuses on examining the problem of localizing user equipment (UE) in the uplink scenario using reconfigurable intelligent surfaces (RIS) based lens. We carry out a thorough analysis of the Fisher information matrix (FIM) and assess the influence of various RIS-based lens configurations using an actual RIS phase-dependent amplitude variations model. Furthermore, to reduce the complexity of the maximum likelihood (ML) estimator, a simple localization algorithm-based angular expansion is presented. Simulation results show superior localization performance when prior location information is available for directional and positional channel configurations. The position error bound (PEB) and the root mean square error (RMSE) are studied to evaluate the localization accuracy of the user utilizing the realistic RIS phase-dependent amplitude model in the near-field region. Furthermore, the achievable data rate is obtained in the same region using the realistic RIS phase-dependent amplitude model. It is noticed that adopting the actual RIS phase-dependent amplitude model under the near-field channel increases the localization error and degrades the data rate performance for amplitude value less than one so, the unity assumption of the RIS phase shift model used widely in the literature is inaccurate.

A Privacy and Energy-Aware Federated Framework for Human Activity Recognition.

Human activity recognition (HAR) using wearable sensors enables continuous monitoring for healthcare applications. However, the conventional centralised training of deep learning models on sensor data poses challenges related to privacy, communication costs, and on-device efficiency. This paper proposes a federated learning framework integrating spiking neural networks (SNNs) with long short-term memory (LSTM) networks for energy-efficient and privacy-preserving HAR. The hybrid spiking-LSTM (S-LSTM) model synergistically combines the event-driven efficiency of SNNs and the sequence modelling capability of LSTMs. The model is trained using surrogate gradient learning and backpropagation through time, enabling fully supervised end-to-end learning. Extensive evaluations of two public datasets demonstrate that the proposed approach outperforms LSTM, CNN, and S-CNN models in accuracy and energy efficiency. For instance, the proposed S-LSTM achieved an accuracy of 97.36% and 89.69% for indoor and outdoor scenarios, respectively. Furthermore, the results also showed a significant improvement in energy efficiency of 32.30%, compared to simple LSTM. Additionally, we highlight the significance of personalisation in HAR, where fine-tuning with local data enhances model accuracy by up to 9% for individual users.

Nature-inspired spider web shaped UHF RFID reader antenna for IoT and healthcare applications.

This paper proposes a nature-inspired spider web-shaped ultra-high frequency (UHF) radio frequency identification (RFID) reader antenna and battery-free sensor-based system for healthcare applications. This antenna design consists of eight concentric decagons of various sizes and five straight microstrip lines.These lines are connected to the ground using 50 [Formula: see text] resistors from both ends, except for one microstrip line that is reserved for connecting a feeding port. The reader antenna design features fairly strong and uniform electric and magnetic field characteristics. It also exhibits wideband characteristics, covering whole UHF RFID band (860-960 MHz) and providing a tag reading volume of 200 [Formula: see text] 200 [Formula: see text] 20 mm[Formula: see text]. Additionally, it has low gain characteristics, which are necessary for the majority of nearfield applications to prevent the misreading of other tags. Moreover, the current distribution in this design is symmetric throughout the structure, effectively resolving orientation sensitivity issues commonly encountered in low-cost linearly polarized tag antennas. The measurement results show that the reader antenna can read medicine pills tagged using low-cost passive/battery-free RFID tags, tagged expensive jewelry, intervenes solution, and blood bags positioned in various orientations. As a result, the proposed reader antenna-based system is a strong contender for near-field RFID, healthcare, and IoT applications.

Configuring reconfigurable intelligent surfaces using a practical codebook approach.

It is proven that the scattering, reflection, and refraction properties of electromagnetic signals can be adapted and managed by using reconfigurable intelligent surfaces (RISs). In this paper, we have investigated the performance of a single-input-single-output (SISO) wideband system in terms of achievable data rate by optimizing the phases of RIS elements and performing a fair power allocation for each subcarrier over the entire bandwidth. A new beamforming codebook is developed from which the maximizing signal-to-noise (SNR) configuration is selected. The channel state information (CSI) along with the selected maximizing SNR configuration is then used by the proposed power algorithm to obtain the optimal configuration of the RIS. To validate our proposed method, it is compared with state-of-the-art semidefinite relaxation (SDR) scheme in terms of performance, complexity and run-time consumption. Our method shows dramatically lower computational complexity than the SDR method and achieves an order of 2.5 increase in the achievable data rate with an optimized RIS compared with an un-configured surface.

