[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Next Issue
Volume 16, March
Previous Issue
Volume 16, January
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
7.1
2.8
Journals
Future Internet
Volume 16
Issue 2
share announcement

Future Internet, Volume 16, Issue 2 (February 2024) – 31 articles

Cover Story (view full-size image): Tiny Machine Learning (TinyML) and Big Data, enhanced by Edge Artificial Intelligence, are essential for effectively managing the extensive data produced by numerous connected devices. Our study introduces a set of TinyML algorithms designed and developed to improve Big Data management in large-scale IoT systems. These algorithms, named TinyCleanEDF, EdgeClusterML, CompressEdgeML, CacheEdgeML, and TinyHybridSenseQ, operate together to enhance data processing, storage, and quality control in IoT networks, utilizing the capabilities of Edge AI. Our experimental evaluation of the proposed techniques includes executing all the algorithms in various numbers of Raspberry Pi devices ranging from one to ten. Ultimately, we anticipate that the proposed algorithms will offer a comprehensive and efficient approach to managing the complexities of IoT, Big Data, and Edge AI. 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
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
21 pages, 3761 KiB  
Article
Energy-Efficient De-Duplication Mechanism for Healthcare Data Aggregation in IoT
by Muhammad Nafees Ulfat Khan, Weiping Cao, Zhiling Tang, Ata Ullah and Wanghua Pan
Future Internet 2024, 16(2), 66; https://doi.org/10.3390/fi16020066 - 19 Feb 2024
Cited by 2 | Viewed by 1770
Abstract
The rapid development of the Internet of Things (IoT) has opened the way for transformative advances in numerous fields, including healthcare. IoT-based healthcare systems provide unprecedented opportunities to gather patients’ real-time data and make appropriate decisions at the right time. Yet, the deployed [...] Read more.
The rapid development of the Internet of Things (IoT) has opened the way for transformative advances in numerous fields, including healthcare. IoT-based healthcare systems provide unprecedented opportunities to gather patients’ real-time data and make appropriate decisions at the right time. Yet, the deployed sensors generate normal readings most of the time, which are transmitted to Cluster Heads (CHs). Handling these voluminous duplicated data is quite challenging. The existing techniques have high energy consumption, storage costs, and communication costs. To overcome these problems, in this paper, an innovative Energy-Efficient Fuzzy Data Aggregation System (EE-FDAS) has been presented. In it, at the first level, it is checked that sensors either generate normal or critical readings. In the first case, readings are converted to Boolean digit 0. This reduced data size takes only 1 digit which considerably reduces energy consumption. In the second scenario, sensors generating irregular readings are transmitted in their original 16 or 32-bit form. Then, data are aggregated and transmitted to respective CHs. Afterwards, these data are further transmitted to Fog servers, from where doctors have access. Lastly, for later usage, data are stored in the cloud server. For checking the proficiency of the proposed EE-FDAS scheme, extensive simulations are performed using NS-2.35. The results showed that EE-FDAS has performed well in terms of aggregation factor, energy consumption, packet drop rate, communication, and storage cost. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>System Model.</p> Full article ">Figure 2
<p>Phases of EE-FDAS.</p> Full article ">Figure 3
<p>The impact of the Number of Attributes on Average Aggregation is explained in (<b>a</b>) while Energy Consumption during the data phase is shown in (<b>b</b>).</p> Full article ">Figure 4
<p>The impact of the Number of Attributes on Average SL is shown in (<b>a</b>) whereas Needed Transmission Slots are displayed in (<b>b</b>).</p> Full article ">Figure 5
<p>The impact of the Number of Attributes on Control overhead is shown in (<b>a</b>) whereas Communication Cost is exhibited in (<b>b</b>).</p> Full article ">Figure 6
<p>In (<b>a</b>), the graph displays Packet Delivery Ratio, representing successfully delivered packets per unit time, while (<b>b</b>) illustrates the Packet Loss Ratio.</p> Full article ">Figure 7
<p>The Storage Cost by Number of packets per unit time.</p> Full article ">
18 pages, 1478 KiB  
Article
IoTwins: Implementing Distributed and Hybrid Digital Twins in Industrial Manufacturing and Facility Management Settings
by Paolo Bellavista and Giuseppe Di Modica
Future Internet 2024, 16(2), 65; https://doi.org/10.3390/fi16020065 - 17 Feb 2024
Cited by 1 | Viewed by 2117
Abstract
A Digital Twin (DT) refers to a virtual representation or digital replica of a physical object, system, process, or entity. This concept involves creating a detailed, real-time digital counterpart that mimics the behavior, characteristics, and attributes of its physical counterpart. DTs have the [...] Read more.
A Digital Twin (DT) refers to a virtual representation or digital replica of a physical object, system, process, or entity. This concept involves creating a detailed, real-time digital counterpart that mimics the behavior, characteristics, and attributes of its physical counterpart. DTs have the potential to improve efficiency, reduce costs, and enhance decision-making by providing a detailed, real-time understanding of the physical systems they represent. While this technology is finding application in numerous fields, such as energy, healthcare, and transportation, it appears to be a key component of the digital transformation of industries fostered by the fourth Industrial revolution (Industry 4.0). In this paper, we present the research results achieved by IoTwins, a European research project aimed at investigating opportunities and issues of adopting DTs in the fields of industrial manufacturing and facility management. Particularly, we discuss a DT model and a reference architecture for use by the research community to implement a platform for the development and deployment of industrial DTs in the cloud continuum. Guided by the devised architectures’ principles, we implemented an open platform and a development methodology to help companies build DT-based industrial applications and deploy them in the so-called Edge/Cloud continuum. To prove the research value and the usability of the implemented platform, we discuss a simple yet practical development use case. Full article
(This article belongs to the Section Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>IoTwins: a distributed and hybrid Digital Twins model.</p> Full article ">Figure 2
<p>IoTwins-RAMI 4.0 mapping.</p> Full article ">Figure 3
<p>Software components of the IoTwins platform deployed in the Cloud.</p> Full article ">Figure 4
<p>Software components of the IoTwins platform deployed in the Edge.</p> Full article ">Figure 5
<p>Digital Twins software components provisioned by the orchestrator.</p> Full article ">
15 pages, 510 KiB  
Article
Online Optimization of Pickup and Delivery Problem Considering Feasibility
by Ryo Matsuoka, Koichi Kobayashi and Yuh Yamashita
Future Internet 2024, 16(2), 64; https://doi.org/10.3390/fi16020064 - 17 Feb 2024
Viewed by 1746
Abstract
A pickup and delivery problem by multiple agents has many applications, such as food delivery service and disaster rescue. In this problem, there are cases where fuels must be considered (e.g., the case of using drones as agents). In addition, there are cases [...] Read more.
A pickup and delivery problem by multiple agents has many applications, such as food delivery service and disaster rescue. In this problem, there are cases where fuels must be considered (e.g., the case of using drones as agents). In addition, there are cases where demand forecasting should be considered (e.g., the case where a large number of orders are carried by a small number of agents). In this paper, we consider an online pickup and delivery problem considering fuel and demand forecasting. First, the pickup and delivery problem with fuel constraints is formulated. The information on demand forecasting is included in the cost function. Based on the orders, the agents’ paths (e.g., the paths from stores to customers) are calculated. We suppose that the target area is given by an undirected graph. Using a given graph, several constraints such as the moves and fuels of the agents are introduced. This problem is reduced to a mixed integer linear programming (MILP) problem. Next, in online optimization, the MILP problem is solved depending on the acceptance of orders. Owing to new orders, the calculated future paths may be changed. Finally, by using a numerical example, we present the effectiveness of the proposed method. Full article
(This article belongs to the Special Issue State-of-the-Art Future Internet Technology in Japan 2022-2023)
►▼ Show Figures

