Search

Conventional in-circuit impedance measurement setup of the inductive coupling approach requires two inductive probes. This article proposes a single-probe setup for the in-circuit impedance measurement, which not only eliminates the probe-to-probe coupling issue but also simplifies the setup with less instrument overhead. The basic principle behind the proposed setup will be elaborated, followed by experimental validation. To ensure the proposed setup can extract the in-circuit impedance with good accuracy, the required signal-to-noise ratio (SNR) of the setup is carefully studied. In addition, to improve the SNR and the ruggedness of the setup, signal amplification and protection devices are incorporated in the measurement. Finally, the proposed setup is deployed to demonstrate its ability to extract the in-circuit impedance of an energized induction motor.

The proactive management of a design can prevent design errors. Design changes can be mitigated by using Augmented Reality (AR) prototype digital models. However, the use of these models in construction is still in its infancy. Thus, this study assessed how AR could be used to overcome traditional pre-contract design communication related issues in construction. The study used a qualitative approach. The data required were collected through interviews conducted with experts in the fields of construction and information technology. Manual content analysis was used to analyse the collected empirical data. The study revealed that design communication issues could be related to architectural design, structural design, service design, and design collaboration. AR toolkit models, 360-degree model visualisation, walkthroughs, model registration, clash detection, and digital information overlays would minimise design communication issues, while AR features would enable design knowledge acquisition and negotiation. However, in construction, technical and social challenges are found to hinder AR implementation. Refereed/Peer-reviewed

While optical character recognition for Latin based scripts have seen near human quality performance, the accuracy for the rounded scripts of South Asia still lags behind. Work on Sinhala OCR has mainly reported on performance on constrained classes of font faces and so been inconclusive. This paper provides a comprehensive series of experiments using conventional machine learning as well as deep learning on texts and font faces of diverse types and in diverse resolutions, in order to present a realistic estimation of the complexity of recognizing the rounded script of Sinhala. While texts of both old and contemporary books can be recognized with over 87% accuracy, those in old newspapers are much harder to recognize owing to poor print quality and resolution.

Kidney development is key to the long-term health of the fetus. Renal volume and vascularity assessed by 3D ultrasound (3D-US) are known markers of wellbeing, however, a lack of real-time image segmentation solutions preclude these measures being used in a busy clinical environment. In this work, we aimed to automate kidney segmentation using fully convolutional neural networks (fCNNs). We used multi-parametric input fusion incorporating 3D B-Mode and power Doppler (PD) volumes, aiming to improve segmentation accuracy. Three different fusion strategies and their performance were assessed versus a single input (B-Mode) network. Early input-level fusion provided the best segmentation accuracy with an average Dice similarity coefficient (DSC) of 0.81 and Hausdorff distance (HD) of 8.96 mm, an improvement of 0.06 DSC and reduction of 1.43 mm HD compared to our baseline network. Compared to manual segmentation for all models, repeatability was assessed by intra-class correlation coefficients (ICC) indicating good to excellent reproducibility (ICC $>=$ 0.93). The framework was extended to support multiple graphics processing units (GPUs) to better handle volumetric data, dense fCNN models, batch normalization and complex fusion networks. This work and available source code provides a framework to increase the parameter space of encoder-decoder style fCNNs across multiple GPUs and shows that application of multi-parametric 3D-US in fCNN training improves segmentation accuracy.

A warm-up helps the individuals to prepare themselves for strenuous workouts. The warm-up is generally executed before participation in any physical activity or sports. The objective of the study was to investigate the effects of sequencing lower-body static and dynamic stretching combinations on agility, speed, and leg power of school-level female Netball players. Three different stretching protocols were performed: (a) Static Stretching combined with Dynamic Stretching (SS+DS), (b) Dynamic Stretching combined with Static Stretching (DS+SS), and (c) Dynamic Stretching combined with Dynamic Stretching (DS+DS). A control warm-up condition without stretching was implemented with a prior aerobic warm-up followed by dynamic activities. Dependent variables included a 30-m Sprint, Agility run, and jump tests to measure speed, agility, and leg power performance. The level of significance was set at 5% for statistical analysis. The method used for statistical analysis was two-way ANOVA. There was no significant impact of stretching protocols on agility (p=0.257), speed (p=0.106), and leg power (p=0.902) of school-level female netball athletes. The results of the analysis allow the authors to retain the hypothesis that a sequence of static and dynamic stretching combinations does not significantly affect the agility, speed, and leg power of school-level female netball athletes.

Although grant-based mechanisms have been a predominant approach for wireless access for years, the additional latency required for initial handshake message exchange and the extra control overhead for packet transmissions have stimulated the emergence of grant-free (GF) transmission. GF access provides a promising mechanism for carrying low and moderate traffic with small data and fits especially well for massive machine type communications (mMTC) applications. Despite a surge of interest in GF access, how to handle heterogeneous mMTC traffic based on GF mechanisms has not been investigated in depth. In this paper, we propose a priority enabled GF access scheme which performs dynamic slot allocation in each 5G new radio subframe to devices with different priority levels on a subframe-by-subframe basis. While high priority traffic has access privilege for slot occupancy, the remaining slots in the same subframe will be allocated to low priority traffic. To evaluate the performance of the proposed scheme, we develop a two-dimensional Markov chain model which integrates these two types of traffic via a pseudo-aggregated process. Furthermore, the model is validated through simulations and the performance of the scheme is evaluated both analytically and by simulations and compared with two other GF access schemes.

When in need for executing complex sets of interrelated calculations on High-Performance Computing (HPC) environments the obvious choice is to use scientific workflows. As workload management software do not support the execution of interrelated tasks, workflow management systems have been introduced to execute workflows on HPC environments. Recently, a new distributed architectural model that offers dynamic workflow execution capabilities to workflow management systems is introduced. It executes workflows on a per-task basis. While this approach facilitates dynamic workflows, it adds a considerable overhead to workflows substantially increasing their makespans. As most workflows are static, task-wise execution of workflows degrades the performance of most workflows. In this paper, we introduce a distributed workflow management system, SwarmForm that introduces task clustering to the new architectural model. SwarmForm is open source and offers better performance than existing distributed workflow management systems by clustering workflow tasks to reduce overheads while allowing the users to choose between task-wise and cluster-wise execution of workflows depending on the workflow nature. The paper proves that SwarmForm enables the use of all the features introduced with the new architectural model while providing better makespans for scientific workflows.