Acoustic Lung Imaging Utilized in Continual Assessment of Patients with Obstructed Airway: A Systematic Review.

Smart respiratory therapy is enabled by continual assessment of lung functions. This systematic review provides an overview of the suitability of equipment-to-patient acoustic imaging in continual assessment of lung conditions. The literature search was conducted using Scopus, PubMed, ScienceDirect, Web of Science, SciELO Preprints, and Google Scholar. Fifteen studies remained for additional examination after the screening process. Two imaging modalities, lung ultrasound (LUS) and vibration imaging response (VRI), were identified. The most common outcome obtained from eleven studies was positive observations of changes to the geographical lung area, sound energy, or both, while positive observation of lung consolidation was reported in the remaining four studies. Two different modalities of lung assessment were used in eight studies, with one study comparing VRI against chest X-ray, one study comparing VRI with LUS, two studies comparing LUS to chest X-ray, and four studies comparing LUS in contrast to computed tomography. Our findings indicate that the acoustic imaging approach could assess and provide regional information on lung function. No technology has been shown to be better than another for measuring obstructed airways; hence, more research is required on acoustic imaging in detecting obstructed airways regionally in the application of enabling smart therapy.

FedBranched: Leveraging Federated Learning for Anomaly-Aware Load Forecasting in Energy Networks.

Increased demand for fast edge computation and privacy concerns have shifted researchers' focus towards a type of distributed learning known as federated learning (FL). Recently, much research has been carried out on FL; however, a major challenge is the need to tackle the high diversity in different clients. Our research shows that using highly diverse data sets in FL can lead to low accuracy of some local models, which can be categorised as anomalous behaviour. In this paper, we present FedBranched, a clustering-based framework that uses probabilistic methods to create branches of clients and assigns their respective global models. Branching is performed using hidden Markov model clustering (HMM), and a round of branching depends on the diversity of the data. Clustering is performed on Euclidean distances of mean absolute percentage errors (MAPE) obtained from each client at the end of pre-defined communication rounds. The proposed framework was implemented on substation-level energy data with nine clients for short-term load forecasting using an artificial neural network (ANN). FedBranched took two clustering rounds and resulted in two different branches having individual global models. The results show a substantial increase in the average MAPE of all clients; the biggest improvement of 11.36% was observed in one client.

Enhancing System Performance through Objective Feature Scoring of Multiple Persons' Breathing Using Non-Contact RF Approach.

Breathing monitoring is an efficient way of human health sensing and predicting numerous diseases. Various contact and non-contact-based methods are discussed in the literature for breathing monitoring. Radio frequency (RF)-based breathing monitoring has recently gained enormous popularity among non-contact methods. This method eliminates privacy concerns and the need for users to carry a device. In addition, such methods can reduce stress on healthcare facilities by providing intelligent digital health technologies. These intelligent digital technologies utilize a machine learning (ML)-based system for classifying breathing abnormalities. Despite advances in ML-based systems, the increasing dimensionality of data poses a significant challenge, as unrelated features can significantly impact the developed system's performance. Optimal feature scoring may appear to be a viable solution to this problem, as it has the potential to improve system performance significantly. Initially, in this study, software-defined radio (SDR) and RF sensing techniques were used to develop a breathing monitoring system. Minute variations in wireless channel state information (CSI) due to breathing movement were used to detect breathing abnormalities in breathing patterns. Furthermore, ML algorithms intelligently classified breathing abnormalities in single and multiple-person scenarios. The results were validated by referencing a wearable sensor. Finally, optimal feature scoring was used to improve the developed system's performance in terms of accuracy, training time, and prediction speed. The results showed that optimal feature scoring can help achieve maximum accuracy of up to 93.8% and 91.7% for single-person and multi-person scenarios, respectively.

Dental caries detection using a semi-supervised learning approach.

