[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Next Issue
Volume 15, October
Previous Issue
Volume 15, August
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 

Future Internet, Volume 15, Issue 9 (September 2023) – 36 articles

Cover Story (view full-size image): Digital twins enhance the intelligent manufacturing process by replicating system behavior, offering real-time analysis, and predictive maintenance during operations. Due to modeling complexity, most of them use purely data-driven approaches without considering physical phenomena. To attenuate modeling error and increase modeling interpretability, this paper proposes a novel approach, using neural ODEs and physical dynamical equations for cooling fan system digital twins. This hybrid modeling approach achieves accurate prediction with fewer parameters and is robust against unexpected input patterns. This work demonstrates great potential for adapting neural networks into a physical or mathematical framework, enabling more intelligent and robust manufacturing processes in the future. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Select all
Export citation of selected articles as:
23 pages, 5452 KiB  
Article
Evaluation of Blockchain Networks’ Scalability Limitations in Low-Powered Internet of Things (IoT) Sensor Networks
by Kithmini Godewatte Arachchige, Philip Branch and Jason But
Future Internet 2023, 15(9), 317; https://doi.org/10.3390/fi15090317 - 21 Sep 2023
Cited by 2 | Viewed by 2612
Abstract
With the development of Internet of Things (IoT) technologies, industries such as healthcare have started using low-powered sensor-based devices. Because IoT devices are typically low-powered, they are susceptible to cyber intrusions. As an emerging information security solution, blockchain technology has considerable potential for [...] Read more.
With the development of Internet of Things (IoT) technologies, industries such as healthcare have started using low-powered sensor-based devices. Because IoT devices are typically low-powered, they are susceptible to cyber intrusions. As an emerging information security solution, blockchain technology has considerable potential for protecting low-powered IoT end devices. Blockchain technology provides promising security features such as cryptography, hash functions, time stamps, and a distributed ledger function. Therefore, blockchain technology can be a robust security technology for securing IoT low-powered devices. However, the integration of blockchain and IoT technologies raises a number of research questions. Scalability is one of the most significant. Blockchain’ scalability of low-powered sensor networks needs to be evaluated to identify the practical application of both technologies in low-powered sensor networks. In this paper, we analyse the scalability limitations of three commonly used blockchain algorithms running on low-powered single-board computers communicating in a wireless sensor network. We assess the scalability limitations of three blockchain networks as we increase the number of nodes. Our analysis shows considerable scalability variations between three blockchain networks. The results indicate that some blockchain networks can have over 800 ms network latency and some blockchain networks may use a bandwidth over 1600 Kbps. This work will contribute to developing efficient blockchain-based IoT sensor networks. Full article
Show Figures

Figure 1

Figure 1
<p>Research architecture.</p>
Full article ">Figure 2
<p>Prototype blockchain sensor network.</p>
Full article ">Figure 3
<p>Latency variability of blockchain networks.</p>
Full article ">Figure 4
<p>Latency standard deviations of blockchain networks.</p>
Full article ">Figure 5
<p>Mean blockchain networks latency variability.</p>
Full article ">Figure 6
<p>Overall mean latency of the networks.</p>
Full article ">Figure 7
<p>Hydrachain blockchain bandwidth usage.</p>
Full article ">Figure 8
<p>Monero blockchain bandwidth usage.</p>
Full article ">Figure 9
<p>Duino coin blockchain bandwidth usage.</p>
Full article ">Figure 10
<p>Mean blockchain network bandwidth usage.</p>
Full article ">Figure 11
<p>Overall mean blockchain network bandwidth usages.</p>
Full article ">Figure 12
<p>Block transaction rate of 7 Hydrachain blockchain nodes.</p>
Full article ">Figure 13
<p>Block transaction rate of 15 Hydrachain blockchain nodes.</p>
Full article ">Figure 14
<p>Block transaction rate of 20 Hydrachain blockchain nodes.</p>
Full article ">Figure 15
<p>Block transaction rate of 7 Monero blockchain nodes.</p>
Full article ">Figure 16
<p>Block transaction rate of 15 Monero blockchain nodes.</p>
Full article ">Figure 17
<p>Block transaction rate of 20 Monero blockchain nodes.</p>
Full article ">Figure 18
<p>Block transaction rate of 7 Duino coin blockchain nodes.</p>
Full article ">Figure 19
<p>Block transaction rate of 15 Duino coin blockchain nodes.</p>
Full article ">Figure 20
<p>Block transaction rate of 20 Duino coin blockchain nodes.</p>
Full article ">Figure 21
<p>Summary of average mean block transaction rates.</p>
Full article ">
21 pages, 11273 KiB  
Article
Technical, Qualitative and Energy Analysis of Wireless Control Modules for Distributed Smart Home Systems
by Andrzej Ożadowicz
Future Internet 2023, 15(9), 316; https://doi.org/10.3390/fi15090316 - 20 Sep 2023
Cited by 1 | Viewed by 2213
Abstract
Distributed smart home systems using wireless communication are increasingly installed and operated in households. Their popularity is due to the ease of installation and configuration. This paper presents a comprehensive technical, quality, and energy analysis of several popular smart home modules. Specifically, it [...] Read more.
Distributed smart home systems using wireless communication are increasingly installed and operated in households. Their popularity is due to the ease of installation and configuration. This paper presents a comprehensive technical, quality, and energy analysis of several popular smart home modules. Specifically, it focuses on verifying their power consumption levels, both in standby and active mode, to assess their impact on the energy efficiency of building installations. This is an important aspect in the context of their continuous operation, as well as in relation to the relatively lower power of loads popular in buildings, such as LED lighting. The author presents the results of measurements carried out for seven different smart home modules controlling seven different types of loads. The analysis of the results shows a significant share of home automation modules in the energy balance; in particular, the appearance of reactive power consumption due to the installation of smart home modules is noteworthy. Bearing in mind all the threads of the analysis and discussion of the results of measurement experiments, a short SWOT analysis is presented, with an indication of important issues in the context of further development of smart systems and the Internet of Things with wireless communication interfaces, dedicated to home and building applications. Full article
(This article belongs to the Special Issue Artificial Intelligence and Blockchain Technology for Smart Cities)
Show Figures

Figure 1

Figure 1
<p>An electrical connection diagram on the measuring stand, with the Sonel PQM-711 analyzer connected to measure one of the smart home modules—Blebox Switchbox.</p>
Full article ">Figure 2
<p>Real view of the measurement stand during the measurement procedure.</p>
Full article ">Figure 3
<p>The Sonoff devices: (<b>a</b>) Sonoff ZBMINI switch module and (<b>b</b>) ZigBee–Wi-Fi bridge.</p>
Full article ">Figure 4
<p>The Sonoff Socket module: (<b>a</b>) front view and (<b>b</b>) rear view.</p>
Full article ">Figure 5
<p>The Sonoff Basic module with the Supla software WiFi Smart Switch DS18B20 v2.0 implemented.</p>
Full article ">Figure 6
<p>Smart home switch modules: (<b>a</b>) Shelly 1 and (<b>b</b>) Blebox Switchbox.</p>
Full article ">Figure 7
<p>Smart home socket modules: (<b>a</b>) tplink and (<b>b</b>) Gosund.</p>
Full article ">Figure 8
<p>Sonel Analysis 4 software (screenshot with the Polish interface and numbers in European nomenclature): (<b>a</b>) measurement data and (<b>b</b>) phasor chart.</p>
Full article ">Figure 9
<p>Graphs of instantaneous power consumption levels for the classic bulb, 60 Watt load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 10
<p>Graphs of instantaneous power consumption levels for the LED lamp, 4 Watt load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 11
<p>Graphs of instantaneous power consumption levels for the LED lamp, 10 Watt load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 12
<p>Graphs of instantaneous power consumption levels for series-connected classic bulb 60 Watt + LED lamp 4 Watt + LED lamp 10 Watt: (<b>a</b>) active power of smart home modules + loads and (<b>b</b>) reactive power of smart home modules + loads.</p>
Full article ">Figure 13
<p>Graphs of instantaneous power consumption levels for the 30 Watt room fan load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 14
<p>Graphs of instantaneous power consumption levels for the kitchen mixer, 76 Watt load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 15
<p>Graphs of instantaneous power consumption levels for the 630 Watt toaster load: (<b>a</b>) active power of smart home modules + load and (<b>b</b>) reactive power of smart home modules + load.</p>
Full article ">Figure 16
<p>Graphs of consumption levels of additional instantaneous active power for smart home modules without load: (<b>a</b>) with the relay in the module switched on and (<b>b</b>) with the relay in the module turned off.</p>
Full article ">Figure 17
<p>Chart of the levels of daily active energy consumption by smart home modules for the various types of loads they support, with the load constantly switched on (relays of smart home modules switched on).</p>
Full article ">Figure 18
<p>Chart of daily active energy consumption levels by smart home modules without loads, with relays on and off.</p>
Full article ">Figure 19
<p>Power factor value for smart home modules with various loads considering near and far distances from the Wi-Fi access point.</p>
Full article ">
19 pages, 1182 KiB  
Article
Force-Based Self-Organizing MANET/FANET with a UAV Swarm
by Fabrice Saffre, Hanno Hildmann and Antti Anttonen
Future Internet 2023, 15(9), 315; https://doi.org/10.3390/fi15090315 - 19 Sep 2023
Cited by 6 | Viewed by 2141
Abstract
This paper introduces a novel distributed algorithm designed to optimize the deployment of access points within Mobile Ad Hoc Networks (MANETs) for better service quality in infrastructure-less environments. The algorithm operates based on local, independent execution by each network node, thus ensuring a [...] Read more.
This paper introduces a novel distributed algorithm designed to optimize the deployment of access points within Mobile Ad Hoc Networks (MANETs) for better service quality in infrastructure-less environments. The algorithm operates based on local, independent execution by each network node, thus ensuring a high degree of scalability and adaptability to changing network conditions. The primary focus is to match the spatial distribution of access points with the distribution of client devices while maintaining strong connectivity to the network root. Using autonomous decision-making and choreographed path-planning, this algorithm bridges the gap between demand-responsive network service provision and the maintenance of crucial network connectivity links. The assessment of the performance of this approach is motivated by using numerical results generated by simulations. Full article
Show Figures