Analysis of metagenomic data is not only challenging because they are acquired from a sample in their natural habitats but also because of the high volume and high dimensionality. The fact that no prior lab based cultivation is carried out in metagenomics makes the inference on the presence of numerous microorganisms all the more challenging, accentuating the need for an informative visualization of this data. In a successful visualization, the congruent reads of the sequences should appear in clusters depending on the diversity and taxonomy of the microorganisms in the sequenced sample. The metagenomic data represented by their oligonucleotide frequency vectors is inherently high dimensional and therefore impossible to visualize as is. This raises the need for a dimensionality reduction technique to convert these higher dimensional sequence data into lower dimensional data for visualization purposes. In this process, preservation of the genomic characteristics must be given highest priority. Currently, for dimensionality reduction purposes in metagenomics, Principal Component Analysis (PCA) which is a linear technique and t-distributed Stochastic Neighbor Embedding (t-SNE), a non-linear technique, are widely used. Albeit their wide use, these techniques are not exceptionally suited to the domain of metagenomics with certain shortcomings and weaknesses. Our research explores the possibility of using autoencoders, a deep learning technique, that has the potential to overcome the prevailing impediments of the existing dimensionality reduction techniques eventually leading to richer visualizations.

A vialess passive metamaterial (MTM)-based Quadrature Phase Shift Keying (QPSK) modulated chipless Radio Frequency Identification (RFID) tag is integrated with a partial discharge (PD) sensor with the objective of PD detection and identification. Individual tags with the sensor can be mounted on specific electrical assets to be monitored. When interrogated by an external reader at 2.4 GHz, the tags respond with signals that contain the ID information and the state of the PD sensor. Any PD events detected can be traced back to the respective electrical assets that generate the PD emissions after a time-domain (TD) analysis, which allows prompt corrective action before the further degradation and premature failure of assets.

The in-circuit impedance serves as a key parameter for evaluating the operating status and health condition of many critical electrical assets. For its noncontact nature and simplicity in on-site implementation, the inductive coupling approach is an attractive method to extract the in-circuit impedance of an electrical system. In some practical situations where there are space constraints, the inductive probes used in this approach are very close to each other and the probe-to-probe coupling between inductive probes cannot be ignored. This article investigates possible errors due to such coupling and proposes an open-short-load calibration technique and a proper measurement process to minimize the error in the measured in-circuit impedance.

Structural Plasticity (SP) in the brain is a process that allows structural neuronal changes, in response to learning. Spiking Neural Networks (SNN) are an emerging form of artificial neural networks that use brain-inspired techniques to learn. However, the application of SP in SNNs, its impact on overall learning, and network behaviour is rarely explored. In the present study, we use an SNN with a single hidden layer, to apply SP in classifying Electroencephalography (EEG) signals of two publicly available datasets. We considered classification accuracy as the learning capability and applied metaheuristics to derive the optimised number of neurons for the hidden layer along with other hyperparameters of the network. The optimised structure was then compared with overgrown and undergrown structures to compare the accuracy, stability, and behaviour of the network properties. Networks with SP yielded ~94% and ~92% accuracies in classifying wrist positions and mental states(stressed vs relaxed) respectively. The same SNN developed for mental stress classification produced ~77% and ~73% accuracies in classifying arousal and valence states. Moreover, the networks with SP demonstrated superior performance stability during iterative random initiations. Interestingly, these networks had a smaller number of inactive neurons and a preference for lowered neuron firing thresholds. This research highlights the importance of systematically selecting the hidden layer neurons over arbitrary settings, particularly for SNNs using Spike Time Dependent Plasticity learning and provides potential findings that may lead to the development of SP learning algorithms for SNNs.

Support Vector Machines (SVMs) are vulnerable to targeted training data manipulations such as poisoning attacks and label flips. By carefully manipulating a subset of training samples, the attacker forces the learner to compute an incorrect decision boundary, thereby causing misclassifications. Considering the increased importance of SVMs in engineering and life-critical applications, we develop a novel defense algorithm that improves resistance against such attacks. Local Intrinsic Dimensionality (LID) is a promising metric that characterizes the outlierness of data samples. In this work, we introduce a new approximation of LID called K-LID that uses kernel distance in the LID calculation, which allows LID to be calculated in high dimensional transformed spaces. We introduce a weighted SVM against such attacks using K-LID as a distinguishing characteristic that de-emphasizes the effect of suspicious data samples on the SVM decision boundary. Each sample is weighted on how likely its K-LID value is from the benign K-LID distribution rather than the attacked K-LID distribution. Experiments with benchmark data sets show that the proposed defense reduces classification error rates substantially (10% on average).