Figure 1

Figure 1
<p>Example of an undirected graph <math display="inline"><semantics> <mrow> <mi mathvariant="script">G</mi> <mo>=</mo> <mo>(</mo> <mi mathvariant="script">V</mi> <mo>,</mo> <mi mathvariant="script">E</mi> <mo>)</mo> </mrow> </semantics></math>, where <math display="inline"><semantics> <mrow> <mo>|</mo> <mi mathvariant="script">S</mi> <mo>|</mo> <mo>=</mo> <mi>N</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>|</mo> <mi mathvariant="script">C</mi> <mo>|</mo> <mo>=</mo> <mi>M</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math>, and <math display="inline"><semantics> <mrow> <mo>|</mo> <mi mathvariant="script">R</mi> <mo>|</mo> <mo>=</mo> <mi>J</mi> <mo>=</mo> <mn>6</mn> </mrow> </semantics></math>.</p> Full article ">Figure 2
<p>Outline of the procedure of online optimization.</p> Full article ">Figure 3
<p>Target area, where an agent can move to an up/down/left/right vertex.</p> Full article ">Figure 4
<p>Results at time 0 with existing method.</p> Full article ">Figure 5
<p>Results at time 0 with proposed method.</p> Full article ">Figure 6
<p>Results at time 5 with existing method.</p> Full article ">Figure 7
<p>Results at time 5 with proposed method.</p> Full article ">Figure 8
<p>Results at time 8 with proposed method.</p> Full article ">
20 pages, 19399 KiB  
Article
Speech Inpainting Based on Multi-Layer Long Short-Term Memory Networks
by Haohan Shi, Xiyu Shi and Safak Dogan
Future Internet 2024, 16(2), 63; https://doi.org/10.3390/fi16020063 - 17 Feb 2024
Cited by 3 | Viewed by 1648
Abstract
Audio inpainting plays an important role in addressing incomplete, damaged, or missing audio signals, contributing to improved quality of service and overall user experience in multimedia communications over the Internet and mobile networks. This paper presents an innovative solution for speech inpainting using [...] Read more.
Audio inpainting plays an important role in addressing incomplete, damaged, or missing audio signals, contributing to improved quality of service and overall user experience in multimedia communications over the Internet and mobile networks. This paper presents an innovative solution for speech inpainting using Long Short-Term Memory (LSTM) networks, i.e., a restoring task where the missing parts of speech signals are recovered from the previous information in the time domain. The lost or corrupted speech signals are also referred to as gaps. We regard the speech inpainting task as a time-series prediction problem in this research work. To address this problem, we designed multi-layer LSTM networks and trained them on different speech datasets. Our study aims to investigate the inpainting performance of the proposed models on different datasets and with varying LSTM layers and explore the effect of multi-layer LSTM networks on the prediction of speech samples in terms of perceived audio quality. The inpainted speech quality is evaluated through the Mean Opinion Score (MOS) and a frequency analysis of the spectrogram. Our proposed multi-layer LSTM models are able to restore up to 1 s of gaps with high perceptual audio quality using the features captured from the time domain only. Specifically, for gap lengths under 500 ms, the MOS can reach up to 3~4, and for gap lengths ranging between 500 ms and 1 s, the MOS can reach up to 2~3. In the time domain, the proposed models can proficiently restore the envelope and trend of lost speech signals. In the frequency domain, the proposed models can restore spectrogram blocks with higher similarity to the original signals at frequencies less than 2.0 kHz and comparatively lower similarity at frequencies in the range of 2.0 kHz~8.0 kHz. Full article
(This article belongs to the Special Issue Deep Learning and Natural Language Processing II)
►▼ Show Figures

