Ubiquitous Devices Research Articles

A Novel Lightweight Human Activity Recognition Method Via L-CTCN.

Wi-Fi-based human activity recognition has attracted significant attention. Deep learning methods are widely used to achieve feature representation and activity sensing. While more learnable parameters in the neural networks model lead to richer feature extraction, it results in significant resource consumption, rendering the model unsuitable for lightweight Internet of Things (IoT) devices. Furthermore, the sensing performance heavily relies on the quality and quantity of data, which is a time-consuming and labor-intensive task. Therefore, there is a need to explore methods that reduce the dependence on the quality and quantity of the dataset while ensuring recognition performance and decreasing model complexity to adapt to ubiquitous lightweight IoT devices. In this paper, we propose a novel Lightweight-Complex Temporal Convolution Network (L-CTCN) for human activity recognition. Specifically, this approach effectively combines complex convolution with a Temporal Convolution Network (TCN). Complex convolution can extract richer information from limited raw complex data, reducing the reliance on the quality and quantity of training samples. Based on the designed TCN framework with 1D convolution and residual blocks, the proposed model can achieve lightweight human activity recognition. Extensive experiments verify the effectiveness of the proposed method. We can achieve an average recognition accuracy of 96.6% with only 0.17 M parameter size. This method performs well under conditions of low sampling rates and a low number of subcarriers and samples.

Forward Secrecy Attack on Privacy Preserving Machine Authenticated Key Agreement for Internet of Things

Internet of things (IoT) is to connect billions of devices and machines via Internet and to have a smart system. Sensors and devices in IoT environment are connected and communicated together. Connecting such a huge number of devices requires high level of security and privacy. A crucial characteristic of ubiquitous IoT devices is their limited resources. In recent times, a scheme for privacy-preserving machine authenticated key agreement scheme (PPMAKA) has been introduced for the IoT environment. It was argued that PPMAKA provides security and privacy at the same time including forward secrecy. Nevertheless, this paper will demonstrate that PPMAKA lacks forward secrecy, a crucial security and privacy feature in the IoT environment. We use Cannetti and Krawzyck threat model for the detailed analysis of PPMAKA. Furthermore, we provide remarks for the future research as it is recommendable to design any security and privacy schemes over IoT environments with lightweight operation and communication property, authenticated key agreement with forward secrecy, anonymity and unlinkability.

DLMDish: Using Applied Deep Learning and Computer Vision to Automatically Classify Mauritian Dishes

The benefits of using an automatic dietary assessment system for accompanying diabetes patients and prediabetic persons to control the risk factor also referred to as the obesity “pandemic” are now widely proven and accepted. However, there is no universal solution as eating habits of people are dependent on context and culture. This project is the cornerstone for future works of researchers and health professionals in the field of automatic dietary assessment of Mauritian dishes. We propose a process to produce a food dataset for Mauritian dishes using the Generative Adversarial Network (GAN) and a fine Convolutional Neural Network (CNN) model for identifying Mauritian food dishes. The outputs and findings of this research can be used in the process of automatic calorie calculation and food recommendation, primarily using ubiquitous devices like mobile phones via mobile applications. Using the Adam optimizer with carefully fixed hyper-parameters, we achieved an Accuracy of 95.66% and Loss of 3.5% as concerns the recognition task.

Electronic Materials Accelerating the Development of Ubiquitous Devices

Electronic Materials Accelerating the Development of Ubiquitous Devices

Enhancing Asynchronous Online Learning with Mobile Augmented Reality Application: A Case Study in Higher Education Business Courses

This case study outlines the development and utilization of a Mobile Augmented Reality (MAR) application to teach an asynchronous online lesson on e-commerce. The MAR technology was leveraged primarily to enhance the presentation of online learning materials, add interactivity to the learning process, and enable students to access the lesson from anywhere via their ubiquitous mobile devices. Two groups, comprising a total of 105 business students, participated in the online lesson during the COVID pandemic. Students' experiences of the implementation were captured through an online survey to find out how the MAR technology was perceived, and what were the benefits and challenges of using the technology. This study contributes to the research on the use of augmented reality in education and offers practical recommendations for teachers to consider when designing and implementing MAR online lessons, including the importance of learner-centered design, careful guidance on technology use, and encouragement of student interaction. The study concludes that augmented reality is a useful tool for improving learning materials and asynchronous online learning practices and MAR applications can be effective for learning with compact materials and micro-credentials.