In this paper, we review Augmented Reality Training systems and how they can be combined with Intelligent Tutoring Systems to improve pedagogical outcomes on spatial learning tasks. Intelligent Tutoring Systems (ITSs) are computer-based systems that provide personalised feedback, learner support and task selection by modelling learner knowledge. They have been shown to significantly improve learning, however currently provide little support for psychomotor tasks that involve spatial learning. For such tasks, ITSs may not be able to present instruction in an intuitive manner for mental model reconstruction. Augmented Reality (AR) overlays 3-Dimensional virtual objects on the real world, and has been shown to be useful for training in physical manipulation tasks. However, most current AR training systems provide only step by step instructions and do not adapt training content to user needs. In this paper we show how combining ITS with AR systems has the potential to overcome the limitations of both approaches and produce a significantly improved learning experience. We provide examples of previous work in this area, and provide guidelines for future AR ITS systems. Refereed/Peer-reviewed

Purpose Woven fabrics have been popularised in use owing to their superior properties and functionality. Today, weavers strive to add value to their product to be competitive and to secure profit in performance fabrics such as technical fabrics, smart fabrics and sportswear fabrics. Over the years, fabrics with special properties such as moisture management have gained higher demand. In this context, multi-layer fabrics provide a reasonable solution to the demand. Design/methodology/approach An attempt was made to develop two-layer fabrics with different compositions and properties. A two-layer woven fabric was produced using handloom weaving, with a hydrophobic inner layer and hydrophilic outer layer, the two layers being attached together using different stitching methods. Different fabric structures and yarn counts were used to achieve the objectives. Findings Experiments carried out verified the suitability of the developed fabric for effective moisture management. It was found that a fabric with a 100% cotton outer layer and 100% polyester inner layer, both layers of 2 × 2 matt weave, showed the best properties. Practical implications In the present COVID-19 pandemic situation, the use of masks in public has become mandatory in many countries. This research will help handloom manufacturers meet the need using simple methods. Originality/value This research uses handloom fabric. As such it provides an opportunity for small and medium enterprises to use available low-cost technology to develop fabric with superior properties.

Biometric recognition for human identification plays a key role in the rapid development of computer vision and pattern recognition research areas. The biometrics, refers to the automatic identification of a person based behavioral characteristics, physiological properties or traits. Signature recognition is one such human identification method and can be performed either in offline or online mode. This paper proposed an offline handwritten signature recognition which is based on image processing technique, scale-invariant feature transform (SIFT) and speeded-up robust features (SURF) features and support vector machines (SVMs). The handwritten signature images were then recognized through the proposed method that involves identification of regions of interest and representation of those regions as SIFT or SURF, construction of codebooks and the multi-class classification of the feature histograms using support vector machines (SVMs). Experiments have been carried out with our dataset of 1600 samples and a recognition rate in excess of 95% was obtained over the ten-fold cross validations.

Although modern society is at the apex of science and technology, the shadows of traditional customs and beliefs in various fields can be seen. When the field of house construction is considered, traditional beliefs had been playing a considerable role which can be seen in Asian countries rather than European countries. Since hardly any engineering-based studies have been done to explore the significance of traditional beliefs in house construction, this paper tends to fill that gap and widen the thinking pattern of the society. The study reveals that 68% of stakeholders of house construction keep faith in traditional beliefs, while 32% do not. Among them, 81% accept not having three or more aligned openings in dwellings, 80.5% believe that it is not favourable to place ridge plate, beam or any load on top of the openings, 75.2% say that more west-facing windows are not suitable for houses while 65.7% say that west-facing verandas and balconies are not suitable for houses and 73.3% believe that using cross-walls for brick walls must be neglected. With respect to the engineering experiments done, not using cross-walls in brick wall construction was proven to be false with respect to the structural testing. Not having ridge plate, beam or any load on top of the openings was proven to be negligible in the presence of reinforced concrete (RC) lintel. But not having three or more aligned openings and not using west-facing verandas, balconies and more windows were accepted based on computational fluid dynamic (CFD) analysis and general explanations based on thermal comfort and structural behaviour of door and window frames. Hence, hasty neglection of traditional beliefs in house constructions by labelling them as superstitions should be changed. The acceptance or rejection of aforesaid beliefs has to be done with respect to the rational content of them, but not being blind slaves of them wasting both money and time.

Depression is a common mental health disorder that affects an individual’s moods, thought processes and behaviours negatively, and disrupts one’s ability to function optimally. In most cases, people with depression try to hide their symptoms and refrain from obtaining professional help due to the stigma related to mental health. The digital footprint we all leave behind, particularly in online support forums, provides a window for clinicians to observe and assess such behaviour in order to make potential mental health diagnoses. Natural language processing (NLP) and Machine learning (ML) techniques are able to bridge the existing gaps in converting language to a machine-understandable format in order to facilitate this. Our objective is to undertake a systematic review of the literature on NLP and ML approaches used for depression identification on Online Support Forums (OSF). A systematic search was performed to identify articles that examined ML and NLP techniques to identify depression disorder from OSF. Articles were selected according to the PRISMA workflow. For the purpose of the review, 29 articles were selected and analysed. From this systematic review, we further analyse which combination of features extracted from NLP and ML techniques are effective and scalable for state-of-the-art Depression Identification. We conclude by addressing some open issues that currently limit real-world implementation of such systems and point to future directions to this end.

Experiential learning is the process of learning through experience or more specifically “learning through reflection on doing”. These experiences can be simulated in Extended Reality (XR) systems– interactive environments generated by computer technology combining real and virtual worlds. Such systems have shown to be excellent tools for providing guidance to teach different skills and activities. In this research, we introduce a comprehensive design space grounded in (i) XR technologies, and (ii) theoretical underpinnings of experiential learning and guidance. We use this design space to describe key design decisions in prior work and elucidate existing gaps. Employing the design space, we intend to build an adaptive, personalised instructional guidance system in XR, which offers various levels of guidance that adapt to users’ learning progress. We utilise the Kolb’s experiential learning cycle considering all four stages and transitions within it. With XR technologies, we offer the environment that can provide guidance to support learners to move spontaneously from one stage to another in the experiential learning cycle and further progress on the learning spiral. It is the latter that we explore in this research by adopting potential learning scenarios such as language learning and spatial navigation. We attempt to investigate how different factors of instructional guidance, such as the amount (minimal guidance, full guidance, and adaptive guidance) and the type (from direct instructions to associative models based examples), would affect the experiential learning process and learning outcomes of different learning scenarios in XR environments.