Figure 1

Figure 1
<p>LSTM cell structure.</p> Full article ">Figure 2
<p>The inpainting process and neural network structure of the proposed speech inpainting model.</p> Full article ">Figure 3
<p>Training losses of the proposed speech inpainting models on single-speaker datasets with different numbers of LSTM layers: (<b>a</b>–<b>e</b>) speech inpainting models with two to six LSTM layers, respectively; (<b>f</b>) the 5–layer LSTM model trained on four multi–speaker datasets.</p> Full article ">Figure 4
<p>Frequency analysis for gap lengths less than 100 ms. The gaps start at 1.62 s and last for 20 ms, 40 ms, 50 ms, and 100 ms, respectively, corresponding to the first to fourth columns on the right side of the figure. The zoomed–in and gap areas are marked with red dashed lines.</p> Full article ">Figure 5
<p>Frequency analysis for gap lengths greater than 100 ms. The gaps start at 1.62 s and last for 200 ms, 500 ms, and 1000 ms, respectively, corresponding to the first to third columns on the right side of the figure. The zoomed–in and gap areas are marked with red dashed lines.</p> Full article ">Figure 6
<p>Frequency analysis for gap lengths less than 100 ms. The gaps start at 2.88 s and last for 20 ms, 40 ms, 50 ms, and 100 ms, respectively, corresponding to the first to fourth columns on the right side of the figure. The zoomed–in and gap areas are marked with red dashed lines.</p> Full article ">Figure 7
<p>Frequency analysis for gap lengths greater than 100 ms. The gaps start at 2.88 s and last for 200 ms, 500 ms, and 1000 ms, respectively, corresponding to the first to third columns on the right side of the figure. The zoomed–in and gap areas are marked with red dashed lines.</p> Full article ">
16 pages, 1186 KiB  
Article
Merging Ontologies and Data from Electronic Health Records
by Salvatore Calcagno, Andrea Calvagna, Emiliano Tramontana and Gabriella Verga
Future Internet 2024, 16(2), 62; https://doi.org/10.3390/fi16020062 - 17 Feb 2024
Cited by 1 | Viewed by 1731
Abstract
The Electronic Health Record (EHR) is a system for collecting and storing patient medical records as data that can be mechanically accessed, hence facilitating and assisting the medical decision-making process. EHRs exist in several formats, and each format lists thousands of keywords to [...] Read more.
The Electronic Health Record (EHR) is a system for collecting and storing patient medical records as data that can be mechanically accessed, hence facilitating and assisting the medical decision-making process. EHRs exist in several formats, and each format lists thousands of keywords to classify patients data. The keywords are specific and are medical jargon; hence, data classification is very accurate. As the keywords constituting the formats of medical records express concepts by means of specific jargon without definitions or references, their proper use is left to clinicians and could be affected by their background, hence the interpretation of data could become slow or less accurate than that desired. This article presents an approach that accurately relates data in EHRs to ontologies in the medical realm. Thanks to ontologies, clinicians can be assisted when writing or analysing health records, e.g., our solution promptly suggests rigorous definitions for scientific terms, and automatically connects data spread over several parts of EHRs. The first step of our approach consists of converting selected data and keywords from several EHR formats into a format easier to parse, then the second step is merging the extracted data with specialised medical ontologies. Finally, enriched versions of the medical data are made available to professionals. The proposed approach was validated by taking samples of medical records and ontologies in the real world. The results have shown both versatility on handling data, precision of query results, and appropriate suggestions for relations among medical records. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Snippet of XML code displaying intrinsic syntactic complexity despite being extracted from a simple, basic example of CDA document. Points marked 1 to 4 highlight the following tags: root, codeSystem, code, and ID, respectively.</p> Full article ">Figure 2
<p>Comparison of two XML files with different structures but the same type of content.</p> Full article ">Figure 3
<p>On the <b>left</b>, the main branches of the Human Disease Ontology and on the <b>right</b> some classes that derive from the first disease, or disease by infectious agent.</p> Full article ">Figure 4
<p>A representation of the class penicillin allergy found in HDO.</p> Full article ">Figure 5
<p>A representation of the class anaphylactic shock found in HDO.</p> Full article ">Figure 6
<p>Example of a patient’s medical record with ID 444222222, showing data taken from three HL7 CDA files (centre) and HDO ontology (accessed on 10 November 2023).</p> Full article ">
24 pages, 1502 KiB  
Article
Enhancing Energy Efficiency in IoT-NDN via Parameter Optimization
by Dennis Papenfuß, Bennet Gerlach, Stefan Fischer and Mohamed Ahmed Hail
Future Internet 2024, 16(2), 61; https://doi.org/10.3390/fi16020061 - 16 Feb 2024
Cited by 1 | Viewed by 1670
Abstract
The IoT encompasses objects, sensors, and everyday items not typically considered computers. IoT devices are subject to severe energy, memory, and computation power constraints. Employing NDN for the IoT is a recent approach to accommodate these issues. To gain a deeper insight into [...] Read more.
The IoT encompasses objects, sensors, and everyday items not typically considered computers. IoT devices are subject to severe energy, memory, and computation power constraints. Employing NDN for the IoT is a recent approach to accommodate these issues. To gain a deeper insight into how different network parameters affect energy consumption, analyzing a range of parameters using hyperparameter optimization seems reasonable. The experiments from this work’s ndnSIM-based hyperparameter setup indicate that the data packet size has the most significant impact on energy consumption, followed by the caching scheme, caching strategy, and finally, the forwarding strategy. The energy footprint of these parameters is orders of magnitude apart. Surprisingly, the packet request sequence influences the caching parameters’ energy footprint more than the graph size and topology. Regarding energy consumption, the results indicate that data compression may be more relevant than expected, and caching may be more significant than the forwarding strategy. The framework for ndnSIM developed in this work can be used to simulate NDN networks more efficiently. Furthermore, the work presents a valuable basis for further research on the effect of specific parameter combinations not examined before. Full article
(This article belongs to the Special Issue Featured Papers in the Section Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>An overview of all the paramters used in the framework and their possible values. The boxes represent parameter categories, and the bold text below them depicts the concrete parameters.</p> Full article ">Figure 2
<p>The structure of the program: The parameter, default and user-defined file go to <tt>config-file-reader.cc</tt>, validating the input. Then <tt>graph-generator.cc</tt> generates the corresponding graph, all of which serves as input for <tt>simulator.cc</tt>. The results are written into some csv file, e.g., <tt>results.csv</tt>.</p> Full article ">Figure 3
<p>A diagram of the <tt>Policy/Scheme</tt> class hierarchy.</p> Full article ">Figure 4
<p>Two examples for the data packet size: The average energy consumption of packets sizes 128, 1024, and 8192 is compared to the overall mean of the corresponding fixed parameter combination. The black line represents the mean (100%).</p> Full article ">Figure 5
<p>Two examples for the caching scheme: (<b>a</b>) shows the probabilities for the repetitive scheme, grid topology and size 64. (<b>b</b>) demonstrates the total energy difference between schemes: The consumed energy per node on average for probabilistic 50% is 287.12 J, while for no caching the consumed energy is 287.38 J.</p> Full article ">Figure 6
<p>Two examples for the caching strategy: For the simple random request type, FIFO and LRU are consistently the two best caching strategies, and FWF is consistently the worst or one of the worst. The black line signifies the average rank over all variable parameters.</p> Full article ">
18 pages, 6477 KiB  
Article
The Microverse: A Task-Oriented Edge-Scale Metaverse
by Qian Qu, Mohsen Hatami, Ronghua Xu, Deeraj Nagothu, Yu Chen, Xiaohua Li, Erik Blasch, Erika Ardiles-Cruz and Genshe Chen
Future Internet 2024, 16(2), 60; https://doi.org/10.3390/fi16020060 - 13 Feb 2024
Cited by 11 | Viewed by 2611
Abstract
Over the past decade, there has been a remarkable acceleration in the evolution of smart cities and intelligent spaces, driven by breakthroughs in technologies such as the Internet of Things (IoT), edge–fog–cloud computing, and machine learning (ML)/artificial intelligence (AI). As society begins to [...] Read more.
Over the past decade, there has been a remarkable acceleration in the evolution of smart cities and intelligent spaces, driven by breakthroughs in technologies such as the Internet of Things (IoT), edge–fog–cloud computing, and machine learning (ML)/artificial intelligence (AI). As society begins to harness the full potential of these smart environments, the horizon brightens with the promise of an immersive, interconnected 3D world. The forthcoming paradigm shift in how we live, work, and interact owes much to groundbreaking innovations in augmented reality (AR), virtual reality (VR), extended reality (XR), blockchain, and digital twins (DTs). However, realizing the expansive digital vista in our daily lives is challenging. Current limitations include an incomplete integration of pivotal techniques, daunting bandwidth requirements, and the critical need for near-instantaneous data transmission, all impeding the digital VR metaverse from fully manifesting as envisioned by its proponents. This paper seeks to delve deeply into the intricacies of the immersive, interconnected 3D realm, particularly in applications demanding high levels of intelligence. Specifically, this paper introduces the microverse, a task-oriented, edge-scale, pragmatic solution for smart cities. Unlike all-encompassing metaverses, each microverse instance serves a specific task as a manageable digital twin of an individual network slice. Each microverse enables on-site/near-site data processing, information fusion, and real-time decision-making within the edge–fog–cloud computing framework. The microverse concept is verified using smart public safety surveillance (SPSS) for smart communities as a case study, demonstrating its feasibility in practical smart city applications. The aim is to stimulate discussions and inspire fresh ideas in our community, guiding us as we navigate the evolving digital landscape of smart cities to embrace the potential of the metaverse. Full article
(This article belongs to the Special Issue State-of-the-Art Future Internet Technology in USA 2022–2023)
►▼ Show Figures