Figure 1

Figure 1
<p>Four screenshots showing a deployed MANET servicing all the clients (the stress is indicated by color, with the continuous spectrum ranging from bright green (where the Link Quality Metric LQM→1), over yellow (for a LQM value of ≈0.5) to red (when LQM→0)). This sequence of screenshots shows the response of a single swarm to successive relocations of the clients.</p>
Full article ">Figure 2
<p>Evolution of the <math display="inline"><semantics> <msub> <mi mathvariant="italic">LQM</mi> <mrow> <mi>root</mi> <mo>,</mo> <mi>c</mi> </mrow> </msub> </semantics></math> variable for all clients <span class="html-italic">c</span> (cf. Equation (<a href="#FD2-futureinternet-15-00315" class="html-disp-formula">2</a>)). A total of 100 independent simulation runs (500 clients, divided into eight dynamic clusters). (<b>a</b>) Clients relocate every 800 time units. (<b>b</b>) Clients relocate every 400. The average is an “average of averages”. The standard deviation is the average of the 100 SD values (1 data-point per simulation run and per time unit). The lowest is the average of the 100 worst LQM values experienced by any client <span class="html-italic">c</span> (idem).</p>
Full article ">Figure 3
<p>Evolution of swarm size (simultaneously airborne drones). Reported are the averages over 100 independent simulation runs. NB: all parameters, including the total number of clients and drone fleet size, are identical in both scenarios. Only the frequency of relocation (the interval between the mobile clients moving to another location) varies (800 vs. 400 time units interval).</p>
Full article ">
28 pages, 10613 KiB  
Article
Analysis of Program Representations Based on Abstract Syntax Trees and Higher-Order Markov Chains for Source Code Classification Task
by Artyom V. Gorchakov, Liliya A. Demidova and Peter N. Sovietov
Future Internet 2023, 15(9), 314; https://doi.org/10.3390/fi15090314 - 18 Sep 2023
Cited by 3 | Viewed by 1947
Abstract
In this paper we consider the research and development of classifiers that are trained to predict the task solved by source code. Possible applications of such task detection algorithms include method name prediction, hardware–software partitioning, programming standard violation detection, and semantic code duplication [...] Read more.
In this paper we consider the research and development of classifiers that are trained to predict the task solved by source code. Possible applications of such task detection algorithms include method name prediction, hardware–software partitioning, programming standard violation detection, and semantic code duplication search. We provide the comparative analysis of modern approaches to source code transformation into vector-based representations that extend the variety of classification and clustering algorithms that can be used for intelligent source code analysis. These approaches include word2vec, code2vec, first-order and second-order Markov chains constructed from abstract syntax trees (AST), histograms of assembly language instruction opcodes, and histograms of AST node types. The vectors obtained with the forementioned approaches are then used to train such classification algorithms as k-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), and multilayer perceptron (MLP). The obtained results show that the use of program vectors based on first-order AST-based Markov chains with an RF-based classifier leads to the highest accuracy, precision, recall, and F1 score. Increasing the order of Markov chains considerably increases the dimensionality of a vector, without any improvements in classifier quality, so we assume that first-order Markov chains are best suitable for real world applications. Additionally, the experimental study shows that first-order AST-based Markov chains are least sensitive to the used classification algorithm. Full article
(This article belongs to the Section Big Data and Augmented Intelligence)
Show Figures

Figure 1

Figure 1
<p>The count of programs grouped by task type.</p>
Full article ">Figure 2
<p>Program embedding into vector space based on the CBOW word2vec model.</p>
Full article ">Figure 3
<p>Program embedding into vector space based on code2vec and PathMiner.</p>
Full article ">Figure 4
<p>(<b>a</b>) The object model of a sample AST constructed using the <b>ast</b> module [<a href="#B41-futureinternet-15-00314" class="html-bibr">41</a>]. (<b>b</b>) The graphical representation of a sample AST obtained with the graphviz library [<a href="#B42-futureinternet-15-00314" class="html-bibr">42</a>].</p>
Full article ">Figure 5
<p>The result of the application of Algorithm 1 to the AST shown in <a href="#futureinternet-15-00314-f004" class="html-fig">Figure 4</a>b.</p>
Full article ">Figure 6
<p>Markov chains of different order constructed from a sample AST shown in <a href="#futureinternet-15-00314-f004" class="html-fig">Figure 4</a>a,b: (<b>a</b>) first-order AST-based Markov chain; (<b>b</b>) second-order AST-based Markov chain.</p>
Full article ">Figure 7
<p>Program embedding into vector space based on <span class="html-italic">n</span>-th order Markov chains.</p>
Full article ">Figure 8
<p>(<b>a</b>) A simple Python program; (<b>b</b>) the CPython bytecode for the simple Python program.</p>
Full article ">Figure 9
<p>Histograms of CPython assembly language instructions obtained using Algorithm 3 for programs listed in the first row of <a href="#futureinternet-15-00314-t002" class="html-table">Table 2</a>. The programs use different algorithms to implement a recurrent formula [<a href="#B19-futureinternet-15-00314" class="html-bibr">19</a>]; the dark blue color highlights overlapping histograms.</p>
Full article ">Figure 10
<p>Program embedding into vector space based on opcode histograms.</p>
Full article ">Figure 11
<p>Histograms of AST node types obtained using Algorithm 4 for programs from the first row of <a href="#futureinternet-15-00314-t002" class="html-table">Table 2</a>. The programs use different approaches to implementing a recurrent formula [<a href="#B19-futureinternet-15-00314" class="html-bibr">19</a>]; the dark blue color highlights overlapping histograms.</p>
Full article ">Figure 12
<p>Program embedding into vector space based on AST node type histograms.</p>
Full article ">Figure 13
<p><span class="html-italic">k</span>-fold cross-validation for evaluating classifier performance.</p>
Full article ">Figure 14
<p>The source code classification framework used in the experiments.</p>
Full article ">Figure 15
<p>The influence of dataset size and program embedding on the KNN classifier quality.</p>
Full article ">Figure 16
<p>The influence of dataset size and program embedding on the SVM classifier quality.</p>
Full article ">Figure 17
<p>The influence of dataset size and program embedding on the RF classifier quality.</p>
Full article ">Figure 18
<p>The influence of dataset size and program embedding on the MLP classifier quality.</p>
Full article ">Figure 19
<p>Classification quality for different pairs of a program embedding and a classifier for a small-sized dataset of source codes containing 100 programs from the original [<a href="#B19-futureinternet-15-00314" class="html-bibr">19</a>] dataset.</p>
Full article ">Figure 20
<p>Classification quality for different pairs of a program embedding and a classifier for a small-sized dataset of source codes containing 1000 programs from the original [<a href="#B19-futureinternet-15-00314" class="html-bibr">19</a>] dataset.</p>
Full article ">Figure 21
<p>The comparison of sensitivity to the used classifier of the considered program embeddings on a small dataset with 100 programs based on the data from <a href="#futureinternet-15-00314-t004" class="html-table">Table 4</a> (lower is better).</p>
Full article ">Figure 22
<p>The comparison of sensitivity to the used classifier of the considered program embeddings on a dataset with 1000 programs based on the data from <a href="#futureinternet-15-00314-t005" class="html-table">Table 5</a> (lower is better).</p>
Full article ">
23 pages, 1006 KiB  
Article
Proof of Flow: A Design Pattern for the Green Energy Market
by Valerio Mandarino, Giuseppe Pappalardo and Emiliano Tramontana
Future Internet 2023, 15(9), 313; https://doi.org/10.3390/fi15090313 - 17 Sep 2023
Cited by 1 | Viewed by 1540
Abstract
The increased penetration of Distributed Energy Resources (DERs) in electricity markets has given rise to a new category of energy players, called Aggregators, whose role is to ensure fair remuneration for energy supplied by DERs, and support the smooth feeding of the intermittent [...] Read more.
The increased penetration of Distributed Energy Resources (DERs) in electricity markets has given rise to a new category of energy players, called Aggregators, whose role is to ensure fair remuneration for energy supplied by DERs, and support the smooth feeding of the intermittent energy produced into the distribution network. This paper presents a software solution, described as a design pattern, that governs the interaction between an Aggregator and DERs, leveraging blockchain technology to achieve a higher degree of decentralization, data integrity and security, through a properly designed, blockchain-based, smart contract. Thus, the proposed solution reduces the reliance on intermediaries acting as authorities, while affording transparency, efficiency and trust to the energy exchange process. Thanks to the underlying blockchain properties, generated events are easily observable and cannot be forged or altered. However, blockchain technology has inherent drawbacks, i.e., mainly the cost of storage and execution, hence our solution provides additional strategies for limiting blockchain usage, without undermining its strengths. Moreover, the design of our smart contract takes care of orchestrating the players, and copes with their potential mutual disagreements, which could arise from different measures of energy, providing an automatic decision process to resolve such disputes. The overall approach results in lower fees for running smart contacts supporting energy players and in a greater degree of fairness assurance. Full article
(This article belongs to the Special Issue Blockchain and Web 3.0: Applications, Challenges and Future Trends)
Show Figures

Figure 1

Figure 1
<p>Treasury Manager pattern after the Service Contract has been updated.</p>
Full article ">Figure 2
<p>The structure of the Proof of Flow design pattern.</p>
Full article ">
2 pages, 168 KiB  
Editorial
Editorial for the Special Issue on “Software Engineering and Data Science”, Volume II
by Davide Tosi
Future Internet 2023, 15(9), 312; https://doi.org/10.3390/fi15090312 - 16 Sep 2023
Viewed by 1172
Abstract
The Special Issue “Software Engineering and Data Science, Volume II” is the natural continuation of its greatly successful predecessor, Volume I [...] Full article
(This article belongs to the Special Issue Software Engineering and Data Science II)
27 pages, 600 KiB  
Review
Automatic Short Text Summarization Techniques in Social Media Platforms
by Fahd A. Ghanem, M. C. Padma and Ramez Alkhatib
Future Internet 2023, 15(9), 311; https://doi.org/10.3390/fi15090311 - 13 Sep 2023
Cited by 6 | Viewed by 3417
Abstract
The rapid expansion of social media platforms has resulted in an unprecedented surge of short text content being generated on a daily basis. Extracting valuable insights and patterns from this vast volume of textual data necessitates specialized techniques that can effectively condense information [...] Read more.
The rapid expansion of social media platforms has resulted in an unprecedented surge of short text content being generated on a daily basis. Extracting valuable insights and patterns from this vast volume of textual data necessitates specialized techniques that can effectively condense information while preserving its core essence. In response to this challenge, automatic short text summarization (ASTS) techniques have emerged as a compelling solution, gaining significant importance in their development. This paper delves into the domain of summarizing short text on social media, exploring various types of short text and the associated challenges they present. It also investigates the approaches employed to generate concise and meaningful summaries. By providing a survey of the latest methods and potential avenues for future research, this paper contributes to the advancement of ASTS in the ever-evolving landscape of social media communication. Full article
Show Figures

Figure 1

Figure 1
<p>The structure of an automatic short text summarization system.</p>
Full article ">
25 pages, 813 KiB  
Review
Exploring Homomorphic Encryption and Differential Privacy Techniques towards Secure Federated Learning Paradigm
by Rezak Aziz, Soumya Banerjee, Samia Bouzefrane and Thinh Le Vinh
Future Internet 2023, 15(9), 310; https://doi.org/10.3390/fi15090310 - 13 Sep 2023
Cited by 20 | Viewed by 6498
Abstract
The trend of the next generation of the internet has already been scrutinized by top analytics enterprises. According to Gartner investigations, it is predicted that, by 2024, 75% of the global population will have their personal data covered under privacy regulations. This alarming [...] Read more.
The trend of the next generation of the internet has already been scrutinized by top analytics enterprises. According to Gartner investigations, it is predicted that, by 2024, 75% of the global population will have their personal data covered under privacy regulations. This alarming statistic necessitates the orchestration of several security components to address the enormous challenges posed by federated and distributed learning environments. Federated learning (FL) is a promising technique that allows multiple parties to collaboratively train a model without sharing their data. However, even though FL is seen as a privacy-preserving distributed machine learning method, recent works have demonstrated that FL is vulnerable to some privacy attacks. Homomorphic encryption (HE) and differential privacy (DP) are two promising techniques that can be used to address these privacy concerns. HE allows secure computations on encrypted data, while DP provides strong privacy guarantees by adding noise to the data. This paper first presents consistent attacks on privacy in federated learning and then provides an overview of HE and DP techniques for secure federated learning in next-generation internet applications. It discusses the strengths and weaknesses of these techniques in different settings as described in the literature, with a particular focus on the trade-off between privacy and convergence, as well as the computation overheads involved. The objective of this paper is to analyze the challenges associated with each technique and identify potential opportunities and solutions for designing a more robust, privacy-preserving federated learning framework. Full article
Show Figures