External fixation is a commonly used method in stabilizing fracture sites. The performance of the fixator depends on how it affects the mechanical properties of the fracture site and is governed by parameters like the fixator type and fixator configuration. Identifying ideal configurations prior to surgery will help surgeons in planning the procedure, limiting the possibility of complications such as non-union. In this study, a framework has been proposed as a surgical pre-planning tool, to assist surgeons compare mechanical properties of a fracture site under different fixator configurations, and thereby identify the optimum solution. A computational tool was identified as the best method for this purpose. Cost and time of computation were given special consideration to reduce complexity in clinical settings. A pilot study was conducted on a section of the proposed framework, where the aim was to understand the feasibility of implementation. In the pilot study, a unilateral uni-planar fixator on a simple diaphyseal transverse fracture was analyzed. During the pilot study the selected fixator was tested and a few models were developed to assess system stability. The models were then compared to identify the optimum model that could be used with the proposed framework. The proposed framework provided a suitable solution for the use case and out of the models developed the simplified finite element model was identified as the best option for the use case.

With recent technology advancements conversational interfaces such as Google Home, Siri, and Amazon Echo have gained much popularity among all kinds of ages. Conversational interfaces include chatbots where it's available on-screen and as a voice assistant. With the increasing number of long-term patients such as elderly and disabled people, it is essential to use the technology to fulfill their needs and make them more independent. Therefore, this study is initiated with the motivation to investigate how to facilitate the communication and psychological requirements of long-term patients using conversational interfaces. For that, a comprehensive Interview study based on long term patients in Sri Lanka was conducted to identify the requirements of the elderly long-term patients and done a thematic analysis on the use of voice integrated personal assistants. Proof of concept designed from the identified needs in a user-centric manner. Further, the Developed solution is quantitatively evaluated by elderly long-term patients mainly under five criteria. Based on the results, appropriateness and acceptance of conversational interfaces in the research domain are presented.

Near-Infrared spectroscopy is used for better vein visualization to make the venipuncture process more efficient. There exist a few models that use the said mechanism. However, they are costly, have accuracy and availability issues, and are limited only to certain types of skin tones. Our objectives were to develop a low-cost mechanism of obtaining near-infrared spectroscopy by using the image-guided technique, low-cost hardware, optimized algorithms, and evaluate its efficiency and usefulness by a clinical trial.We have tested the prototype using different combinations of light sources with different intensities and have analyzed the results. To quantitatively analyze, we have compared the number of visible veins under high intensity and low intensity. The number of visible veins is either same or up to 5 veins higher when 18W is used compared to when an intensity of 60W is used. We have also observed that the darker skin tones that have zero visible veins at normal sight result up to 2-3 veins when the prototype is used. The number of veins increased from 1 to 5, when the device was used, on fairer skin as well. We plan to conduct a clinical trial and test the device on human subjects and get feedback from the end-users, and improve the prototype accordingly.

The role of modelling in predicting the spread of an epidemic is important for health planning and policies. This study aimed to apply a dynamic Susceptible-Exposed-Infected-Recovered-Deaths (SEIRD) model and simulated it under a range of epidemic conditions using python programme language. The predictions were based on different scenarios from without any preventive measures to several different preventive measures under R0 of 4. The model shows that more weight to personal protection can halt the spread of transmission followed by the closure of public places and interprovincial movement restriction. Results after simulating various scenarios indicate that disregarding personal protective measures can have devastating effects on the Sri Lankan population. The importance of strict adherence, maintain and monitoring of self-preventive measures lead to minimizing the death toll from COVID-19. What we already knowO_LICompartmental models can be used to project scenarios with various disease control measures individually or as a useful combination for evidence-based policy formulation. C_LIO_LIEpidemiologists have been using mass action, compartmental models, over a hundred years which are famous for simplicity in both analysis and outcome assessment. C_LIO_LIMathematical modelling plays a vital role in the highest level of policymaking in the fields of health economics, emergency planning, monitoring of surveillance data and, risk assessment and control. C_LI What this article addsO_LIThis paper aimed to describe a dynamic Susceptible-Exposed-Infected-Recovered-Deaths (SEIRD) model and simulate it under a range of epidemiological conditions to give an insight into COVID-19 spread in Sri Lanka. C_LIO_LIThe SEIRD model produces a time frame for preparedness and resource allocation of a country without exceeding the surge capacity which may lead to a disastrous situation which allows authorities to plan potential mortuary capacity and understand the burden on crematoria and burial services. C_LIO_LINew policy discussions need to occur whenever the best available options such as different preventive strategies and knowledge about the epidemiology changes. C_LIO_LIThe proposed model can serve as a tool for health authorities for planning and policymaking to control the pandemic by implementing appropriate policy decisions on time to prevent a disastrous situation. C_LI

Nonlinear regression, although widely used in engineering, financial and security applications for automated decision making, is known to be vulnerable to training data poisoning. Targeted poisoning attacks may cause learning algorithms to fit decision functions with poor predictive performance. This paper presents a new analysis of local intrinsic dimensionality (LID) of nonlinear regression under such poisoning attacks within a Stackelberg game, leading to a practical defense. After adapting a gradient-based attack on linear regression that significantly impairs prediction capabilities to nonlinear settings, we consider a multi-step unsupervised black-box defense. The first step identifies samples that have the greatest influence on the learner's validation error; we then use the theory of local intrinsic dimensionality, which reveals the degree of being an outlier of data samples, to iteratively identify poisoned samples via a generative probabilistic model, and suppress their influence on the prediction function. Empirical validation demonstrates superior performance compared to a range of recent defenses.