Figure 1

Figure 1
<p>Microverse: a hierarchical view.</p> Full article ">Figure 2
<p>SPSS microverse prototype architecture.</p> Full article ">Figure 3
<p>SPSS microverse prototype workflow.</p> Full article ">Figure 4
<p>Screen shots. (<b>a</b>) Designed Android app. (<b>b</b>) Real-time object detection output.</p> Full article ">Figure 5
<p>Screenshots of UE5-based microverse prototype. (<b>a</b>) Both drones are in live-stream mode. (<b>b</b>) The first drone switches to detection mode.</p> Full article ">Figure 6
<p>Screenshots of immersive VR experience.</p> Full article ">Figure 7
<p>(<b>a</b>) Delay in different resolutions at 30 Fps. (<b>b</b>) Streaming throughput in different resolutions at different bit rate settings.</p> Full article ">
16 pages, 3576 KiB  
Article
Digital-Twin-Based Monitoring System for Slab Production Process
by Tianjie Fu, Peiyu Li, Chenke Shi and Youzhu Liu
Future Internet 2024, 16(2), 59; https://doi.org/10.3390/fi16020059 - 13 Feb 2024
Viewed by 1778
Abstract
The growing demand for high-quality steel across various industries has led to an increasing need for superior-grade steel. The quality of slab ingots is a pivotal factor influencing the final quality of steel production. However, the current level of intelligence in the steelmaking [...] Read more.
The growing demand for high-quality steel across various industries has led to an increasing need for superior-grade steel. The quality of slab ingots is a pivotal factor influencing the final quality of steel production. However, the current level of intelligence in the steelmaking industry’s processes is relatively insufficient. Consequently, slab ingot quality inspection is characterized by high-temperature risks and imprecision. The positional accuracy of quality detection is inadequate, and the precise quantification of slab ingot production and quality remains challenging. This paper proposes a digital twin (DT)-based monitoring system for the slab ingot production process that integrates DT technology with slab ingot process detection. A neural network is introduced for defect identification to ensure precise defect localization and efficient recognition. Concurrently, environmental production factors are considered, leading to the introduction of a defect prediction module. The effectiveness of this system is validated through experimental verification. Full article
(This article belongs to the Special Issue The Flexibility and Adaptability of the Digital Twin-Driven Manufacturing System in the Mass Customization Paradigm)
►▼ Show Figures

Figure 1

Figure 1
<p>Monitoring system structure.</p> Full article ">Figure 2
<p>Defect identification module.</p> Full article ">Figure 3
<p>The REPVGG module.</p> Full article ">Figure 4
<p>Attention unit (AU).</p> Full article ">Figure 5
<p>Experimental validation process.</p> Full article ">Figure 6
<p>Effective identification of slab defects.</p> Full article ">Figure 7
<p>System interface.</p> Full article ">Figure 8
<p>Defect database interface.</p> Full article ">
16 pages, 463 KiB  
Article
CROWDMATCH: Optimizing Crowdsourcing Matching through the Integration of Matching Theory and Coalition Games
by Adedamola Adesokan, Rowan Kinney and Eirini Eleni Tsiropoulou
Future Internet 2024, 16(2), 58; https://doi.org/10.3390/fi16020058 - 11 Feb 2024
Cited by 1 | Viewed by 1598
Abstract
This paper tackles the challenges inherent in crowdsourcing dynamics by introducing the CROWDMATCH mechanism. Aimed at enabling crowdworkers to strategically select suitable crowdsourcers while contributing information to crowdsourcing tasks, CROWDMATCH considers incentives, information availability and cost, and the decisions of fellow crowdworkers to [...] Read more.
This paper tackles the challenges inherent in crowdsourcing dynamics by introducing the CROWDMATCH mechanism. Aimed at enabling crowdworkers to strategically select suitable crowdsourcers while contributing information to crowdsourcing tasks, CROWDMATCH considers incentives, information availability and cost, and the decisions of fellow crowdworkers to model the utility functions for both the crowdworkers and the crowdsourcers. Specifically, the paper presents an initial Approximate CROWDMATCH mechanism grounded in matching theory principles, eliminating externalities from crowdworkers’ decisions and enabling each entity to maximize its utility. Subsequently, the Accurate CROWDMATCH mechanism is introduced, which is initiated by the outcome of the Approximate CROWDMATCH mechanism, and coalition game-theoretic principles are employed to refine the matching process by accounting for externalities. The paper’s contributions include the introduction of the CROWDMATCH system model, the development of both Approximate and Accurate CROWDMATCH mechanisms, and a demonstration of their superior performance through comprehensive simulation results. The mechanisms’ scalability in large-scale crowdsourcing systems and operational advantages are highlighted, distinguishing them from existing methods and highlighting their efficacy in empowering crowdworkers in crowdsourcer selection. Full article
(This article belongs to the Special Issue State-of-the-Art Future Internet Technology in USA 2022–2023)
►▼ Show Figures