Figure 1

Figure 1
<p>Attacks on the federated learning process [<a href="#B32-futureinternet-15-00310" class="html-bibr">32</a>].</p>
Full article ">Figure 2
<p>Client process in secure federated learning.</p>
Full article ">Figure 3
<p>Server process in secure federated learning.</p>
Full article ">
17 pages, 813 KiB  
Article
On Evaluating IoT Data Trust via Machine Learning
by Timothy Tadj, Reza Arablouei and Volkan Dedeoglu
Future Internet 2023, 15(9), 309; https://doi.org/10.3390/fi15090309 - 12 Sep 2023
Cited by 2 | Viewed by 1846
Abstract
Data trust in IoT is crucial for safeguarding privacy, security, reliable decision-making, user acceptance, and complying with regulations. Various approaches based on supervised or unsupervised machine learning (ML) have recently been proposed for evaluating IoT data trust. However, assessing their real-world efficacy is [...] Read more.
Data trust in IoT is crucial for safeguarding privacy, security, reliable decision-making, user acceptance, and complying with regulations. Various approaches based on supervised or unsupervised machine learning (ML) have recently been proposed for evaluating IoT data trust. However, assessing their real-world efficacy is hard mainly due to the lack of related publicly available datasets that can be used for benchmarking. Since obtaining such datasets is challenging, we propose a data synthesis method, called random walk infilling (RWI), to augment IoT time-series datasets by synthesizing untrustworthy data from existing trustworthy data. Thus, RWI enables us to create labeled datasets that can be used to develop and validate ML models for IoT data trust evaluation. We also extract new features from IoT time-series sensor data that effectively capture its autocorrelation as well as its cross-correlation with the data of the neighboring (peer) sensors. These features can be used to learn ML models for recognizing the trustworthiness of IoT sensor data. Equipped with our synthesized ground-truth-labeled datasets and informative correlation-based features, we conduct extensive experiments to critically examine various approaches to evaluating IoT data trust via ML. The results reveal that commonly used ML-based approaches to IoT data trust evaluation, which rely on unsupervised cluster analysis to assign trust labels to unlabeled data, perform poorly. This poor performance is due to the underlying assumption that clustering provides reliable labels for data trust, which is found to be untenable. The results also indicate that ML models, when trained on datasets augmented via RWI and using the proposed features, generalize well to unseen data and surpass existing related approaches. Moreover, we observe that a semi-supervised ML approach that requires only about 10% of the data labeled offers competitive performance while being practically more appealing compared to the fully supervised approaches. The related Python code and data are available online. Full article
(This article belongs to the Special Issue Information and Future Internet Security, Trust and Privacy II)
Show Figures

Figure 1

Figure 1
<p>Intel Lab layout with sensor locations.</p>
Full article ">Figure 2
<p>An example actual trustworthy data instance (time-series data collected by a sensor in a 24-h period) and a corresponding untrustworthy data instance synthesized via RWI.</p>
Full article ">Figure 3
<p>An example actual trustworthy data instance (time-series data collected by a sensor in a 24-h period) and a corresponding untrustworthy data instance synthesized via Drift.</p>
Full article ">Figure 4
<p>Visualization of the correlation feature space of the Intel Lab dataset, which is augmented by synthesized RWI and Drift untrustworthy data, using the UMAP algorithm. The <span class="html-italic">x</span> and <span class="html-italic">y</span> axes encode the values in the two embedding dimensions computed by UMAP for the data instances.</p>
Full article ">Figure 5
<p>Data trust classification accuracy of the considered ML-based approaches trained and evaluated on both correlation and DST feature sets that are extracted from the Intel Lab dataset containing untrustworthy data synthesized using the RWI or Drift methods. The results are averaged over ten independent random realizations of data synthesis.</p>
Full article ">Figure 6
<p>Data trust classification accuracy of the considered ML-based approaches when evaluated on unseen data, i.e., trained on a dataset containing untrustworthy data that are synthesized using either RWI or the Drift method and tested on a dataset containing untrustworthy data synthesized using the other method.</p>
Full article ">Figure 7
<p>UMAP visualization of the decision boundaries learned via <span class="html-italic">k</span>-means clustering (<b>left</b>) and an SVM classifier trained on the data labeled via <span class="html-italic">k</span>-means clustering (<b>right</b>) when using a dataset augmented via RWI and the correlation features.</p>
Full article ">Figure 8
<p>UMAP visualization of the trustworthy and untrustworthy decision regions in the correlation feature space determined by the MLP algorithm trained on a dataset containing untrustworthy data synthesized via the Drift method.</p>
Full article ">Figure 9
<p>UMAP visualization of the unique trustworthy and untrustworthy decision regions in the correlation feature space determined by the MLP algorithm trained on datasets augmented by untrustworthy data generated via the RWI or Drift method.</p>
Full article ">
22 pages, 8329 KiB  
Article
Prototyping a Hyperledger Fabric-Based Security Architecture for IoMT-Based Health Monitoring Systems
by Filippos Pelekoudas-Oikonomou, José C. Ribeiro, Georgios Mantas, Georgia Sakellari and Jonathan Gonzalez
Future Internet 2023, 15(9), 308; https://doi.org/10.3390/fi15090308 - 11 Sep 2023
Cited by 5 | Viewed by 2136
Abstract
The Internet of Medical Things (IoMT) has risen significantly in recent years and has provided better quality of life by enabling IoMT-based health monitoring systems. Despite that fact, innovative security mechanisms are required to meet the security concerns of such systems effectively and [...] Read more.
The Internet of Medical Things (IoMT) has risen significantly in recent years and has provided better quality of life by enabling IoMT-based health monitoring systems. Despite that fact, innovative security mechanisms are required to meet the security concerns of such systems effectively and efficiently. Additionally, the industry and the research community have anticipated that blockchain technology will be a disruptive technology that will be able to be integrated into innovative security solutions for IoMT networks since it has the potential to play a big role in: (a) enabling secure data transmission, (b) ensuring IoMT device security, and (c) enabling tamper-proof data storage. Therefore, the purpose of this research work is to design a novel lightweight blockchain-based security architecture for IoMT-based health monitoring systems leveraging the features of the Hyperledger Fabric (HF) Platform, its utilities. and its lightweight blockchain nature in order to: (i) ensure entity authentication, (ii) ensure data confidentiality, and (iii) enable a more energy-efficient blockchain-based security architecture for IoMT-based health monitoring systems while considering the limited resources of IoMT gateways. While security mechanisms for IoT utilizing HF do exist, to the best of our knowledge there is no specific HF-based architecture for IoMT-based health monitoring systems. Full article
(This article belongs to the Special Issue The Future Internet of Medical Things II)
Show Figures

Figure 1

Figure 1
<p>Transaction Workflow in HF.</p>
Full article ">Figure 2
<p>IoMT-based health monitoring system.</p>
Full article ">Figure 3
<p>Overview of the HF-based security architecture for IoMT-based health monitoring systems.</p>
Full article ">Figure 4
<p>The deployed HF network.</p>
Full article ">Figure 5
<p>Creating CAs (step 1).</p>
Full article ">Figure 6
<p>Enrolling the CA admin (step 2).</p>
Full article ">Figure 7
<p>Enrolling the org1 admin (step 2).</p>
Full article ">Figure 8
<p>Creating the genesis block (step 4).</p>
Full article ">Figure 9
<p>Channel creation (step 5).</p>
Full article ">Figure 10
<p>Joining <span class="html-italic">peer0.org1</span> into the channel (step 6).</p>
Full article ">Figure 11
<p>Chaincode installation process (step 7).</p>
Full article ">Figure 12
<p>Chaincode approved and committed on channel (step 7).</p>
Full article ">Figure 13
<p>Docker containers for the performance evaluation.</p>
Full article ">Figure 14
<p>Asset template on the <span class="html-italic">create_asset</span> function.</p>
Full article ">Figure 15
<p>Network Latency (s), (<b>a</b>) per “asset” (<b>b</b>) per “worker”, for <span class="html-italic">create_asset</span>.</p>
Full article ">Figure 16
<p>Network Latency (s), (<b>a</b>) per “asset” (<b>b</b>) per “worker”, for <span class="html-italic">read_asset</span>.</p>
Full article ">Figure 17
<p>Send rate (TPS) and throughput (TPS), (<b>a</b>) per “asset” (<b>b</b>) per “worker”, for <span class="html-italic">create_asset</span>.</p>
Full article ">Figure 18
<p>Send rate (TPS) and throughput (TPS), (<b>a</b>) per “asset” (<b>b</b>) per “worker”, for <span class="html-italic">read_asset</span>.</p>
Full article ">Figure 19
<p>Peer memory usage (MB), (<b>a</b>) per “asset” (<b>b</b>) per “worker”.</p>
Full article ">
27 pages, 2600 KiB  
Article
FL-LoRaMAC: A Novel Framework for Enabling On-Device Learning for LoRa-Based IoT Applications
by Shobhit Aggarwal and Asis Nasipuri
Future Internet 2023, 15(9), 307; https://doi.org/10.3390/fi15090307 - 10 Sep 2023
Cited by 3 | Viewed by 2778
Abstract
The Internet of Things (IoT) enables us to gain access to a wide range of data from the physical world that can be analyzed for deriving critical state information. In this regard, machine learning (ML) is a valuable tool that can be used [...] Read more.
The Internet of Things (IoT) enables us to gain access to a wide range of data from the physical world that can be analyzed for deriving critical state information. In this regard, machine learning (ML) is a valuable tool that can be used to develop models based on observed physical data, leading to efficient analytical decisions, including anomaly detection. In this work, we address some key challenges for applying ML in IoT applications that include maintaining privacy considerations of user data that are needed for developing ML models and minimizing the communication cost for transmitting the data over the IoT network. We consider a representative application of the anomaly detection of ECG signals that are obtained from a set of low-cost wearable sensors and transmitted to a central server using LoRaWAN, which is a popular and emerging low-power wide-area network (LPWAN) technology. We present a novel framework utilizing federated learning (FL) to preserve data privacy and appropriate features for uplink and downlink communications between the end devices and the gateway to optimize the communication cost. Performance results obtained from computer simulations demonstrate that the proposed framework leads to a 98% reduction in the volume of data that is required to achieve the same level of performance as in traditional centralized ML. Full article
(This article belongs to the Special Issue Applications of Wireless Sensor Networks and Internet of Things)
Show Figures