Deep brain stimulation (DBS) is a well-established treatment option for a variety of neurological disorders, including Parkinson’s disease and essential tremor. The symptoms of these disorders are known to be associated with pathological synchronous neural activity in the basal ganglia and thalamus. It is hypothesised that DBS acts to desynchronise this activity, leading to an overall reduction in symptoms. Electrodes with multiple independently controllable contacts are a recent development in DBS technology which have the potential to target one or more pathological regions with greater precision, reducing side effects and potentially increasing both the efficacy and efficiency of the treatment. The increased complexity of these systems, however, motivates the need to understand the effects of DBS when applied to multiple regions or neural populations within the brain. On the basis of a theoretical model, our paper addresses the question of how to best apply DBS to multiple neural populations to maximally desynchronise brain activity. Central to this are analytical expressions, which we derive, that predict how the symptom severity should change when stimulation is applied. Using these expressions, we construct a closed-loop DBS strategy describing how stimulation should be delivered to individual contacts using the phases and amplitudes of feedback signals. We simulate our method and compare it against two others found in the literature: coordinated reset and phase-locked stimulation. We also investigate the conditions for which our strategy is expected to yield the most benefit., Author summary In this paper we use computer models of brain tissue to derive an optimal control algorithm for a recently developed new generation of deep brain stimulation (DBS) devices. DBS is a treatment for a variety of neurological disorders including Parkinson’s disease, essential tremor, depression and pain. There is a growing amount of evidence to suggest that delivering stimulation according to feedback from patients, or closed-loop, has the potential to improve the efficacy, efficiency and side effects of the treatment. An important recent development in DBS technology are electrodes with multiple independently controllable contacts and this paper is a theoretical study into the effects of using this new technology. On the basis of a theoretical model, we devise a closed-loop strategy and address the question of how to best apply DBS across multiple contacts to maximally desynchronise neural populations. We demonstrate using numerical simulation that, for the systems we consider, our methods are more effective than two well-known alternatives, namely phase-locked stimulation and coordinated reset. We also predict that the benefits of using multiple contacts should depend strongly on the intrinsic neuronal response. The insights from this work should lead to a better understanding of how to implement and optimise closed-loop multi-contact DBS systems which in turn should lead to more effective and efficient DBS treatments.

In recent years, the Medical Internet of Things or well known as MIoT has attracted so much attentional from researchers, academia for its potential to alleviate the burden on global healthcare caused by the shortage of medical staff, chronic diseases, and deadly contagious diseases which quickly turned into global pandemics like recent COVID-19. It is believed that soon, MIoT will become the next healthcare enabler contributing to all major aspects of healthcare. Day by day the number of MIoT devices in healthcare growing with a rapid phase and this also will lead to a massive increase of data that the devices are generating. This large volume of data and the devices themselves are posing huge questions about the security and the privacy of the entire MIoT ecosystem. As of now, the cyber-attacks that target healthcare are also increasing with a rapid phase, where comprehensive cyber-attacks that exploit the privacy of MIoT data and devices may even lead to jeopardizing the security of the entire ecosystem and may cause loss of patient lives if no countermeasures are taken. Hence the privacy of MIoT in healthcare is becoming a major threat where no adequate research has not been done so far. With this regard, in this study, we aim to review the privacy aspect of this MIoT and privacy-preserving solutions for MIoT along with the challenges pertaining to the discipline with the anticipated future directions.

In pandemic times, public facilities, such as gymnasiums and other sports facilities are closed. Hence, keeping oneself physically fit becomes particularly difficult as well as getting help from a fitness trainer or instructor is almost impossible. We propose an alternative solution, GymSoles++, a virtual trainer supporting the user with their exercise. This allows users to maintain the correct body posture to avoid injuries when performing exercises such as squats. We combine Google Glass with an unobtrusive sensing approach, an insole-based wearable system that provides feedback on the user’s centre of pressure (CoP) via vibrotactile and visual aids. Previous research has shown that solely visualizing the CoP can significantly improve body posture and thus effectively assist users when performing squats and dead-lifts. In this research, we explored different feedback modalities and conclude that a vibrotactile insole is a practical and effective solution.

Availability and reliability are two essential metrics for the design, deployment, and operation of future ultra-reliable low latency communication (URLLC) networks. Despite a vast amount of research efforts toward URLLC, very little attention has been made on the ultra-reliable communication (URC) aspect of URLLC from a dependability perspective. As an effort toward achieving anytime and anywhere communication, this letter consolidates a dependability theory-based availability concept for individual users by taking into account reliability impairments that affect URC in both spatial and temporal domains . To this end, we perform per-user availability analysis by considering channel status and user mobility patterns in a Poisson Voronoi network.

Recent studies have shown that machine learning can identify individuals with mental illnesses by analyzing their social media posts. Topics and words related to mental health are some of the top predictors. These findings have implications for early detection of mental illnesses. However, they also raise numerous privacy concerns. To fully evaluate the implications for privacy, we analyze the performance of different machine learning models in the absence of tweets that talk about mental illnesses. Our results show that machine learning can be used to make predictions even if the users do not actively talk about their mental illness. To fully understand the implications of these findings, we analyze the features that make these predictions possible. We analyze bag-of-words, word clusters, part of speech n-gram features, and topic models to understand the machine learning model and to discover language patterns that differentiate individuals with mental illnesses from a control group. This analysis confirmed some of the known language patterns and uncovered several new patterns. We then discuss the possible applications of machine learning to identify mental illnesses, the feasibility of such applications, associated privacy implications, and analyze the feasibility of potential mitigations.

Competent transportation infrastructure is one of the prominent factors influencing sustainable economic growth and appreciable living standard in a country. Urban cities habitually equipped with multi-modal traffic flows which encompass various modes of transportation mechanisms such as walking, skating, cycling, motorcycles, three wheelers, private passenger cars, public transport modes like buses or trains, etc. Efficient urban traffic planning has become an exigent task involving complex behavioural analysis of multi-modal traffic scenarios. One possible way of enhancing the travel experience of commuters is by revolutionizing the vehicular communication strategies to devise intelligent routing approaches. Vehicular ad-hoc networks (VANETs) are a special type of mobile ad-hoc networks (MANETs) which can be used to define intelligent conciliator interfaces for vehicle to vehicle (V2V) or vehicle to infrastructure (V2I) communication needs. In this paper, we propose an approach for equilibrating multi-modal traffic environments after road block situations occur through VANET scenario simulations. For this research, we have used OMNeT++ network simulator to simulate the mobile network involved and SUMO microscopic traffic simulator to simulate vehicular movements in parallel and Veins framework as a bridging interface between OMNeT++ and SUMO. We have validated our solution with two heterogeneous metropolitan areas located in the cities of Kandy and Colombo in Sri Lanka. Our solution achieved lower values for waiting time and trip duration of vehicles and qualitatively improving the traffic flow when comparing against the absence of equilibrating module.