Electronic Materials Accelerating the Development of Ubiquitous Devices

Electronic Materials Accelerating the Development of Ubiquitous Devices

CATS: Conditional Adversarial Trajectory Synthesis for privacy-preserving trajectory data publication using deep learning approaches

The prevalence of ubiquitous location-aware devices and mobile Internet enables us to collect massive individual-level trajectory dataset from users. Such trajectory big data bring new opportunities to human mobility research but also raise public concerns with regard to location privacy. In this work, we present the Conditional Adversarial Trajectory Synthesis (CATS), a deep-learning-based GeoAI methodological framework for privacy-preserving trajectory data generation and publication. CATS applies K-anonymity to the underlying spatiotemporal distributions of human movements, which provides a distributional-level strong privacy guarantee. By leveraging conditional adversarial training on K-anonymized human mobility matrices, trajectory global context learning using the attention-based mechanism, and recurrent bipartite graph matching of adjacent trajectory points, CATS is able to reconstruct trajectory topology from conditionally sampled locations and generate high-quality individual-level synthetic trajectory data, which can serve as supplements or alternatives to raw data for privacy-preserving trajectory data publication. The experiment results on over 90k GPS trajectories show that our method has a better performance in privacy preservation, spatiotemporal characteristic preservation, and downstream utility compared with baseline methods, which brings new insights into privacy-preserving human mobility research using generative AI techniques and explores data ethics issues in GIScience.

Society's Attitudes Towards Human Augmentation and Performance Enhancement Technologies (SHAPE) Scale

Human augmentation technologies (ATs) are a subset of ubiquitous on-body devices designed to improve cognitive, sensory, and motor capacities. Although there is a large corpus of knowledge concerning ATs, less is known about societal attitudes towards them and how they shift over time. To that end, we developed The Society's Attitudes Towards Human Augmentation and Performance Enhancement Technologies (SHAPE) Scale, which measures how users of ATs are perceived. To develop the scale, we first created a list of possible scale items based on past work on how people respond to new technologies. The items were then reviewed by experts. Next, we performed exploratory factor analysis to reduce the scale to its final length of thirteen items. Subsequently, we confirmed test-retest validity of our instrument, as well as its construct validity. The SHAPE scale enables researchers and practitioners to understand elements contributing to attitudes toward augmentation technology users. The SHAPE scale assists designers of ATs in designing artifacts that will be more universally accepted.

Traditional surveys versus ecological momentary assessments: Digital citizen science approaches to improve ethical physical activity surveillance among youth.

The role of physical activity (PA) in minimizing non-communicable diseases is well established. Measurement bias can be reduced via ecological momentary assessments (EMAs) deployed via citizen-owned smartphones. This study aims to engage citizen scientists to understand how PA reported digitally by retrospective and prospective measures varies within the same cohort. This study used the digital citizen science approach to collaborate with citizen scientists, aged 13-21 years over eight consecutive days via a custom-built app. Citizen scientists were recruited through schools in Regina, Saskatchewan, Canada in 2018 (August 31-December 31). Retrospective PA was assessed through a survey, which was adapted from three validated PA surveys to suit smartphone-based data collection, and prospective PA was assessed through time-triggered EMAs deployed consecutively every day, from day 1 to day 8, including weekdays and weekends. Data analyses included paired t-tests to understand the difference in PA reported retrospectively and prospectively, and linear regressions to assess contextual and demographic factors associated with PA reported retrospectively and prospectively. Findings showed a significant difference between PA reported retrospectively and prospectively (p = 0.001). Ethnicity (visible minorities: β = - 0.911, 95% C.I. = -1.677, -0.146), parental education (university: β = 0.978, 95% C.I. = 0.308, 1.649), and strength training (at least one day: β = 0.932, 95% C.I. = 0.108, 1.755) were associated with PA reported prospectively. In contrast, the number of active friends (at least one friend: β = 0.741, 95% C.I. = 0.026, 1.458) was associated with retrospective PA. Physical inactivity is the fourth leading cause of mortality globally, which requires accurate monitoring to inform population health interventions. In this digital age, where ubiquitous devices provide real-time engagement capabilities, digital citizen science can transform how we measure behaviours using citizen-owned ubiquitous digital tools to support prevention and treatment of non-communicable diseases.