Figure 1

Figure 1
<p>LoRaWAN architecture.</p>
Full article ">Figure 2
<p>Illustration of a typical machine learning model where all data are transmitted from IoT sensors to a central server for model development and testing.</p>
Full article ">Figure 3
<p>Illustration of a federated learning model, where the initial model is sent to end devices for training using their respective local data. The locally trained models are transmitted to the FL server for aggregation and the process is repeated over multiple rounds of local training and model aggregation.</p>
Full article ">Figure 4
<p>LoRaWAN classes of operation.</p>
Full article ">Figure 5
<p>Illustration of the elongated preamble approach for downlink communication.</p>
Full article ">Figure 6
<p>Operations of FL-LoRaMAC at the end device.</p>
Full article ">Figure 7
<p>Operations of FL-LoRaMAC at the network server.</p>
Full article ">Figure 8
<p>Illustration of the procedure for optimization of communication cost in FL-LoRaMAC.</p>
Full article ">Figure 9
<p>Illustration of dataset division.</p>
Full article ">Figure 10
<p>Traffic volume for machine learning (100 epochs), federated learning (3 communication rounds and 100 epochs), and federated learning with PCA (20 components, 3 communication rounds, and 100 epochs) approaches.</p>
Full article ">Figure 11
<p>Computation time for machine learning (100 epochs) and federated learning (3 communication rounds and 100 epochs) approach.</p>
Full article ">Figure 12
<p>Average performance and corresponding standard deviations with packet loss: (<b>a</b>) recall, (<b>b</b>) precision, and (<b>c</b>) accuracy with the varying loss in network.</p>
Full article ">Figure 13
<p>Performance with various levels of sparsification: (<b>a</b>) recall, (<b>b</b>) precision, and (<b>c</b>) accuracy with the varying sparsification percentages.</p>
Full article ">Figure 14
<p>PDR for high- and low-priority devices for varying total number of devices.</p>
Full article ">Figure 15
<p>Performance comparison of the trained model with QoS and legacy: (<b>a</b>) precision and (<b>b</b>) accuracy with the varying total number of devices in network.</p>
Full article ">
19 pages, 374 KiB  
Article
An Automatic Transformer from Sequential to Parallel Java Code
by Alessandro Midolo and Emiliano Tramontana
Future Internet 2023, 15(9), 306; https://doi.org/10.3390/fi15090306 - 8 Sep 2023
Cited by 2 | Viewed by 1507
Abstract
Sequential programs can benefit from parallel execution to improve their performance. When developing a parallel application, several techniques are employed to achieve the desired behavior: identifying parts that can run in parallel, synchronizing access to shared data, tuning performance, etc. Admittedly, manually transforming [...] Read more.
Sequential programs can benefit from parallel execution to improve their performance. When developing a parallel application, several techniques are employed to achieve the desired behavior: identifying parts that can run in parallel, synchronizing access to shared data, tuning performance, etc. Admittedly, manually transforming a sequential application to make it parallel can be tedious due to the large number of lines of code to inspect, the possibility of errors arising from inaccurate data dependence analysis leading to unpredictable behavior, and inefficiencies when the workload between parallel threads is unbalanced. This paper proposes an automatic approach that analyzes Java source code to identify method calls that are suitable for parallel execution and transforms them so that they run in another thread. The approach is based on data dependence and control dependence analyses to determine the execution flow and data accessed. Based on the proposed method, a tool has been developed to enhance applications by incorporating parallelism, i.e., transforming suitable method calls to execute on parallel threads, and synchronizing data access where needed. The developed tool has been extensively tested to verify the accuracy of its analysis in finding parallel execution opportunities, the correctness of the source code alterations, and the resultant performance gain. Full article
(This article belongs to the Section Smart System Infrastructure and Applications)
Show Figures

Figure 1

Figure 1
<p>Control flow graph of the code shown in Listing 4.</p>
Full article ">Figure 2
<p>Control flow graphs of the code with conditions shown in Listing 5.</p>
Full article ">Figure 3
<p>CFG extracted from method getUserForAuthor() shown in Listing 7.</p>
Full article ">
17 pages, 1136 KiB  
Article
Entering the Metaverse from the JVM: The State of the Art, Challenges, and Research Areas of JVM-Based Web 3.0 Tools and Libraries
by Vlad Bucur and Liviu-Cristian Miclea
Future Internet 2023, 15(9), 305; https://doi.org/10.3390/fi15090305 - 7 Sep 2023
Cited by 3 | Viewed by 1512
Abstract
Web 3.0 is the basis on which the proposed metaverse, a seamless virtual world enabled by computers and interconnected devices, hopes to interact with its users, but beyond the high-level project overview of what Web 3.0 applications try to achieve, the implementation is [...] Read more.
Web 3.0 is the basis on which the proposed metaverse, a seamless virtual world enabled by computers and interconnected devices, hopes to interact with its users, but beyond the high-level project overview of what Web 3.0 applications try to achieve, the implementation is still down to low-level coding details. This article aims to analyze the low-level implementations of key components of Web 3.0 using a variety of frameworks and tools as well as several JVM-based languages. This paper breaks down the low-level implementation of smart contracts and semantic web principles using three frameworks, Corda and Ethereum for smart contracts and Jeda for semantic web, using both Scala and Java as implementing languages all while highlighting differences and similarities between the frameworks used. Full article
Show Figures

Figure 1

Figure 1
<p>Transaction and witness hashing [<a href="#B8-futureinternet-15-00305" class="html-bibr">8</a>].</p>
Full article ">Figure 2
<p>Basic classes and interfaces involved in a Corda smart contract implementation.</p>
Full article ">Figure 3
<p>Basic classes involved in creating an NFT with Web3j.</p>
Full article ">Figure 4
<p>Basic definition of an ontology in Apache Jena.</p>
Full article ">
21 pages, 1310 KiB  
Article
Hospital Readmission and Length-of-Stay Prediction Using an Optimized Hybrid Deep Model
by Alireza Tavakolian, Alireza Rezaee, Farshid Hajati and Shahadat Uddin
Future Internet 2023, 15(9), 304; https://doi.org/10.3390/fi15090304 - 6 Sep 2023
Cited by 4 | Viewed by 2833
Abstract
Hospital readmission and length-of-stay predictions provide information on how to manage hospital bed capacity and the number of required staff, especially during pandemics. We present a hybrid deep model called the Genetic Algorithm-Optimized Convolutional Neural Network (GAOCNN), with a unique preprocessing method to [...] Read more.
Hospital readmission and length-of-stay predictions provide information on how to manage hospital bed capacity and the number of required staff, especially during pandemics. We present a hybrid deep model called the Genetic Algorithm-Optimized Convolutional Neural Network (GAOCNN), with a unique preprocessing method to predict hospital readmission and the length of stay required for patients of various conditions. GAOCNN uses one-dimensional convolutional layers to predict hospital readmission and the length of stay. The parameters of the layers are optimized via a genetic algorithm. To show the performance of the proposed model in patients with various conditions, we evaluate the model under three healthcare datasets: the Diabetes 130-US hospitals dataset, the COVID-19 dataset, and the MIMIC-III dataset. The diabetes 130-US hospitals dataset has information on both readmission and the length of stay, while the COVID-19 and MIMIC-III datasets just include information on the length of stay. Experimental results show that the proposed model’s accuracy for hospital readmission was 97.2% for diabetic patients. Furthermore, the accuracy of the length-of-stay prediction was 89%, 99.4%, and 94.1% for the diabetic, COVID-19, and ICU patients, respectively. These results confirm the superiority of the proposed model compared to existing methods. Our findings offer a platform for managing the healthcare funds and resources for patients with various diseases. Full article
(This article belongs to the Special Issue Internet of Things (IoT) for Smart Living and Public Health)
Show Figures

Figure 1

Figure 1
<p>Structure of the proposed model.</p>
Full article ">Figure 2
<p>The flowchart of the proposed GA.</p>
Full article ">Figure 3
<p>Distributions of the readmission in diabetic patients.</p>
Full article ">Figure 4
<p>LOS distribution: (<b>a</b>) Diabetes, (<b>b</b>) COVID-19, (<b>c</b>) MIMIC-III.</p>
Full article ">Figure 5
<p>Relation between age and LOS based different genders for instances with the highest LOS distributions: (<b>a</b>) Diabetes, (<b>b</b>) COVID-19, (<b>c</b>) MIMIC-III.</p>
Full article ">Figure 6
<p>Distributions of the LOS in the original and balanced datasets: (<b>a</b>) Diabetes, (<b>b</b>) COVID-19, (<b>c</b>) MIMIC-III.</p>
Full article ">Figure 6 Cont.
<p>Distributions of the LOS in the original and balanced datasets: (<b>a</b>) Diabetes, (<b>b</b>) COVID-19, (<b>c</b>) MIMIC-III.</p>
Full article ">Figure 7
<p>The normalized confusion matrix; (<b>a</b>) readmission prediction in diabetic patients, (<b>b</b>) LOS prediction in diabetic patients.</p>
Full article ">Figure 8
<p>The normalized confusion matrix for the LOS prediction in COVID-19 patients.</p>
Full article ">Figure 9
<p>The normalized confusion matrix for the LOS prediction in ICU patients.</p>
Full article ">Figure 10
<p>Best selected features based on accuracy with wrapper feature selection.</p>
Full article ">
19 pages, 12249 KiB  
Article
Internet of Robotic Things (IoRT) and Metaheuristic Optimization Techniques Applied for Wheel-Legged Robot
by Mateusz Malarczyk, Grzegorz Kaczmarczyk, Jaroslaw Szrek and Marcin Kaminski
Future Internet 2023, 15(9), 303; https://doi.org/10.3390/fi15090303 - 6 Sep 2023
Viewed by 1682
Abstract
This paper presents the operation of a remotely controlled, wheel-legged robot. The developed Wi-Fi connection framework is established on a popular ARM microcontroller board. The implementation provides a low-cost solution that is in congruence with the newest industrial standards. Additionally, the problem of [...] Read more.
This paper presents the operation of a remotely controlled, wheel-legged robot. The developed Wi-Fi connection framework is established on a popular ARM microcontroller board. The implementation provides a low-cost solution that is in congruence with the newest industrial standards. Additionally, the problem of limb structure and motor speed control is solved. The design process of the mechanical structure is enhanced by a nature-inspired metaheuristic optimization algorithm. An FOC-based BLDC motor speed control strategy is selected to guarantee dynamic operation of the drive. The paper provides both the theoretical considerations and the obtained prototype experimental results. Full article
(This article belongs to the Special Issue Internet of Things (IoT) for Smart Living and Public Health)
Show Figures

Figure 1

Figure 1
<p>The topology of microcontroller system used for remote control of wheel-legged robot.</p>
Full article ">Figure 2
<p>The principles of field-oriented control.</p>
Full article ">Figure 3
<p>Field-oriented control structure.</p>
Full article ">Figure 4
<p>The position control structure applied for the chassis.</p>
Full article ">Figure 5
<p>Required degrees of freedom.</p>
Full article ">Figure 6
<p>Four-bar linkage suspension structure of analyzed robot.</p>
Full article ">Figure 7
<p>Flowchart of the flower pollination algorithm—the pollination mechanism is presented in yellow blocks.</p>
Full article ">Figure 8
<p>The developed prototype of the LegVan3 robot, (<b>a</b>)—the down position (regular driving position), and (<b>b</b>)—the elevated position.</p>
Full article ">Figure 9
<p>Changes of cost function value during calculations of the FPA.</p>
Full article ">Figure 10
<p>Speed transient of the BLDC motor implemented in the robot.</p>
Full article ">Figure 11
<p>Screenshot of developed website.</p>
Full article ">Figure 12
<p>Data processing in the ESP32.</p>
Full article ">Figure 13
<p>The kinematics analysis of robot (control part).</p>
Full article ">
19 pages, 9235 KiB  
Article
A Hybrid Neural Ordinary Differential Equation Based Digital Twin Modeling and Online Diagnosis for an Industrial Cooling Fan
by Chao-Chung Peng and Yi-Ho Chen
Future Internet 2023, 15(9), 302; https://doi.org/10.3390/fi15090302 - 4 Sep 2023
Cited by 3 | Viewed by 1866
Abstract
Digital twins can reflect the dynamical behavior of the identified system, enabling self-diagnosis and prediction in the digital world to optimize the intelligent manufacturing process. One of the key benefits of digital twins is the ability to provide real-time data analysis during operation, [...] Read more.
Digital twins can reflect the dynamical behavior of the identified system, enabling self-diagnosis and prediction in the digital world to optimize the intelligent manufacturing process. One of the key benefits of digital twins is the ability to provide real-time data analysis during operation, which can monitor the condition of the system and prognose the failure. This allows manufacturers to resolve the problem before it happens. However, most digital twins are constructed using discrete-time models, which are not able to describe the dynamics of the system across different sampling frequencies. In addition, the high computational complexity due to significant memory storage and large model sizes makes digital twins challenging for online diagnosis. To overcome these issues, this paper proposes a novel structure for creating the digital twins of cooling fan systems by combining with neural ordinary differential equations and physical dynamical differential equations. Evaluated using the simulation data, the proposed structure not only shows accurate modeling results compared to other digital twins methods but also requires fewer parameters and smaller model sizes. The proposed approach has also been demonstrated using experimental data and is robust in terms of measurement noise, and it has proven to be an effective solution for online diagnosis in the intelligent manufacturing process. Full article
(This article belongs to the Special Issue Digital Twins in Intelligent Manufacturing)
Show Figures