Concept Maps and wikis were examined as tools that compliment the engineering design process and better facilitate collaboration between geographically dispersed design teams. Concept maps are a visual knowledge capture medium that can be used to promote the understanding of the design problem and in providing an organizational reference for the entire project. Wikis are editable websites that give the users the freedom to add content in any manner, thus providing a project management platform for design teams. This thesis reports on five studies by the author in which concept maps and wikis are used as both tools for the design process as well as collaborative tools in various modes of interaction. Recommendations to improve concept maps and wiki technologies were determined based on the feedback from the participants in these studies. The wiki was used extensively as a project management tool but required some push for the participants to use. The hierarchical layout structure has been shown to be the favoured pattern by participants in the engineering field. Concept maps are also shown to be a clearer visualization tool than another accepted tool, Unified Modelling Language.

The development of computer-based diagramming tools to support the early stages of engineering design and to improve the capacity to innovate is described. We believe that the human brain is currently the best available device to perform early designing. Thus, good tools will facilitate designers' abilities to 'see' information patterns, reflect on them, and achieve insights they may not have achieved otherwise. Our work suggests that diagram layout and style are at least as important as the textual content for their rapid comprehension. These features can be easier to manage in software than in textual descriptions, leading to tools that are more robust and usable. We are working along several lines of enquiry intended to explore particular aspects of the matter, including using existent tools such as concept maps, and developing alternative tools to test ideas about supporting early design by diagramming. While we have not formally evaluated much of our work, anecdotal evidence is encouraging.

Attention is the fundamental element of effective learning, memory, and interaction. Learning however, with the evolvement of technologies in the modern digital age, has surpassed traditional learning systems to more convenient online or e-learning systems. Nevertheless, unlike in the traditional learning systems, attention detection of a student in an e-learning environment remains one of the barely explored areas in Human Computer Interaction. This study proposes a multimodal ensemble solution to detect the level of attentiveness of a student in an e-learning environment, with the use of computer vision, natural language processing, and deep learning to overcome the barriers in identifying user attention in e-learning. The proposed multimodal captures, processes, and predicts user attentiveness levels of individual students, which are subsequently aggregated through an ensemble model to derive an overall outcome of better accuracy than individual model outcomes. The final outcome of the ensemble model produces a range of percentages, within which the attentiveness level of the student lies during a single online lesson. This range is consequently delivered to the users through an Application Programming Interface.

Objective. Deep brain stimulation is a treatment for medically refractory essential tremor. To improve the therapy, closed-loop approaches are designed to deliver stimulation according to the system’s state, which is constantly monitored by recording a pathological signal associated with symptoms (e.g. brain signal or limb tremor). Since the space of possible closed-loop stimulation strategies is vast and cannot be fully explored experimentally, how to stimulate according to the state should be informed by modeling. A typical modeling goal is to design a stimulation strategy that aims to maximally reduce the Hilbert amplitude of the pathological signal in order to minimize symptoms. Isostables provide a notion of amplitude related to convergence time to the attractor, which can be beneficial in model-based control problems. However, how isostable and Hilbert amplitudes compare when optimizing the amplitude response to stimulation in models constrained by data is unknown. Approach. We formulate a simple closed-loop stimulation strategy based on models previously fitted to phase-locked deep brain stimulation data from essential tremor patients. We compare the performance of this strategy in suppressing oscillatory power when based on Hilbert amplitude and when based on isostable amplitude. We also compare performance to phase-locked stimulation and open-loop high-frequency stimulation. Main results. For our closed-loop phase space stimulation strategy, stimulation based on isostable amplitude is significantly more effective than stimulation based on Hilbert amplitude when amplitude field computation time is limited to minutes. Performance is similar when there are no constraints, however constraints on computation time are expected in clinical applications. Even when computation time is limited to minutes, closed-loop phase space stimulation based on isostable amplitude is advantageous compared to phase-locked stimulation, and is more efficient than high-frequency stimulation. Significance. Our results suggest a potential benefit to using isostable amplitude more broadly for model-based optimization of stimulation in neurological disorders.

Augmenting the human body using wearable technology can be particularly interesting to sense context. The user’s context includes the mental and physical state, which is inferable by detecting facial and head related gestures. For the recognition of these gestures, we propose instrumenting a pair of glasses with Capacitive Sensing (CapSense) technology. We demonstrate proximity sensing with CapSense for mobile use despite its commonly known limitations in context of mobility. Moreover, we demonstrate how to incorporate transparent sensing electrodes into the glass and copper electrodes into the frame while being potentially invisible in a future specs product. We demonstrate an untethered battery-powered glasses prototype, CapGlasses, to sense facial expressions and head gestures. We selected a set of 12 gestures and ran a study with 12 users. We obtained an average accuracy of 89.6% by a user-dependent machine learning model. We focused on providing clear documentation to enable a straightforward replication of our technology.

Local 5G Operator (L5GO) concept is one of the most prominent versatile applications of the 5G in the near future. The popularity of L5GOs will trigger a greater number of roaming and offloading events between mobile operators. However, existing static and the operator-assisted roaming and offloading procedures are inefficient for L5GO ecosystem due to poor service quality, data privacy issues, data transferring delays, excessive costs for intermediary parties and existence of roaming fraud. To address these challenges, we propose a blockchain / Distributed Ledger Technology (DLT) based service platform for L5GOs to facilitate efficient roaming and offload services. As the key contribution, blockchain-based smart contract scheme is proposed to establish dynamic and automated agreements between operators. By using smart contracts, we introduce several novel features such as universal wallet for subscribers, service quality based L5GO rating system, user-initiated roaming process and the roaming fraud prevention system to improve the operational quality of a L5GO. A prototype of the proposed platform is emulated with the Ethereum blockchain platform and Rinkeby Testnet to evaluate the performance and justify the feasibility of the proposal. Upon an extensive evaluation on the prototype, it was observed that the proposed platform offered benefits such as cost effective, more secure and reliable experience.