Figure 1

Figure 1
<p>Crowdworkers’ (<b>a</b>) approximate (Equation (<a href="#FD3-futureinternet-16-00058" class="html-disp-formula">3</a>)), and (<b>b</b>) accurate utility (Equation (<a href="#FD1-futureinternet-16-00058" class="html-disp-formula">1</a>)).</p> Full article ">Figure 2
<p>Number of matched crowdworkers to crowdsourcers and crowdsourcers’ utility under the (<b>a</b>) Approximate CROWDMATCH algorithm, and (<b>b</b>) Accurate CROWDMATCH algorithm.</p> Full article ">Figure 3
<p>Total crowdworkers’ utility per associated crowdsourcer under the (<b>a</b>) Approximate, and (<b>b</b>) Accurate CROWDMATCH algorithms.</p> Full article ">Figure 4
<p>Scalability analysis for an increasing number of crowdworkers: (<b>a</b>) execution time, and (<b>b</b>) average crowdworkers’ utility under the Approximate and Accurate CROWDMATCH algorithms.</p> Full article ">Figure 5
<p>Scalability analysis for an increasing number of crowdsourcers: (<b>a</b>) execution time, and (<b>b</b>) average crowdworkers’ utility under the Approximate and Accurate CROWDMATCH Algorithms.</p> Full article ">Figure 6
<p>Comparative evaluation: (<b>a</b>) average crowdsourcers’ utility and (<b>b</b>) average crowdworkers’ utility under the (i) Accurate CROWDMATCH mechanism, (ii) Highest Reward, (iii) Random Matching, and (iv) Stochastic Learning Automata (SLA) models.</p> Full article ">
26 pages, 17847 KiB  
Article
Distributed Mobility Management Support for Low-Latency Data Delivery in Named Data Networking for UAVs
by Mohammed Bellaj, Najib Naja and Abdellah Jamali
Future Internet 2024, 16(2), 57; https://doi.org/10.3390/fi16020057 - 10 Feb 2024
Cited by 1 | Viewed by 1931
Abstract
Named Data Networking (NDN) has emerged as a promising architecture to overcome the limitations of the conventional Internet Protocol (IP) architecture, particularly in terms of mobility, security, and data availability. However, despite the advantages it offers, producer mobility management remains a significant challenge [...] Read more.
Named Data Networking (NDN) has emerged as a promising architecture to overcome the limitations of the conventional Internet Protocol (IP) architecture, particularly in terms of mobility, security, and data availability. However, despite the advantages it offers, producer mobility management remains a significant challenge for NDN, especially for moving vehicles and emerging technologies such as Unmanned Aerial Vehicles (UAVs), known for their high-speed and unpredictable movements, which makes it difficult for NDN to maintain seamless communication. To solve this mobility problem, we propose a Distributed Mobility Management Scheme (DMMS) to support UAV mobility and ensure low-latency content delivery in NDN architecture. DMMS utilizes decentralized Anchors to forward proactively the consumer’s Interest packets toward the producer’s predicted location when handoff occurs. Moreover, it introduces a new forwarding approach that combines the standard and location-based forwarding strategy to improve forwarding efficiency under producer mobility without changing the network structure. Using a realistic scenario, DMMS is evaluated and compared against two well-known solutions, namely MAP-ME and Kite, using the ndnSIM simulations. We demonstrate that DMMS achieves better results compared to Kite and MAP-ME solutions in terms of network cost and consumer quality-of-service metrics. Full article
(This article belongs to the Topic Future Internet Architecture: Difficulties and Opportunities)
►▼ Show Figures