The Impact of Artificial Intelligence (AI) on Customer Relationship Management: A Qualitative Study

Ever since the commercialization of the Internet in the '90s, technology has been evolving faster than ever with the advent of cloud computing, social media, ubiquitous mobile devices, the Internet of Things (IoT), blockchain, and more. A staggering number of three billion internet users, five billion mobile users, and six billion devices are now connected through this massive global network of networks, facilitating customer information exchange and interaction never before seen in history. Driven by recent technological advances in computing power, big data, high-speed internet connection, and easier access to models built with advanced algorithms, Artificial Intelligence (AI) is the next wave of innovation, which has already come into widespread awareness in the consumer world with the emergence of virtual assistants and chatbots (e.g., Amazon's Alexa, Apple's Siri, Google's Assistant), image recognition (e.g., Facebook Photos, Google ImageNet), personalized recommendations (e.g., Netflix, Amazon) and autonomous driving (e.g., Tesla, Google Waymo). This qualitative research study intends to learn about the impact of AI on customer relationship management (CRM), specifically in the area of customer service of problem resolution. Most prior research focuses on the AI technologies leveraged in CRM systems, such as machine learning, natural language processing, voice recognition, chatbots, data analytics, and cloud infrastructure. Few extant studies have used a qualitative research methodology to gather data from industry experts to truly understand the impact of AI technologies on customer relationship management, especially in the area of customer service and problem resolution. This study aims to fill this research gap. This research contributes to the literature on AI in the context of CRM and is of value to both academics and practitioners as it provides a detailed analysis and documentation of the impact of AI on the customer service domain.

Electronic Materials Accelerating the Development of Ubiquitous Devices

Electronic Materials Accelerating the Development of Ubiquitous Devices

CAN APPLE LIDAR CAMERAS BE RELIABLY USED FOR COASTAL MONITORING?

Smart phones are becoming ubiquitous devices that can be exploited for citizen science applications, including coastal monitoring (Harley et al., 2019). In 2020, Apple released new versions of their iPhone (12+) and iPad Pro (2020+) that included built-in LIDAR functionality. These were accompanied by a series of commercial applications (e.g., 3DScanner, Scaniverse, and many others) built to capture high resolution objects at close range using mesh or point-cloud formulations. The built-in GPS of the iPhone/iPads however is of low accuracy (~5m horizontal) compared to the demands for scientific surveys. Fortuitously, commercial hand-held RTK-GNSS units (e.g., EOS Arrow Gold, Pix4D viDoc, RedCatch) that can talk directly to the smart phones and provide surveygrade positioning have also been recently developed. The Pix4D viDoc and RedCatch devices are targeted at photogrammetry applications rather than the internal LIDAR scanner, while the EOS Arrow family is application independent. In the past year, several papers have been published examining the ability of smartphones equipped with LIDAR to be used in various geophysical applications ( Luetzenburg et al. 2021; Mokros et al. 2021; Spreafico et al. 2021). This then begs the question – can these smart phones equipped with LIDAR be reliably used for coastal monitoring with existing (free) commercial applications?

Racially fair pupillometry measurements for RGB smartphone cameras using the far red spectrum

Pupillometry is a measurement of pupil dilation commonly performed as part of neurological assessments. Prior work have demonstrated the potential for pupillometry in screening or diagnosing a number of neurological disorders including Alzheimer’s Disease, Schizophrenia, and Traumatic Brain Injury. Unfortunately, the expense and inaccessibility of specialized pupilometers that image in the near infrared spectrum limit the measurement to high resource clinics or institutions. Ideally, this measurement could be available via ubiquitous devices like smartphones or tablets with integrated visible spectrum imaging systems. In the visible spectrum of RGB cameras, the melanin in the iris absorbs light such that it is difficult to distinguish the pupil aperature that appears black. In this paper, we propose a novel pupillometry technique to enable smartphone RGB cameras to effectively differentiate the pupil from the iris. The proposed system utilizes a 630 nm long-pass filter to image in the far red (630–700 nm) spectrum, where the melanin in the iris reflects light to appear brighter in constrast to the dark pupil. Using a convolutional neural network, the proposed system measures pupil diameter as it dynamically changes in a frame by frame video. Comparing across 4 different smartphone models, the pupil-iris contrast of N = 12 participants increases by an average of 451% with the proposed system. In a validation study of N = 11 participants comparing the relative pupil change in the proposed system to a Neuroptics PLR-3000 Pupillometer during a pupillary light response test, the prototype system acheived a mean absolute error of 2.4%.