5G is a promising technology that has the potential to support verticals and applications such as Industrial Internet of Things IoT (IIoT), smart cities, autonomous vehicles, remote surgeries, virtual and augmented realities, and so on. These verticals have a diverse set of network connectivity requirements, and it is challenging to deliver customized services for each by using a common 5G infrastructure. Thus, the operation of Local 5G operator (L5GO) networks or private 5G networks are a viable option to tackle this challenge. A L5GO network is a localized small cell network which can offer tailored service delivery. The adaptation of network softwarization in 5G allows vertical owners to deploy and operate L5GO networks. However, the deployment of L5GOs raises various issues related to management of subscribers, roaming users, spectrum, security, and also the infrastructure. This paper proposes a blockchain-based platform to address these issues. The paper introduces a set of blockchain-based modularized functions such as service rating systems, bidding techniques, and selection functions, which can be used to deploy different services for L5GOs. Exploitation of blockchain technology ensures availability, non-reliance on trusted third parties, secure transfer payments, and stands to gain many more advantages. The performance and the viability of the proposed platform are analyzed by using simulations and a prototype implementation. European Commission Academy of Finland

Sequencing in metagenomes opens new ways of analyzing genomics in microbial communities in their habitats. Analyzing metagenomes has been a challenge not only because they are acquired from unknown collections without any prior lab-based cultivation, but also the volume. The value of these metagenomic data can be greatly enhanced by integrating with an informative and interactive visualization tool. Need for such tools is growing but still insufficient in number. MetaG, a standalone metagenomic analysis tool which converts higher dimensional sequence data into lower dimensional visualizations through deep learning techniques, is uniquely designed to produce rich lower dimensional representations of oligonucleotide frequency vectors of the population-level genomic diversity of the microbial organisms. Users can opt their preference in dimensionality reduction, between Principal Component Analysis (PCA), t- distributed Stochastic Neighbor Embedding (t-SNE) and autoencoders. Additionally, MetaG provides information on the presence, diversity and abundance of organisms by analyzing alignments against a database that contains taxonomy information. It caters more flexibility with number of visualizations, providing insights of complex microbial communities. MetaG, written in Python is free and open-source with the code publicly accessible.

Omniscient is a system that enables higher-level management of massive organizations to remotely monitor and scrutinize the activities that take place in the branches from the head office itself by providing exclusive insight in the form of detailed reports on the employees' behaviour and performance daily, weekly and monthly. The system further monitors the branch and provides reports on any suspicious behaviour and also on the customers' activity within the branch premises. Omniscient rates the customer's level of satisfaction by capturing the customer's facial expressions and analyzing their emotions while they are being served. The employee face and dress recognition models have accuracies of 90.90% and 87.00% respectively while, employee activity detection has an accuracy of 89.00%. Customer emotion and miscellaneous activities detection models have the accuracies of 91.50% and 83.00% respectively. All of the aforementioned procedures were made possible by systematically analyzing the IP camera video footage obtained throughout the day to analyze the work productivity and performance of the branch as accurately as possible using deep learning and modern visual computing techniques like CNN, OpenCV, Haar Cascade classifier, face recognition, Dlib and Darknet.

Sri Lankan Agriculture sector can be considered as a crucial component as it contributes 18% of country GDP. As native farmers still cling to inapplicable traditional theorems and practices to track customer's vegetable consumption trends, they failed to assure a “good price” for their harvest. Also, the plants are prone to many diseases and pests' attacks which causes loss of the harvest. Unreliable problem identification, poor knowledge on application of fertilizers and pesticides have caused the farmers to lose their profits. As a solution to mitigate these problems, this study has built a computerized system with a vegetable price prediction system and a plant disease, pest identification system. Taking Potato as an example, the parameters of the time series model were analyzed through experiment and has built the price predictor using ARIMA model. Also, with advanced Image processing and CNN techniques Plant disease, pest identifier has built. Desirable results of the entire system have been achieved with more than 94%-97% rate of accuracy. The ultimate goal of this study is to achieve the optimal growth of the sector by navigating the users for a quality and effective decision making by reliable market trends and problem identification.

With close to 4 billion origin-destination passenger journeys worldwide, airline operations have become a crucial factor in the global economy. With the increasing number of journeys and passengers, managing the daily operations of airlines have become a complicated task. We have proposed a real-time decision optimization platform for airline operations with the following subsystems: (1) determine the optimum path for a flight, (2) optimum fleet assignment, (3) optimum gate allocation, (4) optimum crew allocation. We have used an approximation (heuristics) based optimization approach: Genetic Programming (GP) to implement the modules. The results of our proposed platform illustrate that, the decision-making process of Airline Operations Control Center (AOCC) can be optimized, and dynamic change requirements can be accommodated.

Direct sampling methods are increasingly being used to solve the inverse medium scattering problem to estimate the shape of the scattering object. A simple direct method using one incident wave and multiple measurements was proposed by Ito, Jin and Zou [1]. In this report, we performed some analytic and numerical studies of the direct sampling method. The method was found to be effective in general. However, there are a few exceptions exposed in the investigation. Analytic solutions in different situations were studied to verify the viability of the method while numerical tests were used to validate the effectiveness of the method.

Opinions hindered in social media, and online news feeds to control the predictability of financial markets such as Forex and cause huge profits and losses for the investors. Due to this reason, financial market prediction through opinion mining has gained significant attention from researchers recently. Current approaches relied more on basic text mining algorithms and overlooked the knowledge generated through the integration of the multiple data sources. To overcome this deficiency, this study harnesses the knowledge required for decisive currency-pair prediction by analyzing forex news and social media data using ensemble classification. Three classification approaches namely Support Vector Machine (SVM), Naive Bayes and K-Nearest Neighbor (KNN) are used in the ensemble classifier. Based on the classifier, impactful posts/comments are discovered to facilitate forex market prediction.