Figure 1

Figure 1
<p>Illustration of a cooling fan driving system [<a href="#B13-futureinternet-15-00302" class="html-bibr">13</a>].</p>
Full article ">Figure 2
<p>The structure of the NARX model.</p>
Full article ">Figure 3
<p>Structure of the neural ODE model for the cooling fan system.</p>
Full article ">Figure 4
<p>The training process of the hybrid neural ODE model.</p>
Full article ">Figure 5
<p>Simulation data for training. (<b>a</b>) Excitation input torque; (<b>b</b>) the response of fan speed.</p>
Full article ">Figure 6
<p>Simulation data for testing. (<b>a</b>) Excitation input torque; (<b>b</b>) the response of fan speed.</p>
Full article ">Figure 7
<p>The fitting performance of different fan dynamics models in [<a href="#B25-futureinternet-15-00302" class="html-bibr">25</a>].</p>
Full article ">Figure 8
<p>Physically based model fitting performance in high-speed conditions.</p>
Full article ">Figure 9
<p>Training fitting result for simulation data. (<b>a</b>) Fan speed output. (<b>b</b>) Modeling error.</p>
Full article ">Figure 10
<p>Testing fitting results for simulation data. (<b>a</b>) Fan speed output. (<b>b</b>) Modeling error.</p>
Full article ">Figure 11
<p>Developed fan tray system used for the experiment [<a href="#B13-futureinternet-15-00302" class="html-bibr">13</a>]. (<b>a</b>) Front view of the fan system; (<b>b</b>) microprocessor for input excitation.</p>
Full article ">Figure 12
<p>Experimental data for training. (<b>a</b>) Excitation input voltage. (<b>b</b>) The response of fan speed.</p>
Full article ">Figure 13
<p>Experimental data for testing. (<b>a</b>) Excitation input voltage. (<b>b</b>) The response of fan speed.</p>
Full article ">Figure 14
<p>Training fitting result on experimental data. (<b>a</b>) Fan speed output. (<b>b</b>) Modeling error.</p>
Full article ">Figure 15
<p>Testing fitting results on experimental data. (<b>a</b>) Fan speed output. (<b>b</b>) Modeling error.</p>
Full article ">Figure 16
<p>Testing fitting result for NARX model on experimental data. (<b>a</b>) Fan speed output. (<b>b</b>) Modeling error.</p>
Full article ">Figure 17
<p>The anomalies that occurred in the fan system. (<b>a</b>) Fan inlet covered by an object. (<b>b</b>) Disturbance of the rotor.</p>
Full article ">Figure 18
<p>Error distribution between the digital twin and measurement under healthy conditions.</p>
Full article ">Figure 19
<p>Anomaly detection result. (<b>a</b>) Covered inlet. (<b>b</b>) Disturbance on the rotor. (<b>c</b>) Manual modulation with anomalies.</p>
Full article ">Figure 19 Cont.
<p>Anomaly detection result. (<b>a</b>) Covered inlet. (<b>b</b>) Disturbance on the rotor. (<b>c</b>) Manual modulation with anomalies.</p>
Full article ">
24 pages, 1339 KiB  
Article
Wireless Energy Harvesting for Internet-of-Things Devices Using Directional Antennas
by Hsiao-Ching Chang, Hsing-Tsung Lin and Pi-Chung Wang
Future Internet 2023, 15(9), 301; https://doi.org/10.3390/fi15090301 - 3 Sep 2023
Cited by 1 | Viewed by 2695
Abstract
With the rapid development of the Internet of Things, the number of wireless devices is increasing rapidly. Because of the limited battery capacity, these devices may suffer from the issue of power depletion. Radio frequency (RF) energy harvesting technology can wirelessly charge devices [...] Read more.
With the rapid development of the Internet of Things, the number of wireless devices is increasing rapidly. Because of the limited battery capacity, these devices may suffer from the issue of power depletion. Radio frequency (RF) energy harvesting technology can wirelessly charge devices to prolong their lifespan. With the technology of beamforming, the beams generated by an antenna array can select the direction for wireless charging. Although a good charging-time schedule should be short, energy efficiency should also be considered. In this work, we propose two algorithms to optimize the time consumption for charging devices. We first present a greedy algorithm to minimize the total charging time. Then, a differential evolution (DE) algorithm is proposed to minimize the energy overflow and improve energy efficiency. The DE algorithm can also gradually increase fully charged devices. The experimental results show that both the proposed greedy and DE algorithms can find a schedule of a short charging time with the lowest energy overflow. The DE algorithm can further improve the performance of data transmission to promote the feasibility of potential wireless sensing and charging applications by reducing the number of fully charged devices at the same time. Full article
Show Figures

Figure 1

Figure 1
<p>The architecture of RF energy harvesting.</p>
Full article ">Figure 2
<p>Simplified antenna model.</p>
Full article ">Figure 3
<p>Sector model. (<b>a</b>) The First Sector (<b>b</b>) The Second Sector</p>
Full article ">Figure 4
<p>An example of the proposed greedy algorithm.</p>
Full article ">Figure 5
<p>The step of the DE algorithm.</p>
Full article ">Figure 6
<p>The representation of NP individuals.</p>
Full article ">Figure 7
<p>The convergence of our DE algorithm.</p>
Full article ">Figure 8
<p>The scaling factor of our DE algorithm.</p>
Full article ">Figure 9
<p>The hybridization probability of our DE algorithm.</p>
Full article ">Figure 10
<p>The charging time for different numbers of devices.</p>
Full article ">Figure 11
<p>The overflow energy for different numbers of devices.</p>
Full article ">Figure 12
<p>The charging time for different transmission angles.</p>
Full article ">Figure 13
<p>The overflow energy for different transmission angles.</p>
Full article ">Figure 14
<p>The charging time of different offset.</p>
Full article ">Figure 15
<p>The overflow energy of different offset.</p>
Full article ">Figure 16
<p>The topology types, where the orange node in each topology is the AP and the blues nodes are rechargeable devices.</p>
Full article ">Figure 17
<p>The charging time of different topologies.</p>
Full article ">Figure 18
<p>The overflow energy for different topologies.</p>
Full article ">Figure 19
<p>The charging time for different charging ratios.</p>
Full article ">Figure 20
<p>The overflow energy for different charging ratios.</p>
Full article ">Figure 21
<p>The number of fully charged devices in different iterations. (<b>a</b>) Ratio = 1, (<b>b</b>) Ratio = 0.8, (<b>c</b>) Ratio = 0.6.</p>
Full article ">Figure 22
<p>The data rate of fully charged devices in different iterations. (<b>a</b>) Ratio = 1, (<b>b</b>) Ratio = 0.8, (<b>c</b>) Ratio = 0.6.</p>
Full article ">Figure 23
<p>The total data rate of fully charged devices in different iterations. (<b>a</b>) Ratio = 1, (<b>b</b>) Ratio = 0.8, (<b>c</b>) Ratio = 0.6.</p>
Full article ">Figure 23 Cont.
<p>The total data rate of fully charged devices in different iterations. (<b>a</b>) Ratio = 1, (<b>b</b>) Ratio = 0.8, (<b>c</b>) Ratio = 0.6.</p>
Full article ">
18 pages, 3680 KiB  
Article
Application of ChatGPT-Based Digital Human in Animation Creation
by Chong Lan, Yongsheng Wang, Chengze Wang, Shirong Song and Zheng Gong
Future Internet 2023, 15(9), 300; https://doi.org/10.3390/fi15090300 - 2 Sep 2023
Cited by 11 | Viewed by 5672
Abstract
Traditional 3D animation creation involves a process of motion acquisition, dubbing, and mouth movement data binding for each character. To streamline animation creation, we propose combining artificial intelligence (AI) with a motion capture system. This integration aims to reduce the time, workload, and [...] Read more.
Traditional 3D animation creation involves a process of motion acquisition, dubbing, and mouth movement data binding for each character. To streamline animation creation, we propose combining artificial intelligence (AI) with a motion capture system. This integration aims to reduce the time, workload, and cost associated with animation creation. By utilizing AI and natural language processing, the characters can engage in independent learning, generating their own responses and interactions, thus moving away from the traditional method of creating digital characters with pre-defined behaviors. In this paper, we present an approach that employs a digital person’s animation environment. We utilized Unity plug-ins to drive the character’s mouth Blendshape, synchronize the character’s voice and mouth movements in Unity, and connect the digital person to an AI system. This integration enables AI-driven language interactions within animation production. Through experimentation, we evaluated the correctness of the natural language interaction of the digital human in the animated scene, the real-time synchronization of mouth movements, the potential for singularity in guiding users during digital human animation creation, and its ability to guide user interactions through its own thought process. Full article
(This article belongs to the Topic AI Chatbots: Threat or Opportunity?)
Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Flowchart of data conversion between ASR, TTS local deployment, and APIs.</p>
Full article ">Figure 2
<p>Schematic diagram of a digital human voice data-driven mouth.</p>
Full article ">Figure 3
<p>Schematic of action data creation in Unity.</p>
Full article ">Figure 4
<p>SALSA with RandomEyes matching parameter map for driving digital people.</p>
Full article ">Figure 5
<p>Schematic of Baidu Speech Recognition script and parameter adjustment for history button.</p>
Full article ">Figure 6
<p>Application system overall structure diagram.</p>
Full article ">Figure 7
<p>ChatGPT digital man operation principle flowchart.</p>
Full article ">Figure 8
<p>ChatGPT digital human interaction test.</p>
Full article ">Figure 9
<p>ChatGPT digital human–character interaction history.</p>
Full article ">
18 pages, 442 KiB  
Article
Precoding for RIS-Assisted Multi-User MIMO-DQSM Transmission Systems
by Francisco R. Castillo-Soria, J. Alberto Del Puerto-Flores, Cesar A. Azurdia-Meza, Vinoth Babu Kumaravelu, Jorge Simón and Carlos A. Gutierrez
Future Internet 2023, 15(9), 299; https://doi.org/10.3390/fi15090299 - 2 Sep 2023
Cited by 4 | Viewed by 1727
Abstract
This paper presents two precoding techniques for a reconfigurable intelligent surface (RIS)-assisted multi-user (MU) multiple-input multiple-output (MIMO) double quadrature spatial modulation (DQSM) downlink transmission system. Instead of being applied at the remote RIS, the phase shift vector is applied at the base station [...] Read more.
This paper presents two precoding techniques for a reconfigurable intelligent surface (RIS)-assisted multi-user (MU) multiple-input multiple-output (MIMO) double quadrature spatial modulation (DQSM) downlink transmission system. Instead of being applied at the remote RIS, the phase shift vector is applied at the base station (BS) by using a double precoding stage. Results show that the proposed RIS-MU-MIMO-DQSM system has gains of up to 17 dB in terms of bit error rate (BER) and a reduction in detection complexity of 51% when compared with the conventional MU-MIMO system based on quadrature amplitude modulation (QAM). Compared with a similar system based on amplify and forward (AF) relay-assisted technique, the proposed system has a gain of up to 18 dB in terms of BER under the same conditions and parameters. Full article
Show Figures