Early diagnosis of dental caries progression can prevent invasive treatment and enable preventive treatment. In this regard, dental radiography is a widely used tool to capture dental visuals that are used for the detection and diagnosis of caries. Different deep learning (DL) techniques have been used to automatically analyse dental images for caries detection. However, most of these techniques require large-scale annotated data to train DL models. On the other hand, in clinical settings, such medical images are scarcely available and annotations are costly and time-consuming. To this end, we present an efficient self-training-based method for caries detection and segmentation that leverages a small set of labelled images for training the teacher model and a large collection of unlabelled images for training the student model. We also propose to use centroid cropped images of the caries region and different augmentation techniques for the training of self-supervised models that provide computational and performance gains as compared to fully supervised learning and standard self-supervised learning methods. We present a fully labelled dental radiographic dataset of 141 images that are used for the evaluation of baseline and proposed models. Our proposed self-supervised learning strategy has provided performance improvement of approximately 6% and 3% in terms of average pixel accuracy and mean intersection over union, respectively as compared to standard self-supervised learning. Data and code will be made available to facilitate future research.

Contactless WiFi Sensing and Monitoring for Future Healthcare - Emerging Trends, Challenges, and Opportunities.

WiFi sensing has received recent and significant interest from academia, industry, healthcare professionals, and other caregivers (including family members) as a potential mechanism to monitor our aging population at a distance without deploying devices on users' bodies. In particular, these methods have the potential to detect critical events such as falls, sleep disturbances, wandering behavior, respiratory disorders, and abnormal cardiac activity experienced by vulnerable people. The interest in such WiFi-based sensing systems arises from practical advantages including its ease of operation indoors as well as ready compliance from monitored individuals. Unlike other sensing methods, such as wearables, camera-based imaging, and acoustic-based solutions, WiFi technology is easy to implement and unobtrusive. This paper reviews the current state-of-the-art research on collecting and analyzing channel state information extracted using ubiquitous WiFi signals, describing a range of healthcare applications and identifying a series of open research challenges, including untapped areas of research and related trends. This work aims to provide an overarching view in understanding the technology and discusses its use-cases from a perspective that considers hardware, advanced signal processing, and data acquisition.

Multi-Person Breathing Detection With Switching Antenna Array Based on WiFi Signal.

WiFi sensing, an emerging sensing technology, has been widely used in vital sign monitoring. However, most respiration monitoring studies have focused on single-person tasks. In this paper, we propose a multi-person breathing sensing system based on WiFi signals. Specifically, we use radio frequency (RF) switch to extend the antennas to form switching antenna array. A reference channel is introduced in the receiver, which is connected to the transmitter by cable and attenuator. The phase offset introduced by asynchronous transceiver devices can be eliminated by using the ratio of the channel frequency response (CFR) between the antenna array and the reference channel. In order to realize multi-person breathing perception, we use beamforming technology to conduct two-dimensional scanning of the whole scene. After eliminating static clutter, we combine frequency domain and angle of arrival (AOA) domain analysis to construct the AOA and frequency (AOA-FREQ) spectrogram. Finally, the respiratory frequency and position of each target are obtained by clustering. Experimental results show that the proposed system can not only estimate the direction and respiration rate of multi-person, but also monitor abnormal respiration in multi-person scenarios. The proposed low-cost, non-contact, rapid multi-person respiratory detection technology can meet the requirements of long-term home health monitoring.

Evaluation of deep learning models in contactless human motion detection system for next generation healthcare.

Recent decades have witnessed the growing importance of human motion detection systems based on artificial intelligence (AI). The growing interest in human motion detection systems is the advantages of automation in the monitoring of patients remotely and giving warnings to doctors promptly. Currently, wearable devices are frequently used for human motion detection systems. However, such devices have several limitations, such as the elderly not wearing devices due to lack of comfort or forgetfulness and/or battery limitations. To overcome the problems of wearable devices, we propose an AI-driven human motion detection system (deep learning-based system) using channel state information (CSI) extracted from Radio Frequency (RF) signals. The main contribution of this paper is to improve the performance of the deep learning models through techniques, including structure modification and dimension reduction of the original data. In this work, We firstly collected the CSI data with the center frequency 5.32 GHz and implemented the structure of the basic deep learning network in our previous work. After that, we changed the basic deep learning network by increasing the depth, increasing the width, adapting some advanced network structures, and reducing dimensions. After finishing those modifications, we observed the results and analyzed how to further improve the deep learning performance of this contactless AI-enabled human motion detection system. It can be found that reducing the dimension of the original data can work better than modifying the structure of the deep learning model.