Dogs are one of the most popular pets in the world. It is usual that pet owners are always concerned about the health and the wellbeing of their pets. The activity levels of the dogs vary from each other based on breed and age. Tracking the behavioral changes using image processing and machine learning concepts and notifying the pet owners via a mobile application is the main objective of this research. Breed recognition has been done applying deep learning concepts to the user-uploaded video or the photograph of the dog. This research mainly focuses on walking, running, resting, and barking activity patterns of the dog. A surveillance camera and sensors were the main equipment for data collection. The audio feature of the surveillance camera is used to identity the barking behavior of the dog. Dogs from different ages belonging to Pomeranian and German Shepherd breeds have been selected for this experiment. Transfer learning with ResNet50, Inception V3, and support vector machines have been used to recognize and classify the activities of the dogs. The research study was able to achieve the accuracy levels as follows: - breed recognition - 89%+, walking pattern recognition - 99.5%, resting pattern recognition - 97% and barking pattern recognition - 60%. With the above accuracy levels, the research was able to identify the unusual behaviour of the dogs.

Flood emergency management has been a major issue in the last few decades as it can disrupt human lives as well as the economy and property damage. A flood occurs when the overflow of water that melts relatively dry land. In the hydrology discipline, floods are a field of study and they are the most common and widespread unpredictable weather occurrence of natural sources. Floods can look quite different because anywhere from a few inches of water to several feet is affected by flooding. They can also come on suddenly, or slowly increase. Therefore frequent identification of flood impact levels is very important. This study aims to create a multimedia big data platform with a deep learning approach for flood emergency management. It uses multimedia data as they are freely available as social media data (Twitter and Facebook), satellite image data, crowdsourcing, and sensor network data for mining purposes. As this research based on deep learning and image processing cutting edge technologies, authorities can identify impact level using satellite images and provide a real-time warning for the people or people can use this for self-estimation of flood risk level when they want in their day to day life like detecting passable or low-risk roads in flooding time. The deep neural network plays a major role in feature extraction and data augmentation helps to increase the number of images in the dataset. This study provides a comparative study between VGG16, VGG19, Densetnet169, and MobileNet deep learning models and evaluates the performance by using training and testing data. Dataset compromises of over 1500 data and the conclusions drawn from work prove that the MobileNet model worked with 86% accuracy with high performance. In the latter part of the paper, it will describe future recommendations.

According to aesthetic evaluations, hair is the most unique feature which can enhance the facial features of a person. Beauty experts have identified that 70% of overall face appearance completely depends on the haircut or hairstyle. The physical attributes such as the haircut is a major determinant of women's psychology. This is the essence of why a haircut which is matching a woman's face is necessary articulation. But selecting the right haircut or hairstyle is one of the most difficult decisions to take in a woman's life. This paper presents a novel framework to select the most suitable hairstyle or haircut by classifying the face shape. The author considers the shape of the face, beauty experts knowledge related to hair cuts and hairstyles and the length of the hair to develop a model to recommend the most suitable hairstyle or haircut. The author focused to recommend the haircuts and hairstyles for women which is a subsection of this large research area. According to beauty experts identifying the shape of the face is the most important step before selecting the right hairstyle or haircut. The proposed model has the ability to classify the face shape when a user uploaded a portrait of herself. Machine Learning libraries were used to identify the landmarks of the face image and classify the face in the correct shape. Naive Bayes classification algorithm has used to recommend the most suitable hairstyle or haircut according to the detected face shape., hair length and information collected from the hair experts. User has given an option to share the recommended hair style or haircut with the beautician via “The Beauty Quest” Salon network platform. Five thousand images were trained, and python language has used as the programming language. The accuracy of the face shape classification model is 91% and the accuracy of the hair recommendation is also 83%.

Scientific workflows are widely used in High Performance Computing environments to carry out complex sets of interrelated calculations. As executing tasks in a workflow in the form of individual jobs substantially increase the makespan of the workflow, tasks are grouped into several clusters before executing. Task clustering helps to minimize the makespan of workflows by decreasing overheads in job execution as less number of jobs are required to be executed when tasks are clustered. However, clustering tasks with the sole intension of reducing the makespan of the workflow can lead to serious resource underutilization if tasks having significantly different resource requirements are clustered together. Therefore, it is critical to identify the trade-off between the makespan and the resource utilization to ensure task clustering reduces the makespan of a workflow while maximizing its resource utilization. This paper introduces a new task clustering algorithm called Resource Aware Clustering algorithm which uses a novel metric called Resource Aware Clustering Coefficient that aims to maximize both the makespan improvement and the resource utilization of a workflow. The proposed algorithm improves the resource utilization of the considered workflows by more than 34% compared to other baseline task clustering algorithms while having a competitive makespan improvement compared to their unclustered workflows.

News articles are important for providing timely, historic information. However, the Internet is replete with text that may contain irrelevant or unhelpful information, therefore means of processing it and distilling content is important and useful to human readers as well as information extracting tools. Some common questions we may want to answer are “what is this article about?” and “who wrote it?”. In this work we compare machine learning models for evaluating two common NLP tasks, topic and authorship attribution, on the 2017 Vox Media dataset. Additionally, we use the models to classify on a subsection, about ~20%, of the original text which show to be better for classification than the provided blurbs. Because of the large number of topics, we take into account topic overlap and address it via top-n accuracy and hierarchical groupings of topics. We also consider edge cases in authorship by classifying on inter-topic and intra-topic author distributions. Our results show that both topics and authors readily identifiable consistently perform best when using neural networks rather than support vector, random forests, or naive Bayes classifiers, although the latter methods perform acceptably.