Figure 1

Figure 1
<p>RIS-Assisted MU-DQSM system model.</p>
Full article ">Figure 2
<p>DQSM modulation block.</p>
Full article ">Figure 3
<p>Users under MUI are illuminated by two or more beams.</p>
Full article ">Figure 4
<p>Strategy I (ZF-BD precoding technique).</p>
Full article ">Figure 5
<p>Strategy II (joint-BD precoder).</p>
Full article ">Figure 6
<p>BER Performance comparison for a SE of 8 bpcu/user, L = 2, 4 users, and the uncorrelated fading channel.</p>
Full article ">Figure 7
<p>BER performance comparison for a SE of 12 bpcu/user, L = 4, 8 users, and the uncorrelated fading channel.</p>
Full article ">Figure 8
<p>BER performance comparison for a SE of 8 bpcu/user, L = 2, 4 users, and a correlation factor of <math display="inline"><semantics> <mrow> <mi>ρ</mi> <mo>=</mo> <mn>0.7</mn> </mrow> </semantics></math>.</p>
Full article ">Figure 9
<p>BER performance comparison for a SE of 12 bpcu/user, L = 4, 8 users, and a correlation factor of <math display="inline"><semantics> <mrow> <mi>ρ</mi> <mo>=</mo> <mn>0.7</mn> </mrow> </semantics></math>.</p>
Full article ">
20 pages, 1231 KiB  
Article
Intelligent Unsupervised Network Traffic Classification Method Using Adversarial Training and Deep Clustering for Secure Internet of Things
by Weijie Zhang, Lanping Zhang, Xixi Zhang, Yu Wang, Pengfei Liu and Guan Gui
Future Internet 2023, 15(9), 298; https://doi.org/10.3390/fi15090298 - 1 Sep 2023
Viewed by 2064
Abstract
Network traffic classification (NTC) has attracted great attention in many applications such as secure communications, intrusion detection systems. The existing NTC methods based on supervised learning rely on sufficient labeled datasets in the training phase, but for most traffic datasets, it is difficult [...] Read more.
Network traffic classification (NTC) has attracted great attention in many applications such as secure communications, intrusion detection systems. The existing NTC methods based on supervised learning rely on sufficient labeled datasets in the training phase, but for most traffic datasets, it is difficult to obtain label information in practical applications. Although unsupervised learning does not rely on labels, its classification accuracy is not high, and the number of data classes is difficult to determine. This paper proposes an unsupervised NTC method based on adversarial training and deep clustering with improved network traffic classification (NTC) and lower computational complexity in comparison with the traditional clustering algorithms. Here, the training process does not require data labels, which greatly reduce the computational complexity of the network traffic classification through pretraining. In the pretraining stage, an autoencoder (AE) is used to reduce the dimension of features and reduce the complexity of the initial high-dimensional network traffic data features. Moreover, we employ the adversarial training model and a deep clustering structure to further optimize the extracted features. The experimental results show that our proposed method has robust performance, with a multiclassification accuracy of 92.2%, which is suitable for classification with a large number of unlabeled data in actual application scenarios. This paper only focuses on breakthroughs in the algorithm stage, and future work can be focused on the deployment and adaptation in practical environments. Full article
(This article belongs to the Special Issue Information and Future Internet Security, Trust and Privacy II)
Show Figures

Figure 1

Figure 1
<p>Flowchart of unsupervised-learning-aided NTC method.</p>
Full article ">Figure 2
<p>The detailed process of data preprocessing.</p>
Full article ">Figure 3
<p>The model framework of the proposed DC-CAAE method. The model is divided into two steps, CAAE and DC. First, the CAAE network is used to preliminarily train the feature extractor, then DC constraints are added to fine-tune the feature extractor so that the extracted features are oriented toward clustering tasks.</p>
Full article ">Figure 4
<p>The framework of the proposed CAAE method.</p>
Full article ">Figure 5
<p>The framework of the deep clustering algorithm.</p>
Full article ">Figure 6
<p>Convergence trend of overall cluster loss during training.</p>
Full article ">Figure 7
<p>t-SNE feature visualizations for different clustering algorithms. (<b>a</b>) CAE; (<b>b</b>) CVAE; (<b>c</b>) CAAE; (<b>d</b>) DC-CAE; (<b>e</b>) DC-CVAE; (<b>f</b>) DC-CAAE.</p>
Full article ">Figure 8
<p>The change in average distortion at different cluster centers.</p>
Full article ">Figure 9
<p>The change in silhouette coefficient at different cluster centers.</p>
Full article ">Figure 10
<p>Performance of the DC-CAAE model at <math display="inline"><semantics> <mrow> <mi>k</mi> <mo>=</mo> <mfenced separators="" open="{" close="}"> <mn>5</mn> <mo>,</mo> <mn>10</mn> <mo>,</mo> <mn>15</mn> <mo>,</mo> <mn>20</mn> <mo>,</mo> <mn>25</mn> </mfenced> </mrow> </semantics></math>.</p>
Full article ">Figure 11
<p>Performance of the DC-CAAE model at <math display="inline"><semantics> <mrow> <mi>k</mi> <mo>=</mo> <mfenced separators="" open="{" close="}"> <mn>18</mn> <mo>,</mo> <mn>19</mn> <mo>,</mo> <mn>20</mn> <mo>,</mo> <mn>21</mn> <mo>,</mo> <mn>22</mn> </mfenced> </mrow> </semantics></math>.</p>
Full article ">
19 pages, 2089 KiB  
Article
Explainable Lightweight Block Attention Module Framework for Network-Based IoT Attack Detection
by Furkat Safarov, Mainak Basak, Rashid Nasimov, Akmalbek Abdusalomov and Young Im Cho
Future Internet 2023, 15(9), 297; https://doi.org/10.3390/fi15090297 - 1 Sep 2023
Cited by 5 | Viewed by 1679
Abstract
In the rapidly evolving landscape of internet usage, ensuring robust cybersecurity measures has become a paramount concern across diverse fields. Among the numerous cyber threats, denial of service (DoS) and distributed denial of service (DDoS) attacks pose significant risks, as they can render [...] Read more.
In the rapidly evolving landscape of internet usage, ensuring robust cybersecurity measures has become a paramount concern across diverse fields. Among the numerous cyber threats, denial of service (DoS) and distributed denial of service (DDoS) attacks pose significant risks, as they can render websites and servers inaccessible to their intended users. Conventional intrusion detection methods encounter substantial challenges in effectively identifying and mitigating these attacks due to their widespread nature, intricate patterns, and computational complexities. However, by harnessing the power of deep learning-based techniques, our proposed dense channel-spatial attention model exhibits exceptional accuracy in detecting and classifying DoS and DDoS attacks. The successful implementation of our proposed framework addresses the challenges posed by imbalanced data and exhibits its potential for real-world applications. By leveraging the dense channel-spatial attention mechanism, our model can precisely identify and classify DoS and DDoS attacks, bolstering the cybersecurity defenses of websites and servers. The high accuracy rates achieved across different datasets reinforce the robustness of our approach, underscoring its efficacy in enhancing intrusion detection capabilities. As a result, our framework holds promise in bolstering cybersecurity measures in real-world scenarios, contributing to the ongoing efforts to safeguard against cyber threats in an increasingly interconnected digital landscape. Comparative analysis with current intrusion detection methods reveals the superior performance of our model. We achieved accuracy rates of 99.38%, 99.26%, and 99.43% for Bot-IoT, CICIDS2017, and UNSW_NB15 datasets, respectively. These remarkable results demonstrate the capability of our approach to accurately detect and classify various types of DoS and DDoS assaults. By leveraging the inherent strengths of deep learning, such as pattern recognition and feature extraction, our model effectively overcomes the limitations of traditional methods, enhancing the accuracy and efficiency of intrusion detection systems. Full article
Show Figures

Figure 1

Figure 1
<p>System pipeline.</p>
Full article ">Figure 2
<p>Demonstrate ABC algorithm mechanism.</p>
Full article ">Figure 3
<p>Proposed model architecture.</p>
Full article ">Figure 4
<p>Multi-class classification on Bot-IoT dataset.</p>
Full article ">Figure 5
<p>Accuracy comparison of the proposed approach on Bot-IoT dataset.</p>
Full article ">Figure 6
<p>Multi-class classification on the CICIDS2017 dataset.</p>
Full article ">Figure 7
<p>Multi-class classification on the UNSW_NB15 dataset.</p>
Full article ">Figure 8
<p>Ablation of proposed model with existing techniques on the UNSW_NB15 dataset.</p>
Full article ">
16 pages, 743 KiB  
Article
FREDY: Federated Resilience Enhanced with Differential Privacy
by Zacharias Anastasakis, Terpsichori-Helen Velivassaki, Artemis Voulkidis, Stavroula Bourou, Konstantinos Psychogyios, Dimitrios Skias and Theodore Zahariadis
Future Internet 2023, 15(9), 296; https://doi.org/10.3390/fi15090296 - 1 Sep 2023
Cited by 1 | Viewed by 1449
Abstract
Federated Learning is identified as a reliable technique for distributed training of ML models. Specifically, a set of dispersed nodes may collaborate through a federation in producing a jointly trained ML model without disclosing their data to each other. Each node performs local [...] Read more.
Federated Learning is identified as a reliable technique for distributed training of ML models. Specifically, a set of dispersed nodes may collaborate through a federation in producing a jointly trained ML model without disclosing their data to each other. Each node performs local model training and then shares its trained model weights with a server node, usually called Aggregator in federated learning, as it aggregates the trained weights and then sends them back to its clients for another round of local training. Despite the data protection and security that FL provides to each client, there are still well-studied attacks such as membership inference attacks that can detect potential vulnerabilities of the FL system and thus expose sensitive data. In this paper, in order to prevent this kind of attack and address private data leakage, we introduce FREDY, a differential private federated learning framework that enables knowledge transfer from private data. Particularly, our approach has a teachers–student scheme. Each teacher model is trained on sensitive, disjoint data in a federated manner, and the student model is trained on the most voted predictions of the teachers on public unlabeled data which are noisy aggregated in order to guarantee the privacy of each teacher’s sensitive data. Only the student model is publicly accessible as the teacher models contain sensitive information. We show that our proposed approach guarantees the privacy of sensitive data against model inference attacks while it combines the federated learning settings for the model training procedures. Full article
(This article belongs to the Special Issue Privacy and Security in Computing Continuum and Data-Driven Workflows)
Show Figures