Graph neural network-based cell switching for energy optimization in ultra-dense heterogeneous networks.

The development of ultra-dense heterogeneous networks (HetNets) will cause a significant rise in energy consumption with large-scale base station (BS) deployments, requiring cellular networks to be more energy efficient to reduce operational expense and promote sustainability. Cell switching is an effective method to achieve the energy efficiency goals, but traditional heuristic cell switching algorithms are computationally demanding with limited generalization abilities for ultra-dense HetNet applications, motivating the usage of machine learning techniques for adaptive cell switching. Graph neural networks (GNNs) are powerful deep learning models with strong generalization abilities but receive little attention for cell switching. This paper proposes a GNN-based cell switching solution (GBCSS) that has a smaller computational complexity than existing heuristic algorithms. The presented performance evaluation uses the Milan telecommunication dataset based on real-world call detail records, comparing GBCSS with a traditional exhaustive search (ES) algorithm, a state-of-the-art learning-based algorithm, and the baseline without cell switching. Results indicate that GBCSS achieves a 10.41% energy efficiency gain when compared with the baseline and achieves 75.76% of the optimal performance obtained with ES algorithm. The results also demonstrate GBCSS' significant scalability and generalization abilities to differing load conditions and the number of BSs, suggesting this approach is well-suited to ultra-dense HetNet deployment.

Towards the Digital Twin (DT) of Narrow-Band Internet of Things (NBIoT) Wireless Communication in Industrial Indoor Environment.

A study of the behavior of NB-IoT wireless communication in an industrial indoor environment was conducted in this paper. With Wireless Insite software, a scenario in the industrial sector was simulated and modeled. Our research examined how this scenario or environment affected the communication parameters of NB-IoT's physical layer. In this context, throughput levels among terminals as well as between terminals and transceiver towers, the power received at signal destination points, signal-to-noise ratios (SNRs) in the environment, and distances between terminals and transceivers are considered. These simulated results are also compared with the calculated or theoretical values of these parameters. The results show the effect of the industrial setting on wireless communication. The differences between the theoretical and simulated values are also established.

A data driven approach in less expensive robust transmitting coverage and power optimization.

This paper aims the development of a new reduced-cost algorithm for a multi-objective robust transmitter placement under uncertainty. Toward this end, we propose a new hybrid Kriging/Grey Wolf Optimizer (GWO) approach combined with robust design optimization to estimate the set of Pareto frontier by searching robustness as well as accuracy (lower objective function) in a design space. We consider minimization of the energy power consumption for transmitting as well as maximization of signal coverage in a multi-objective robust optimization model. The reliability of the model to control signal overlap for multiple transmitting antennas is also provided. To smooth computational cost, the proposed method instead of evaluating all receiver test points in each optimization iteration approximates signal coverages using Kriging interpolation to obtain optimal transmitter positions. The results demonstrate the utility and the efficiency of the proposed method in rendering the robust optimal design and analyzing the sensitivity of the transmitter placement problem under practically less-expensive computational efforts (350% and 320% less than computational time elapsed using standalone GWO and NSGAII respectively).

Mitigation pilot contamination based on matching technique for uplink cell-free massive MIMO systems.

In this paper, the cell-free massive multiple input multiple output (MIMO) network is affected by the pilot contamination phenomenon when a large number of users and a small number of available pilots exists, the quality of service (QoS) will deteriorate due to the low accuracy of the channel estimation because some of users will use the same pilot. Therefore, we address this problem by presenting two novel schemes of pilot assignment and pilot power control design based on the matching technique for the uplink of cell-free massive MIMO systems to maximize spectral efficiency. We first formulate an assignment optimization problem in order to find the best possible pilot sequence to be used by utilizing genetic algorithm (GA) and then propose a Hungarian matching algorithm to solve this formulated problem. Regarding the power control design, we formulate a minimum-weighted assignment problem to assign pilot power control coefficients to the estimated channel's minimum mean-squared error by considering the access point (AP) selection. Then, we also propose the Hungarian algorithm to solve this problem. Simulation results show that our proposed schemes outperform the state-of-the-art techniques concerning both the pilot assignment and the pilot power control design by achieving a 15% improvement in the spectral efficiency. Finally, the computational complexity analysis is provided for the proposed schemes compared with the state-of-the-art techniques.