Figure 1

Figure 1
<p>Anchor-based distributed mobility management architecture.</p> Full article ">Figure 2
<p>Mobility management scenario: The MP sends the CP packet when it connects to the network for the first time (step 0). The MP send the RP packet prior to handoff (step 2).</p> Full article ">Figure 3
<p>Example of the forwarding of Interest and Data packets.</p> Full article ">Figure 4
<p>Pipeline of the prediction–update procedure of the KF.</p> Full article ">Figure 5
<p>WMN topology of the city of Barcelona, Spain with Wi-Fi Access points (cyan dots).</p> Full article ">Figure 6
<p>(<b>a</b>) An extract of Barcelona WMN topology; (<b>b</b>) the corresponding 5 × 5 topology used with rendezvous Anchor in red.</p> Full article ">Figure 7
<p>Comparison of the actual trajectory of the producer with the estimated trajectory using the Kalman filter.</p> Full article ">Figure 8
<p>Average Interest packet loss rates vs. mobile producer speed.</p> Full article ">Figure 9
<p>(<b>a</b>) Average path stretch vs. mobile producer speed, (<b>b</b>) average hop count vs. mobile producer speed.</p> Full article ">Figure 10
<p>Average data packet delay vs. mobile producer speed.</p> Full article ">Figure 11
<p>Signaling overheads vs. producer speed.</p> Full article ">Figure 12
<p>Handoff latency vs. producer speed.</p> Full article ">Figure 13
<p>(<b>a</b>) Average delay vs. mobile producer speed, (<b>b</b>) average path stretch vs. mobile producer speed.</p> Full article ">
24 pages, 8449 KiB  
Article
A Secure Opportunistic Network with Efficient Routing for Enhanced Efficiency and Sustainability
by Ayman Khalil and Besma Zeddini
Future Internet 2024, 16(2), 56; https://doi.org/10.3390/fi16020056 - 8 Feb 2024
Cited by 4 | Viewed by 1717
Abstract
The intersection of cybersecurity and opportunistic networks has ushered in a new era of innovation in the realm of wireless communications. In an increasingly interconnected world, where seamless data exchange is pivotal for both individual users and organizations, the need for efficient, reliable, [...] Read more.
The intersection of cybersecurity and opportunistic networks has ushered in a new era of innovation in the realm of wireless communications. In an increasingly interconnected world, where seamless data exchange is pivotal for both individual users and organizations, the need for efficient, reliable, and sustainable networking solutions has never been more pressing. Opportunistic networks, characterized by intermittent connectivity and dynamic network conditions, present unique challenges that necessitate innovative approaches for optimal performance and sustainability. This paper introduces a groundbreaking paradigm that integrates the principles of cybersecurity with opportunistic networks. At its core, this study presents a novel routing protocol meticulously designed to significantly outperform existing solutions concerning key metrics such as delivery probability, overhead ratio, and communication delay. Leveraging cybersecurity’s inherent strengths, our protocol not only fortifies the network’s security posture but also provides a foundation for enhancing efficiency and sustainability in opportunistic networks. The overarching goal of this paper is to address the inherent limitations of conventional opportunistic network protocols. By proposing an innovative routing protocol, we aim to optimize data delivery, minimize overhead, and reduce communication latency. These objectives are crucial for ensuring seamless and timely information exchange, especially in scenarios where traditional networking infrastructures fall short. By large-scale simulations, the new model proves its effectiveness in the different scenarios, especially in terms of message delivery probability, while ensuring reasonable overhead and latency. Full article
(This article belongs to the Collection 5G/6G Networks for the Internet of Things: Communication Technologies and Challenges)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed FT-OLSR scheme.</p> Full article ">Figure 2
<p>Schematic of two required tables.</p> Full article ">Figure 3
<p>Acked Messages table flowchart.</p> Full article ">Figure 4
<p>Message overhead.</p> Full article ">Figure 5
<p>Phase 1 functionality.</p> Full article ">Figure 6
<p>Phase 2 functionality.</p> Full article ">Figure 7
<p>Phase 3 functionality.</p> Full article ">Figure 8
<p>Phase 4 functionality.</p> Full article ">Figure 9
<p>Detection time.</p> Full article ">Figure 10
<p>Delivery probability with respect to the number of messages.</p> Full article ">Figure 11
<p>Overhead ratio with respect to the number of messages.</p> Full article ">Figure 12
<p>Latency with respect to the number of messages.</p> Full article ">Figure 13
<p>Delivery probability with respect to the number of users.</p> Full article ">Figure 14
<p>Overhead ratio with respect to the number of users.</p> Full article ">Figure 15
<p>Latency with respect to the number of users.</p> Full article ">Figure 16
<p>Delivery probability with respect to the transmission speed.</p> Full article ">Figure 17
<p>Overhead ratio with respect to the transmission speed.</p> Full article ">Figure 18
<p>Latency with respect to the transmission speed.</p> Full article ">Figure 19
<p>Delivery probability with respect to the buffer size.</p> Full article ">Figure 20
<p>Overhead ratio with respect to the buffer size.</p> Full article ">Figure 21
<p>Latency with respect to the buffer size.</p> Full article ">Figure 22
<p>Delivery probability with respect to the number of copies.</p> Full article ">Figure 23
<p>Overhead ratio with respect to the number of copies.</p> Full article ">Figure 24
<p>Latency with respect to the number of copies.</p> Full article ">Figure 25
<p>Ping–pong application probability results.</p> Full article ">Figure 26
<p>Delivery probability WiFi-Direct case.</p> Full article ">Figure 27
<p>Overhead ratio WiFi-Direct case.</p> Full article ">Figure 28
<p>Latency WiFi-Direct case.</p> Full article ">
17 pages, 855 KiB  
Article
Automated Identification of Sensitive Financial Data Based on the Topic Analysis
by Meng Li, Jiqiang Liu and Yeping Yang
Future Internet 2024, 16(2), 55; https://doi.org/10.3390/fi16020055 - 8 Feb 2024
Viewed by 1440
Abstract
Data governance is an extremely important protection and management measure throughout the entire life cycle of data. However, there are still data governance issues, such as data security risks, data privacy breaches, and difficulties in data management and access control. These problems lead [...] Read more.
Data governance is an extremely important protection and management measure throughout the entire life cycle of data. However, there are still data governance issues, such as data security risks, data privacy breaches, and difficulties in data management and access control. These problems lead to a risk of data breaches and abuse. Therefore, the security classification and grading of data has become an important task to accurately identify sensitive data and adopt appropriate maintenance and management measures with different sensitivity levels. This work started from the problems existing in the current data security classification and grading work, such as inconsistent classification and grading standards, difficult data acquisition and sorting, and weak semantic information of data fields, to find the limitations of the current methods and the direction for improvement. The automatic identification method of sensitive financial data proposed in this paper is based on topic analysis and was constructed by incorporating Jieba word segmentation, word frequency statistics, the skip-gram model, K-means clustering, and other technologies. Expert assistance was sought to select appropriate keywords for enhanced accuracy. This work used the descriptive text library and real business data of a Chinese financial institution for training and testing to further demonstrate its effectiveness and usefulness. The evaluation indicators illustrated the effectiveness of this method in the classification of data security. The proposed method addressed the challenge of sensitivity level division in texts with limited semantic information, which overcame the limitations on model expansion across different domains and provided an optimized application model. All of the above pointed out the direction for the real-time updating of the method. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Overall architecture of data security classification.</p> Full article ">Figure 2
<p>Accuracy rate changes with the number of iterations in the process of model fitting.</p> Full article ">Figure 3
<p>Changes in sum squared errors with the number of iterations in the process of model fitting.</p> Full article ">Figure 4
<p>Confusion matrix of the model on the training set.</p> Full article ">Figure 5
<p>Intra-cluster average distance of each category.</p> Full article ">Figure 6
<p>Comparison of the intra-cluster average distance and the inter-cluster average distance.</p> Full article ">Figure 7
<p>Confusion matrix of the model on test set 1.</p> Full article ">Figure 8
<p>Confusion matrix of the model on test set 2.</p> Full article ">Figure 9
<p>Top 40 frequency distribution of words.</p> Full article ">
3 pages, 155 KiB  
Editorial
Modern Trends in Multi-Agent Systems
by Martin Kenyeres, Ivana Budinská, Ladislav Hluchý and Agostino Poggi
Future Internet 2024, 16(2), 54; https://doi.org/10.3390/fi16020054 - 8 Feb 2024
Viewed by 2602
Abstract
The term “multi-agent system” is generally understood as an interconnected set of independent entities that can effectively solve complex and time-consuming problems exceeding the individual abilities of common problem solvers [...] Full article
(This article belongs to the Special Issue Modern Trends in Multi-Agent Systems)
4 pages, 141 KiB  
Editorial
State-of-the-Art Future Internet Technology in Italy 2022–2023
by Massimo Cafaro, Italo Epicoco and Marco Pulimeno
Future Internet 2024, 16(2), 53; https://doi.org/10.3390/fi16020053 - 6 Feb 2024
Viewed by 1462
Abstract
This Special Issue aims to provide a comprehensive overview of the current state of the art in Future Internet Technology in Italy [...] Full article
(This article belongs to the Special Issue State-of-the-Art Future Internet Technology in Italy 2022–2023)
22 pages, 1501 KiB  
Article
A QoS-Aware IoT Edge Network for Mobile Telemedicine Enabling In-Transit Monitoring of Emergency Patients
by Adwitiya Mukhopadhyay, Aryadevi Remanidevi Devidas, Venkat P. Rangan and Maneesha Vinodini Ramesh
Future Internet 2024, 16(2), 52; https://doi.org/10.3390/fi16020052 - 6 Feb 2024
Cited by 1 | Viewed by 2108
Abstract
Addressing the inadequacy of medical facilities in rural communities and the high number of patients affected by ailments that need to be treated immediately is of prime importance for all countries. The various recent healthcare emergency situations bring out the importance of telemedicine [...] Read more.
Addressing the inadequacy of medical facilities in rural communities and the high number of patients affected by ailments that need to be treated immediately is of prime importance for all countries. The various recent healthcare emergency situations bring out the importance of telemedicine and demand rapid transportation of patients to nearby hospitals with available resources to provide the required medical care. Many current healthcare facilities and ambulances are not equipped to provide real-time risk assessment for each patient and dynamically provide the required medical interventions. This work proposes an IoT-based mobile medical edge (IM2E) node to be integrated with wearable and portable devices for the continuous monitoring of emergency patients transported via ambulances and it delves deeper into the existing challenges, such as (a) a lack of a simplified patient risk scoring system, (b) the need for architecture that enables seamless communication for dynamically varying QoS requirements, and (c)the need for context-aware knowledge regarding the effect of end-to-end delay and the packet loss ratio (PLR) on the real-time monitoring of health risks in emergency patients. The proposed work builds a data path selection model to identify the most effective path through which to route the data packets in an effective manner. The signal-to-noise interference ratio and the fading in the path are chosen to analyze the suitable path for data transmission. Full article
(This article belongs to the Special Issue Novel 5G Deployment Experience and Performance Results)
►▼ Show Figures