Electronic Materials Accelerating the Development of Ubiquitous Devices

Electronic Materials Accelerating the Development of Ubiquitous Devices

Biosignals meet Adaptive Systems

This paper presents a perspective on Biosignal-Adaptive Systems (BAS) which automatically adapt to user needs by continuously interpreting their biosignals and by providing transparent feedback, thereby keeping the user in the loop. The major hallmark of the described BAS is the low latency with which biosignals are processed, interpreted, and applied to perform rapid system adaptation, providing the user in the loop with immediate feedback on the BAS’s understanding of his or her condition. In contrast to explicit user input or the interpretation of observable behavior, the rapid system adaptation relies on biosignals, which in context of a concrete application can be interpreted as implicit signals of user needs. Recently, great advances have been made in sensor integration into smart devices, making it possible to collect vasts amounts of multimodal biosignal data. Furthermore, powerful machine learning methods enable rapidly processing and learning from such data. We argue that the time has come to harness the full spectrum of low-latency processing of biosignals to understand user needs and to apply this information to deliver adaptive systems accordingly. However, this will just be the beginning: real-time signal processing in combination with ubiquitous devices that are always connected to huge processing and storage capacities allow systems to provide users (and bystanders) with instant and transparent feedback and adaptations for the recognized needs. In the future, such systems could run 24/7 to assist users @home, @work, and @play from the cradle to the grave. Thus, BAS must be human-centric to curate data, archive information, learn from experience, and adapt to individual users throughout their lifespan. In this position paper, we present the concept of BAS with its key building blocks, provide selected examples of BAS from our research, and articulate selected challenges for future research.

Dual facet passivation of silver halometallate for eco-friendly silver bismuth sulfide near IR photodetector

Ternary chalcogenide silver bismuth sulfide nanocrystals (AgBiS2 NCs) have taken great strides in the past few years to emerge as one of the better eco-friendly alternatives to compete with the prevalent toxic semiconductor materials such as lead sulfide quantum dots (PbS QDs) in the near-infrared (NIR) region. Nevertheless, their implementation in photodetectors has been scarce due to high dark current and complicated solid-state ligand exchange fabrication steps involved, resulting in a lower overall detectivity. The performance is further deemed to be stunted due to the difficulty associated with the passivation of the charge-neutral (100) facet of larger AgBiS2 NCs efficiently. In this work, we aimed to develop a mixed ‘halometallate’ ligand approach, wherein we introduce silver bromide (AgBr) as an ancillary ligand to silver iodide (AgI), passivating both (100) and (111) facets of cubic AgBiS2 solids in a facile solution-phase ligand exchange step to obtain highly dispersible colloidal ink. Decreased bond length, bond angle (Br-Ag-Br), and ionic size of [AgBr2-] anion induces less compressive strain compared to [AgI2–], culminating in higher molecular stability on the AgBiS2 surface. This dual passivation reduces the dark current to 6.01 × 10-7 A cm−2 and a high specific detectivity of 1.8 × 1012 Jones at 800 nm is achieved, comparable to ubiquitous PbS QD devices. We also demonstrate diminished in-gap carrier density population, enhanced light detection, and ultrafast microsecond response at higher wavelengths operating under high bias (-1V) photoconductive mode. This study illustrates the role of optimum surface coverage in eliminating the deleterious non-radiative recombination defect centers by introducing additional ligands in solution-processed AgBiS2 NC and the viability of the mixed ligand approach for stable eco-friendly NIR photodetectors.

An edge re-ordering based acceleration architecture for improving data locality in graph analytics applications

Data structure is the key in Edge Computing where various types of data are continuously generated by ubiquitous devices. Within all common data structures, graphs are used to express relationships and dependencies among human identities, objects, and locations. They are also expected to become one of the most important data infrastructures in the near future. Furthermore, as graph processing often requires random accesses to vast memory spaces, conventional memory hierarchies with caches cannot work efficiently. To alleviate such memory access bottlenecks in graph processing, we present a solution through vertex accesses scheduling and edge array re-ordering, in parallel with the execution of graph processing applications to improve both temporal and spatial locality of memory accesses, especially for edge-centric graph analytics which are popular means in handling dynamic graphs. Our proposed architecture is evaluated and tested through both trace-based cache simulations and cycle-accurate FPGA-based prototyping. Evaluation results show that our proposal has a potential of significantly reducing the Last Level Cache (LLC) misses by 62.60% in general among PageRank and BFS algorithms. Meanwhile, evaluations with the FPGA prototype successfully reduce the quantity of Miss-Per-Kilo-Instructions (MPKI) for LLC by 56.27% on average.