Privacy-Preserving British Sign Language Recognition Using Deep Learning.

Sign language is a means of communication between the deaf community and normal hearing people who use hand gestures, facial expressions, and body language to communicate. It has the same level of complexity as spoken language, but it does not employ the same sentence structure as English. The motions in sign language comprise a range of distinct hand and finger articulations that are occasionally synchronized with the head, face, and body. Existing sign language recognition systems are mainly camera-based, which have fundamental limitations of poor lighting conditions, potential training challenges with longer video sequence data, and serious privacy concerns. This study presents a first of its kind, contact-less and privacy-preserving British sign language (BSL) Recognition system using Radar and deep learning algorithms. Six most common emotions are considered in this proof of concept study, namely confused, depressed, happy, hate, lonely, and sad. The collected data is represented in the form of spectrograms. Three state-of-the-art deep learning models, namely, InceptionV3, VGG19, and VGG16 models then extract spatiotemporal features from the spectrogram. Finally, BSL emotions are accurately identified by classifying the spectrograms into considered emotion signs. Comparative simulation results demonstrate that a maximum classifying accuracy of 93.33% is obtained on all classes using the VGG16 model.

Pushing the limits of remote RF sensing by reading lips under the face mask.

The problem of Lip-reading has become an important research challenge in recent years. The goal is to recognise speech from lip movements. Most of the Lip-reading technologies developed so far are camera-based, which require video recording of the target. However, these technologies have well-known limitations of occlusion and ambient lighting with serious privacy concerns. Furthermore, vision-based technologies are not useful for multi-modal hearing aids in the coronavirus (COVID-19) environment, where face masks have become a norm. This paper aims to solve the fundamental limitations of camera-based systems by proposing a radio frequency (RF) based Lip-reading framework, having an ability to read lips under face masks. The framework employs Wi-Fi and radar technologies as enablers of RF sensing based Lip-reading. A dataset comprising of vowels A, E, I, O, U and empty (static/closed lips) is collected using both technologies, with a face mask. The collected data is used to train machine learning (ML) and deep learning (DL) models. A high classification accuracy of 95% is achieved on the Wi-Fi data utilising neural network (NN) models. Moreover, similar accuracy is achieved by VGG16 deep learning model on the collected radar-based dataset.

High Gain Compact UWB Antenna for Ground Penetrating Radar Detection and Soil Inspection.

An ultrawide bandwidth (UWB) antenna for ground-penetrating radar (GPR) applications is designed to check soil moisture and provide good-quality images of metallic targets hidden in the soil. GPR is a promising technology for detecting and identifying buried objects, such as landmines, and investigating soil in terms of moisture content and contamination. A paddle-shaped microstrip antenna is created by cutting a rectangular patch at one of its diametrical edges fed by the coplanar waveguide technique. The antenna is loaded by stubs, shorting pins, and a split-ring resonator (SRR) metamaterial structure to increase the antenna's gain and enhance the bandwidth (BW) towards both the lower and higher end of the working BW. The antenna's performance in soil inspection is studied in terms of the operating frequency range, different types of soil, different distances (e.g., 50 cm) between the antenna arrays and soil, S-parameters, and gain. Following this, the antenna's ability to find a metallic target in the soil is tested, considering different array numbers, multi-targets, and locations. The antenna is designed on a thin layer of economic polytetrafluoroethylene (PTFE) substrate with dimensions 50 × 39 × 0.508 mm3 and works in the frequency range 1.9-9.2 GHz. In addition, two more resonances at 0.9 and 1.8 GHz are also achieved; hence, the antenna works for more than two application bands, such as the ISM- and L-bands. The measurement results validated excellent agreement with the simulated results. Furthermore, the recommended antenna offering a high gain of about 10.8 dBi and maximum efficiency above 97% proved able to discriminate between hidden objects and even recognize their shapes. Moreover, the reconstructed images show that the antenna can detect an object in the soil at any location.