Figure 1

Figure 1
<p>IoT-Enabled Traffic-Aware Telemedicine Architecture (ITTA).</p> Full article ">Figure 2
<p>IoT-based Telemedicine Services Framework (ITSF).</p> Full article ">Figure 3
<p>Representation of Scenario 1.</p> Full article ">Figure 4
<p>Representation of Scenario 2.</p> Full article ">Figure 5
<p>Packet loss ratio (PLR) experienced by each data type in Scenario 1.</p> Full article ">Figure 6
<p>Delay experienced by each data type in Scenario 1.</p> Full article ">Figure 7
<p>Packet loss ratio (PLR) experienced by each data type in Scenario 2.</p> Full article ">Figure 8
<p>Delay experienced by each data type in Scenario 2.</p> Full article ">
15 pages, 2026 KiB  
Article
Optimizing Session-Aware Recommenders: A Deep Dive into GRU-Based Latent Interaction Integration
by Ming-Yen Lin, Ping-Chun Wu and Sue-Chen Hsueh
Future Internet 2024, 16(2), 51; https://doi.org/10.3390/fi16020051 - 1 Feb 2024
Viewed by 1778
Abstract
This study introduces session-aware recommendation models, leveraging GRU (Gated Recurrent Unit) and attention mechanisms for advanced latent interaction data integration. A primary advancement is enhancing latent context, a critical factor for boosting recommendation accuracy. We address the existing models’ rigidity by dynamically blending [...] Read more.
This study introduces session-aware recommendation models, leveraging GRU (Gated Recurrent Unit) and attention mechanisms for advanced latent interaction data integration. A primary advancement is enhancing latent context, a critical factor for boosting recommendation accuracy. We address the existing models’ rigidity by dynamically blending short-term (most recent) and long-term (historical) preferences, moving beyond static period definitions. Our approaches, pre-combination (LCII-Pre) and post-combination (LCII-Post), with fixed (Fix) and flexible learning (LP) weight configurations, are thoroughly evaluated. We conducted extensive experiments to assess our models’ performance on public datasets such as Amazon and MovieLens 1M. Notably, on the MovieLens 1M dataset, LCII-PreFix achieved a 1.85% and 2.54% higher Recall@20 than II-RNN and BERT4Rec+st+TSA, respectively. On the Steam dataset, LCII-PostLP outperformed these models by 18.66% and 5.5%. Furthermore, on the Amazon dataset, LCII showed a 2.59% and 1.89% improvement in Recall@20 over II-RNN and CAII. These results affirm the significant enhancement our models bring to session-aware recommendation systems, showcasing their potential for both academic and practical applications in the field. Full article
(This article belongs to the Special Issue Deep Learning in Recommender Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Long-term vs. short-term preferences by varying window scope.</p> Full article ">Figure 2
<p>LCII Architecture.</p> Full article ">Figure 3
<p>Representation Fusion module—LCII-Pre.</p> Full article ">Figure 4
<p>Representation Fusion module—LCII-Post.</p> Full article ">Figure 5
<p>Representation Fusion module—LCII.</p> Full article ">
14 pages, 3418 KiB  
Article
Enhancing Smart City Safety and Utilizing AI Expert Systems for Violence Detection
by Pradeep Kumar, Guo-Liang Shih, Bo-Lin Guo, Siva Kumar Nagi, Yibeltal Chanie Manie, Cheng-Kai Yao, Michael Augustine Arockiyadoss and Peng-Chun Peng
Future Internet 2024, 16(2), 50; https://doi.org/10.3390/fi16020050 - 31 Jan 2024
Cited by 3 | Viewed by 3003
Abstract
Violent attacks have been one of the hot issues in recent years. In the presence of closed-circuit televisions (CCTVs) in smart cities, there is an emerging challenge in apprehending criminals, leading to a need for innovative solutions. In this paper, the propose a [...] Read more.
Violent attacks have been one of the hot issues in recent years. In the presence of closed-circuit televisions (CCTVs) in smart cities, there is an emerging challenge in apprehending criminals, leading to a need for innovative solutions. In this paper, the propose a model aimed at enhancing real-time emergency response capabilities and swiftly identifying criminals. This initiative aims to foster a safer environment and better manage criminal activity within smart cities. The proposed architecture combines an image-to-image stable diffusion model with violence detection and pose estimation approaches. The diffusion model generates synthetic data while the object detection approach uses YOLO v7 to identify violent objects like baseball bats, knives, and pistols, complemented by MediaPipe for action detection. Further, a long short-term memory (LSTM) network classifies the action attacks involving violent objects. Subsequently, an ensemble consisting of an edge device and the entire proposed model is deployed onto the edge device for real-time data testing using a dash camera. Thus, this study can handle violent attacks and send alerts in emergencies. As a result, our proposed YOLO model achieves a mean average precision (MAP) of 89.5% for violent attack detection, and the LSTM classifier model achieves an accuracy of 88.33% for violent action classification. The results highlight the model’s enhanced capability to accurately detect violent objects, particularly in effectively identifying violence through the implemented artificial intelligence system. Full article
(This article belongs to the Special Issue Challenges in Real-Time Intelligent Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>The proposed model of attack alerting system—setup involving an image-to-image generation model, object detection model (YOLO v7), pose estimation model (MediaPipe), and action classification using the LSTM model.</p> Full article ">Figure 2
<p>Generated synthetic image samples of image-to-image stable diffusion.</p> Full article ">Figure 3
<p>LSTM model architecture.</p> Full article ">Figure 4
<p>The proposed YOLO v7 model—real-time predicted results in different attack-type conditions: (<b>a</b>,<b>d</b>) show a detected violent object, a baseball bat; (<b>b</b>) shows a detected violent object, a gun; (<b>c</b>,<b>e</b>) show a detected violent object, a knife.</p> Full article ">Figure 5
<p>Violent attack pose estimation using MediaPipe: (<b>a</b>,<b>d</b>) show an attacking action with a baseball bat; (<b>b</b>,<b>e</b>) show an attacking action with a knife; (<b>c</b>) show an attacking action with a gun.</p> Full article ">Figure 6
<p>(<b>a</b>) The confusion matrix; (<b>b</b>) the precision recall curve for attack types or divergent classes.</p> Full article ">Figure 7
<p>The YOLO v7 main metrics of the training and validation loss and accuracy of the model.</p> Full article ">Figure 8
<p>Performance of violence action classification in MediaPipe using LSTM.</p> Full article ">
19 pages, 3392 KiB  
Article
A New Dynamic Game-Based Pricing Model for Cloud Environment
by Hamid Saadatfar, Hamid Gholampour Ahangar and Javad Hassannataj Joloudari
Future Internet 2024, 16(2), 49; https://doi.org/10.3390/fi16020049 - 31 Jan 2024
Cited by 1 | Viewed by 1705
Abstract
Resource pricing in cloud computing has become one of the main challenges for cloud providers. The challenge is determining a fair and appropriate price to satisfy users and resource providers. To establish a justifiable price, it is imperative to take into account the [...] Read more.
Resource pricing in cloud computing has become one of the main challenges for cloud providers. The challenge is determining a fair and appropriate price to satisfy users and resource providers. To establish a justifiable price, it is imperative to take into account the circumstances and requirements of both the provider and the user. This research tries to provide a pricing mechanism for cloud computing based on game theory. The suggested approach considers three aspects: the likelihood of faults, the interplay among virtual machines, and the amount of energy used, in order to determine a justifiable price. In the game that is being proposed, the provider is responsible for determining the price of the virtual machine that can be made available to the user on each physical machine. The user, on the other hand, has the authority to choose between the virtual machines that are offered in order to run their application. The whole game is implemented as a function of the resource broker component. The proposed mechanism is simulated and evaluated using the CloudSim simulator. Its performance is compared with several previous recent mechanisms. The results indicate that the suggested mechanism has successfully identified a more rational price for both the user and the provider, consequently enhancing the overall profitability of the cloud system. Full article
(This article belongs to the Special Issue Cloud Computing and High Performance Computing (HPC) Advances for Next Generation Internet)
►▼ Show Figures