A Review on Smartphone Keystroke Dynamics as a Digital Biomarker for Understanding Neurocognitive Functioning.

Can digital technologies provide a passive unobtrusive means to observe and study cognition outside of the laboratory? Previously, cognitive assessments and monitoring were conducted in a laboratory or clinical setting, allowing for a cross-sectional glimpse of cognitive states. In the last decade, researchers have been utilizing technological advances and devices to explore ways of assessing cognition in the real world. We propose that the virtual keyboard of smartphones, an increasingly ubiquitous digital device, can provide the ideal conduit for passive data collection to study cognition. Passive data collection occurs without the active engagement of a participant and allows for near-continuous, objective data collection. Most importantly, this data collection can occur in the real world, capturing authentic datapoints. This method of data collection and its analyses provide a more comprehensive and potentially more suitable insight into cognitive states, as intra-individual cognitive fluctuations over time have shown to be an early manifestation of cognitive decline. We review different ways passive data, centered around keystroke dynamics, collected from smartphones, have been used to assess and evaluate cognition. We also discuss gaps in the literature where future directions of utilizing passive data can continue to provide inferences into cognition and elaborate on the importance of digital data privacy and consent.

Ageing-aware battery discharge prediction with deep learning

Electrochemical batteries are ubiquitous devices in our society. When employed in mission-critical applications, the ability to precisely predict their end-of-discharge under highly variable operating conditions is of paramount importance in order to support operational decision-making and to fully exploit the entire battery’s lifetime. While there are accurate predictive models of the processes underlying the charge and discharge phases, the modelling of ageing remains an open challenge. This lack of understanding often leads to inaccurate models or the need for time-consuming calibration procedures whenever the battery ages or its conditions change significantly. This represents a major obstacle to the real-world deployment of efficient and robust battery management systems. In this paper, we introduce Dynaformer, a novel deep learning architecture which is able to simultaneously infer the ageing state from a limited number of voltage/current samples and predict the full voltage discharge curve for real batteries with high precision. In the first step of our evaluation, we investigate the performance of the proposed framework on simulated data. In the second step, we demonstrate that a minimal amount of fine-tuning allows Dynaformer to bridge the simulation-to-real gap between simulations and real data collected from a set of batteries. The proposed methodology enables the utilization of battery-powered systems until the end of discharge in a controlled and predictable way, thereby significantly prolonging the operating cycles and reducing costs.

Dap-FL: Federated Learning Flourishes by Adaptive Tuning and Secure Aggregation

Federated learning (FL), an attractive and promising distributed machine learning paradigm, has sparked extensive interest in exploiting tremendous data stored on ubiquitous mobile devices. However, conventional FL suffers severely from resource heterogeneity, as clients with weak computational and communication capabilities may be unable to complete local training using the same local training hyper-parameters. In this paper, we propose Dap-FL, a deep deterministic policy gradient (DDPG)-assisted adaptive FL system, in which local learning rates and local training epochs are adaptively adjusted by all resource-heterogeneous clients through locally deployed DDPG-assisted adaptive hyper-parameter selection schemes. Particularly, the rationality of the proposed hyper-parameter selection scheme is confirmed through rigorous mathematical proof. Besides, due to the thoughtlessness of security consideration of adaptive FL systems in previous studies, we introduce the Paillier cryptosystem to aggregate local models in a secure and privacy-preserving manner. Rigorous analyses show that the proposed Dap-FL system could protect clients' private local models against chosen-plaintext attacks and chosen-message attacks in a widely used honest-but-curious participants and active adversaries security model. More importantly, through ingenious and extensive experiments, the proposed Dap-FL achieves higher model prediction accuracy than two state-of-the-art RL-assisted FL methods, i.e., 6.03% higher than DDPG-based FL and 7.85% higher than DQN-based FL. In addition, experimental results also show that the proposed Dap-FL achieves higher global model prediction accuracy and faster convergence rates than conventional FL, and the comprehensiveness of the adjusted local training hyper-parameters is validated.