Figure 1

Figure 1
<p>Structure of Federated Learning.</p>
Full article ">Figure 2
<p>Proposed FREDY architecture.</p>
Full article ">Figure 3
<p>Teacher and student model architectures.</p>
Full article ">Figure 4
<p>Test accuracy of aggregated model. (<b>a</b>) Test accuracy of aggregated model for each federated round during FL training of teachers on CIFAR10 dataset. (<b>b</b>) Test accuracy of aggregated model for each federated round during FL training of teachers on MNIST dataset.</p>
Full article ">Figure 5
<p>Student test accuracy for 9000 queries, without noise injection.</p>
Full article ">Figure 6
<p>Student test accuracy under various scenarios. (<b>a</b>) Test accuracy of the student model per epoch with 15 teacher models and <math display="inline"><semantics> <mrow> <mi>ϵ</mi> <mo>=</mo> </mrow> </semantics></math> 0.2 noise injection per query on CIFAR10 dataset. (<b>b</b>) Test accuracy of the student model per epoch for the three teacher ensembles with 9000 queries and <math display="inline"><semantics> <mrow> <mi>ϵ</mi> <mo>=</mo> </mrow> </semantics></math> 0.1 noise injection per query on CIFAR10 dataset.</p>
Full article ">Figure 7
<p>Test accuracy of the student model for three MNIST and CIFAR-10 teacher ensembles with 9000 queries and varying <math display="inline"><semantics> <mi>ϵ</mi> </semantics></math> value per query.</p>
Full article ">
22 pages, 3262 KiB  
Article
FLAME-VQA: A Fuzzy Logic-Based Model for High Frame Rate Video Quality Assessment
by Štefica Mrvelj and Marko Matulin
Future Internet 2023, 15(9), 295; https://doi.org/10.3390/fi15090295 - 1 Sep 2023
Cited by 1 | Viewed by 1399
Abstract
In the quest to optimize user experience, network, and service, providers continually seek to deliver high-quality content tailored to individual preferences. However, predicting user perception of quality remains a challenging task, given the subjective nature of human perception and the plethora of technical [...] Read more.
In the quest to optimize user experience, network, and service, providers continually seek to deliver high-quality content tailored to individual preferences. However, predicting user perception of quality remains a challenging task, given the subjective nature of human perception and the plethora of technical attributes that contribute to the overall viewing experience. Thus, we introduce a Fuzzy Logic-bAsed ModEl for Video Quality Assessment (FLAME-VQA), leveraging the LIVE-YT-HFR database containing 480 video sequences and subjective ratings of their quality from 85 test subjects. The proposed model addresses the challenges of assessing user perception by capturing the intricacies of individual preferences and video attributes using fuzzy logic. It operates with four input parameters: video frame rate, compression rate, and spatio-temporal information. The Spearman Rank–Order Correlation Coefficient (SROCC) and Pearson Correlation Coefficient (PCC) show a high correlation between the output and the ground truth. For the training, test, and complete dataset, SROCC equals 0.8977, 0.8455, and 0.8961, respectively, while PCC equals 0.9096, 0.8632, and 0.9086, respectively. The model outperforms comparative models tested on the same dataset. Full article
(This article belongs to the Special Issue QoS in Wireless Sensor Network for IoT Applications)
Show Figures

Figure 1

Figure 1
<p>Extracted frames from the sequences of the LIVE-YT-HFR database: (<b>a</b>) a person running; (<b>b</b>) three persons running; (<b>c</b>) a person doing flips; (<b>d</b>) runners jumping hurdles; (<b>e</b>) a long-jump video; (<b>f</b>) a bobblehead spinning; (<b>g</b>) a shelf filled with books; (<b>h</b>) underwater ball bounce; (<b>i</b>) two persons passing a ball; (<b>j</b>) a street cyclist; (<b>k</b>) a hamster running a wheel; (<b>l</b>) a lamppost in a park; (<b>m</b>) falling leaves; (<b>n</b>) a spinning top; (<b>o</b>) underwater bubbles; (<b>p</b>) water waves splashing.</p>
Full article ">Figure 2
<p>Results of the FCM clustering. For each input variable, three fuzzy clusters were used. (<b>a</b>) Video FPS cluster names: low, medium, and high frame rate; (<b>b</b>) Video CRF cluster names: low, medium, and high compression; (<b>c</b>) SI cluster names: low, medium, and high complexity; (<b>d</b>) TI cluster names: low, medium, and high frame diversity.</p>
Full article ">Figure 3
<p>Membership functions of our fuzzy clusters. (<b>a</b>) Video FPS; (<b>b</b>) Video CRF; (<b>c</b>) SI; (<b>d</b>) TI.</p>
Full article ">Figure 4
<p>Membership functions that describe the fuzzy output.</p>
Full article ">Figure 5
<p>FIS block diagram of the FLAME-VQA model.</p>
Full article ">Figure 6
<p>Comparison of the MOS values between the model output and: (<b>a</b>) training dataset; (<b>b</b>) test dataset; (<b>c</b>) complete dataset.</p>
Full article ">Figure 7
<p>Three-dimensional views of the model’s output in correlation with a pair of input parameters: (<b>a</b>) Video CRF and Video FPS; (<b>b</b>) Video FPS and Video TI.</p>
Full article ">
2 pages, 146 KiB  
Editorial
Advances Techniques in Computer Vision and Multimedia
by Yang Wang
Future Internet 2023, 15(9), 294; https://doi.org/10.3390/fi15090294 - 1 Sep 2023
Viewed by 1225
Abstract
Computer vision has experienced significant advancements and great success in areas closely related to human society, which aims to enable computer systems to automatically see, recognize, and understand the visual world by simulating the mechanism of human vision [...] Full article
(This article belongs to the Special Issue Advances Techniques in Computer Vision and Multimedia)
16 pages, 1050 KiB  
Review
Enhancing E-Learning with Blockchain: Characteristics, Projects, and Emerging Trends
by Mahmoud Bidry, Abdellah Ouaguid and Mohamed Hanine
Future Internet 2023, 15(9), 293; https://doi.org/10.3390/fi15090293 - 28 Aug 2023
Cited by 11 | Viewed by 3684
Abstract
Blockchain represents a decentralized and distributed ledger technology, ensuring transparent and secure transaction recording across networks. This innovative technology offers several benefits, including increased security, trust, and transparency, making it suitable for a wide range of applications. In the last few years, there [...] Read more.
Blockchain represents a decentralized and distributed ledger technology, ensuring transparent and secure transaction recording across networks. This innovative technology offers several benefits, including increased security, trust, and transparency, making it suitable for a wide range of applications. In the last few years, there has been a growing interest in investigating the potential of Blockchain technology to enhance diverse fields, such as e-learning. In this research, we undertook a systematic literature review to explore the potential of Blockchain technology in enhancing the e-learning domain. Our research focused on four main questions: (1) What potential characteristics of Blockchain can contribute to enhancing e-learning? (2) What are the existing Blockchain projects dedicated to e-learning? (3) What are the limitations of existing projects? (4) What are the future trends in Blockchain-related research that will impact e-learning? The results showed that Blockchain technology has several characteristics that could benefit e-learning. We also discussed immutability, transparency, decentralization, security, and traceability. We also identified several existing Blockchain projects dedicated to e-learning and discussed their potential to revolutionize learning by providing more transparency, security, and effectiveness. However, our research also revealed many limitations and challenges that could be addressed to achieve Blockchain technology’s potential in e-learning. Full article
(This article belongs to the Special Issue Future Prospects and Advancements in Blockchain Technology)
Show Figures

Figure 1

Figure 1
<p>PRISMA diagram illustrating our literature research methodology.</p>
Full article ">Figure 2
<p>Document sources.</p>
Full article ">Figure 3
<p>Retrieved and included papers.</p>
Full article ">Figure 4
<p>Document types.</p>
Full article ">Figure 5
<p>Word cloud of the most frequently used keywords in this research.</p>
Full article ">Figure 6
<p>Bar chart graph of the most frequently used keywords in this research.</p>
Full article ">
19 pages, 5975 KiB  
Article
Autism Screening in Toddlers and Adults Using Deep Learning and Fair AI Techniques
by Ishaani Priyadarshini
Future Internet 2023, 15(9), 292; https://doi.org/10.3390/fi15090292 - 28 Aug 2023
Cited by 7 | Viewed by 2933
Abstract
Autism spectrum disorder (ASD) has been associated with conditions like depression, anxiety, epilepsy, etc., due to its impact on an individual’s educational, social, and employment. Since diagnosis is challenging and there is no cure, the goal is to maximize an individual’s ability by [...] Read more.
Autism spectrum disorder (ASD) has been associated with conditions like depression, anxiety, epilepsy, etc., due to its impact on an individual’s educational, social, and employment. Since diagnosis is challenging and there is no cure, the goal is to maximize an individual’s ability by reducing the symptoms, and early diagnosis plays a role in improving behavior and language development. In this paper, an autism screening analysis for toddlers and adults has been performed using fair AI (feature engineering, SMOTE, optimizations, etc.) and deep learning methods. The analysis considers traditional deep learning methods like Multilayer Perceptron (MLP), Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM), and also proposes two hybrid deep learning models, i.e., CNN–LSTM with Particle Swarm Optimization (PSO), and a CNN model combined with Gated Recurrent Units (GRU–CNN). The models have been validated using multiple performance metrics, and the analysis confirms that the proposed models perform better than the traditional models. Full article
(This article belongs to the Special Issue Machine Learning Perspective in the Convolutional Neural Network Era)
Show Figures

Figure 1

Figure 1
<p>Overall Architecture of the CNN-LSTM-PSO.</p>
Full article ">Figure 2
<p>Flow chart depicting the operations in CNN–LSTM–PSO.</p>
Full article ">Figure 3
<p>GRU–CNN architecture for classification.</p>
Full article ">Figure 4
<p>Datasets depicting features for Autism Screening in toddlers and adults.</p>
Full article ">Figure 5
<p>ASD Data distribution for toddlers and adults.</p>
Full article ">Figure 5 Cont.
<p>ASD Data distribution for toddlers and adults.</p>
Full article ">Figure 6
<p>ASD cases for toddlers and adults (ethnicity).</p>
Full article ">Figure 6 Cont.
<p>ASD cases for toddlers and adults (ethnicity).</p>
Full article ">Figure 7
<p>Ethnicity vs. gender for toddlers and adults.</p>
Full article ">Figure 7 Cont.
<p>Ethnicity vs. gender for toddlers and adults.</p>
Full article ">Figure 8
<p>ASD cases with jaundice.</p>
Full article ">Figure 9
<p>ASD cases with jaundice.</p>
Full article ">Figure 10
<p>Gender-based distribution for ASD cases in toddlers.</p>
Full article ">Figure 11
<p>Gender-based distribution for ASD cases in Adults.</p>
Full article ">Figure 12
<p>Performance evaluation of ML models for ASD (toddlers).</p>
Full article ">Figure 13
<p>Performance evaluation of ML models for ASD (Adults).</p>
Full article ">
24 pages, 14748 KiB  
Article
A Novel SDWSN-Based Testbed for IoT Smart Applications
by Duaa Zuhair Al-Hamid, Pejman A. Karegar and Peter Han Joo Chong
Future Internet 2023, 15(9), 291; https://doi.org/10.3390/fi15090291 - 28 Aug 2023
Cited by 3 | Viewed by 1500
Abstract
Wireless sensor network (WSN) environment monitoring and smart city applications present challenges for maintaining network connectivity when, for example, dynamic events occur. Such applications can benefit from recent technologies such as software-defined networks (SDNs) and network virtualization to support network flexibility and offer [...] Read more.
Wireless sensor network (WSN) environment monitoring and smart city applications present challenges for maintaining network connectivity when, for example, dynamic events occur. Such applications can benefit from recent technologies such as software-defined networks (SDNs) and network virtualization to support network flexibility and offer validation for a physical network. This paper aims to present a testbed-based, software-defined wireless sensor network (SDWSN) for IoT applications with a focus on promoting the approach of virtual network testing and analysis prior to physical network implementation to monitor and repair any network failures. Herein, physical network implementation employing hardware boards such as Texas Instruments CC2538 (TI CC2538) and TI CC1352R sensor nodes is presented and designed based on virtual WSN- based clustering for stationary and dynamic networks use cases. The key performance indicators such as evaluating node (such as a gateway node to the Internet) connection capability based on packet drop and energy consumption virtually and physically are discussed. According to the test findings, the proposed software-defined physical network benefited from “prior-to-implementation” analysis via virtualization, as the performance of both virtual and physical networks is comparable. Full article
(This article belongs to the Special Issue QoS in Wireless Sensor Network for IoT Applications)
Show Figures

Figure 1

Figure 1
<p>The interaction between the physical network and the cloud.</p>
Full article ">Figure 2
<p>The main activities for WSN functions.</p>
Full article ">Figure 3
<p>General communication messages for network structure.</p>
Full article ">Figure 4
<p>SmartRF06 evaluation board.</p>
Full article ">Figure 5
<p>TI CC2538 board with micro-USB.</p>
Full article ">Figure 6
<p>The Launchpad SensorTag kit CC1352R.</p>
Full article ">Figure 7
<p>The LaunchPad development board CC1352R.</p>
Full article ">Figure 8
<p>Debugging the LaunchPad SensorTag CC1352R using development board CC1352R.</p>
Full article ">Figure 9
<p>Raspberry Pi board.</p>
Full article ">Figure 10
<p>Indoor physical network setup.</p>
Full article ">Figure 11
<p>Outdoor physical node mobility testing.</p>
Full article ">Figure 12
<p>Pseudocode for configuring TI CC2538 leaf node.</p>
Full article ">Figure 13
<p>Pseudocode for configuring TI CC2538 coordinator node.</p>
Full article ">Figure 14
<p>A sample of the RSSI data logged to text file.</p>
Full article ">Figure 15
<p>Power function in C code.</p>
Full article ">Figure 16
<p>Proposed network architecture for physical network implementation.</p>
Full article ">Figure 17
<p>The network configuration for two development boards as sensor nodes and two SensorTags as sensors.</p>
Full article ">Figure 18
<p>Over-the-air debugging (OAD) and sensing reading on the Starter app.</p>
Full article ">Figure 19
<p>Temperature and RSSI data received by the coordinator node.</p>
Full article ">Figure 20
<p>The physical network architecture for TI LaunchPad modules and RPi.</p>
Full article ">Figure 21
<p>Python script.</p>
Full article ">Figure 22
<p>The received data from Temp and RSSI of each SensorTag on RPi terminal.</p>
Full article ">Figure 23
<p>The physical ground network formation for a large-scale network with <math display="inline"><semantics> <mrow> <mi mathvariant="bold-italic">σ</mi> </mrow> </semantics></math> = 1.</p>
Full article ">Figure 24
<p>The physical ground network formation for a large-scale network with <math display="inline"><semantics> <mrow> <mi mathvariant="bold-italic">σ</mi> </mrow> </semantics></math> = 2.</p>
Full article ">Figure 25
<p>Router capacity based on packet loss.</p>
Full article ">Figure 26
<p>Energy consumption of physical ground network in terms of message rate and gateway nodes’ spread factor.</p>
Full article ">Figure 27
<p>Packets received by coordinators (gateways) from distributed leaf nodes during physical ground network communication.</p>
Full article ">
15 pages, 455 KiB  
Article
Short-Term Mobile Network Traffic Forecasting Using Seasonal ARIMA and Holt-Winters Models
by Irina Kochetkova, Anna Kushchazli, Sofia Burtseva and Andrey Gorshenin
Future Internet 2023, 15(9), 290; https://doi.org/10.3390/fi15090290 - 28 Aug 2023
Cited by 10 | Viewed by 3330
Abstract
Fifth-generation (5G) networks require efficient radio resource management (RRM) which should dynamically adapt to the current network load and user needs. Monitoring and forecasting network performance requirements and metrics helps with this task. One of the parameters that highly influences radio resource management [...] Read more.
Fifth-generation (5G) networks require efficient radio resource management (RRM) which should dynamically adapt to the current network load and user needs. Monitoring and forecasting network performance requirements and metrics helps with this task. One of the parameters that highly influences radio resource management is the profile of user traffic generated by various 5G applications. Forecasting such mobile network profiles helps with numerous RRM tasks such as network slicing and load balancing. In this paper, we analyze a dataset from a mobile network operator in Portugal that contains information about volumes of traffic in download and upload directions in one-hour time slots. We apply two statistical models for forecasting download and upload traffic profiles, namely, seasonal autoregressive integrated moving average (SARIMA) and Holt-Winters models. We demonstrate that both models are suitable for forecasting mobile network traffic. Nevertheless, the SARIMA model is more appropriate for download traffic (e.g., MAPE [mean absolute percentage error] of 11.2% vs. 15% for Holt-Winters), while the Holt-Winters model is better suited for upload traffic (e.g., MAPE of 4.17% vs. 9.9% for SARIMA and Holt-Winters, respectively). Full article
(This article belongs to the Special Issue 5G Wireless Communication Networks II)
Show Figures

Figure 1

Figure 1
<p>“Web Applications” traffic (group of applications No. 1).</p>
Full article ">Figure 2
<p>“Terminals” traffic (group of applications No. 2).</p>
Full article ">Figure 3
<p>“VoIP” traffic (group of applications No. 3).</p>
Full article ">Figure 4
<p>Download traffic forecast using SARIMA model.</p>
Full article ">Figure 5
<p>Download traffic forecast using Holt-Winters model.</p>
Full article ">Figure 6
<p>Upload traffic forecast using SARIMA model.</p>
Full article ">Figure 7
<p>Upload traffic forecast using Holt-Winters model.</p>
Full article ">Figure 8
<p>Absolute error for traffic forecast.</p>
Full article ">
35 pages, 1386 KiB  
Article
3D Path Planning Algorithms in UAV-Enabled Communications Systems: A Mapping Study
by Jorge Carvajal-Rodriguez, Marco Morales and Christian Tipantuña
Future Internet 2023, 15(9), 289; https://doi.org/10.3390/fi15090289 - 27 Aug 2023
Cited by 3 | Viewed by 5855
Abstract
Unmanned Aerial Vehicles (UAVs) equipped with communication technologies have gained significant attention as a promising solution for providing wireless connectivity in remote, disaster-stricken areas lacking communication infrastructure. However, enabling UAVs to provide communications (e.g., UAVs acting as flying base stations) in real scenarios [...] Read more.
Unmanned Aerial Vehicles (UAVs) equipped with communication technologies have gained significant attention as a promising solution for providing wireless connectivity in remote, disaster-stricken areas lacking communication infrastructure. However, enabling UAVs to provide communications (e.g., UAVs acting as flying base stations) in real scenarios requires the integration of various technologies and algorithms. In particular 3D path planning algorithms are crucial in determining the optimal path free of obstacles so that UAVs in isolation or forming networks can provide wireless coverage in a specific region. Considering that most of the existing proposals in the literature only address path planning in a 2D environment, this paper systematically studies existing path-planning solutions in UAVs in a 3D environment in which optimization models (optimal and heuristics) have been applied. This paper analyzes 37 articles selected from 631 documents from a search in the Scopus database. This paper also presents an overview of UAV-enabled communications systems, the research questions, and the methodology for the systematic mapping study. In the end, this paper provides information about the objectives to be minimized or maximized, the optimization variables used, and the algorithmic strategies employed to solve the 3D path planning problem. Full article
Show Figures

Figure 1

Figure 1
<p>Classification of UAVs [<a href="#B1-futureinternet-15-00289" class="html-bibr">1</a>].</p>
Full article ">Figure 2
<p>3D communication network (high-altitude long-range communication and low-altitude high-density communication) [<a href="#B3-futureinternet-15-00289" class="html-bibr">3</a>].</p>
Full article ">Figure 3
<p>Path planning of a network of UAVs for communications [<a href="#B1-futureinternet-15-00289" class="html-bibr">1</a>].</p>
Full article ">Figure 4
<p>Classification of 3D UAVs path planning algorithms.</p>
Full article ">Figure 5
<p>Process flow and tasks of the systematic mapping study, based on [<a href="#B24-futureinternet-15-00289" class="html-bibr">24</a>].</p>
Full article ">Figure 6
<p>Search string with keywords in Scopus.</p>
Full article ">Figure 7
<p>Search string with automatic inclusion–exclusion criteria.</p>
Full article ">Figure 8
<p>Decision tree for manual selection of papers, based on [<a href="#B24-futureinternet-15-00289" class="html-bibr">24</a>].</p>
Full article ">Figure 9
<p>Classification scheme.</p>
Full article ">Figure 10
<p>Scope of the study, based on [<a href="#B24-futureinternet-15-00289" class="html-bibr">24</a>].</p>
Full article ">Figure 11
<p>Scope of the study, based on [<a href="#B24-futureinternet-15-00289" class="html-bibr">24</a>].</p>
Full article ">Figure 12
<p>Optimization objective functions vs. analyzed papers.</p>
Full article ">Figure 13
<p>Static or dynamic environments that have been considered in path planning.</p>
Full article ">Figure 14
<p>Variables to optimize in the analyzed articles. The optimized multiple variables are observed in lead color, and the predominant variables in blue and orange.</p>
Full article ">Figure 15
<p>Objective to be Min-max optimized vs. Optimization Variable.</p>
Full article ">
17 pages, 3268 KiB  
Article
Spot Market Cloud Orchestration Using Task-Based Redundancy and Dynamic Costing
by Vyas O’Neill and Ben Soh
Future Internet 2023, 15(9), 288; https://doi.org/10.3390/fi15090288 - 27 Aug 2023
Viewed by 1363
Abstract
Cloud computing has become ubiquitous in the enterprise environment as its on-demand model realizes technical and economic benefits for users. Cloud users demand a level of reliability, availability, and quality of service. Improvements to reliability generally come at the cost of additional replication. [...] Read more.
Cloud computing has become ubiquitous in the enterprise environment as its on-demand model realizes technical and economic benefits for users. Cloud users demand a level of reliability, availability, and quality of service. Improvements to reliability generally come at the cost of additional replication. Existing approaches have focused on the replication of virtual environments as a method of improving the reliability of cloud services. As cloud systems move towards microservices-based architectures, a more granular approach to replication is now possible. In this paper, we propose a cloud orchestration approach that balances the potential cost of failure with the spot market running cost, optimizing the resource usage of the cloud system. We present the results of empirical testing we carried out using a simulator to compare the outcome of our proposed approach to a control algorithm based on a static reliability requirement. Our empirical testing showed an improvement of between 37% and 72% in total cost over the control, depending on the specific characteristics of the cloud models tested. We thus propose that in clouds where the cost of failure can be reasonably approximated, our approach may be used to optimize the cloud redundancy configuration to achieve a lower total cost. Full article
Show Figures

Figure 1

Figure 1
<p>Comparison of virtual machine (<b>left</b>) and container (<b>right</b>) architectures.</p>
Full article ">Figure 2
<p>Stateless container architecture.</p>
Full article ">Figure 3
<p>Classification tree of approaches to cloud resilience.</p>
Full article ">Figure 4
<p>Our Task-Based Redundancy (TBR) model for cloud microservices. The diagram shows a single task composed of several microservices. The microservices shown may also be dependencies for other independent tasks in the cloud system.</p>
Full article ">Figure 5
<p>High-level architecture of the open-source Cloud Reliability Simulator, showing our proposed spot market functionality.</p>
Full article ">Figure 6
<p>Control algorithm behavior in Experiment 1 with immediate detection and remediation of failures.</p>
Full article ">Figure 7
<p>Experimental algorithm behavior in Experiment 1 with immediate detection and remediation of failures.</p>
Full article ">Figure 8
<p>Control algorithm behavior in Experiment 2 with immediate detection and remediation of failures.</p>
Full article ">Figure 9
<p>Experimental algorithm behavior in Experiment 2 with immediate detection and remediation of failures.</p>
Full article ">
Previous Issue
Next Issue
Back to TopTop