Figure 1

Figure 1
<p>The architecture of cloud computing system.</p> Full article ">Figure 2
<p>The relationship between price and demand.</p> Full article ">Figure 3
<p>The tree of the proposed game.</p> Full article ">Figure 4
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] from the viewpoint of average profit per user.</p> Full article ">Figure 5
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] from the viewpoint of the provider’s total revenue.</p> Full article ">Figure 6
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] from the viewpoint of the provider’s total profit.</p> Full article ">Figure 7
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] from the viewpoint of the number of users’ requests that can be met.</p> Full article ">Figure 8
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] under different system loads from the viewpoint of average user’s profit.</p> Full article ">Figure 9
<p>Comparison of the proposed method with [<a href="#B16-futureinternet-16-00049" class="html-bibr">16</a>,<a href="#B39-futureinternet-16-00049" class="html-bibr">39</a>] under different system loads from the viewpoint of the provider’s total profit.</p> Full article ">
28 pages, 8697 KiB  
Article
Efficient Privacy-Aware Forwarding for Enhanced Communication Privacy in Opportunistic Mobile Social Networks
by Azizah Assiri and Hassen Sallay
Future Internet 2024, 16(2), 48; https://doi.org/10.3390/fi16020048 - 31 Jan 2024
Viewed by 1652
Abstract
Opportunistic mobile social networks (OMSNs) have become increasingly popular in recent years due to the rise of social media and smartphones. However, message forwarding and sharing social information through intermediary nodes on OMSNs raises privacy concerns as personal data and activities become more [...] Read more.
Opportunistic mobile social networks (OMSNs) have become increasingly popular in recent years due to the rise of social media and smartphones. However, message forwarding and sharing social information through intermediary nodes on OMSNs raises privacy concerns as personal data and activities become more exposed. Therefore, maintaining privacy without limiting efficient social interaction is a challenging task. This paper addresses this specific problem of safeguarding user privacy during message forwarding by integrating a privacy layer on the state-of-the-art OMSN routing decision models that empowers users to control their message dissemination. Mainly, we present three user-centric privacy-aware forwarding modes guiding the selection of the next hop in the forwarding path based on social metrics such as common friends and exchanged messages between OMSN nodes. More specifically, we define different social relationship strengths approximating real-world scenarios (familiar, weak tie, stranger) and trust thresholds to give users choices on trust levels for different social contexts and guide the routing decisions. We evaluate the privacy enhancement and network performance through extensive simulations using ONE simulator for several routing schemes (Epidemic, Prophet, and Spray and Wait) and different movement models (random way, bus, and working day). We demonstrate that our modes can enhance privacy by up to 45% in various network scenarios, as measured by the reduction in the likelihood of unintended message propagation, while keeping the message-delivery process effective and efficient. 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>System model overview.</p> Full article ">Figure 2
<p>Privacy modes’ flowcharts (<b>a</b>) for the trust threshold-based mode, (<b>b</b>) for stranger then familiar selection mode, and (<b>c</b>) for familiar then stranger selection mode.</p> Full article ">Figure 3
<p>The result of trust threshold 30% and random way movement model.</p> Full article ">Figure 4
<p>The result of trust threshold 30% and the bus movement model.</p> Full article ">Figure 5
<p>The result of trust thresholds 30% and working day movement model.</p> Full article ">Figure 6
<p>The results of thresholds 50% and random way model.</p> Full article ">Figure 7
<p>The result of thresholds 50 and bus model.</p> Full article ">Figure 8
<p>The result of thresholds 50 and working day.</p> Full article ">Figure 9
<p>The result of thresholds 80% and the random way model.</p> Full article ">Figure 10
<p>The result of thresholds 80 and bus model.</p> Full article ">Figure 11
<p>The result of thresholds 80% and the working day model.</p> Full article ">Figure 12
<p>The result of stranger mode with random way model.</p> Full article ">Figure 13
<p>The result of stranger mode with bus model.</p> Full article ">Figure 14
<p>The result of stranger mode with working day model.</p> Full article ">Figure 15
<p>The result of familiar mode with the random way model.</p> Full article ">Figure 16
<p>The result of familiar mode with bus model.</p> Full article ">Figure 17
<p>The result of familiar mode with working day model.</p> Full article ">Figure 18
<p>The result of the data collection process in the random way model.</p> Full article ">Figure 19
<p>The result of the data collection process with the bus model.</p> Full article ">Figure 20
<p>The result data collection process with the working day model.</p> Full article ">
44 pages, 38595 KiB  
Article
Enhancing Urban Resilience: Smart City Data Analyses, Forecasts, and Digital Twin Techniques at the Neighborhood Level
by Andreas F. Gkontzis, Sotiris Kotsiantis, Georgios Feretzakis and Vassilios S. Verykios
Future Internet 2024, 16(2), 47; https://doi.org/10.3390/fi16020047 - 30 Jan 2024
Cited by 11 | Viewed by 5802
Abstract
Smart cities, leveraging advanced data analytics, predictive models, and digital twin techniques, offer a transformative model for sustainable urban development. Predictive analytics is critical to proactive planning, enabling cities to adapt to evolving challenges. Concurrently, digital twin techniques provide a virtual replica of [...] Read more.
Smart cities, leveraging advanced data analytics, predictive models, and digital twin techniques, offer a transformative model for sustainable urban development. Predictive analytics is critical to proactive planning, enabling cities to adapt to evolving challenges. Concurrently, digital twin techniques provide a virtual replica of the urban environment, fostering real-time monitoring, simulation, and analysis of urban systems. This study underscores the significance of real-time monitoring, simulation, and analysis of urban systems to support test scenarios that identify bottlenecks and enhance smart city efficiency. This paper delves into the crucial roles of citizen report analytics, prediction, and digital twin technologies at the neighborhood level. The study integrates extract, transform, load (ETL) processes, artificial intelligence (AI) techniques, and a digital twin methodology to process and interpret urban data streams derived from citizen interactions with the city’s coordinate-based problem mapping platform. Using an interactive GeoDataFrame within the digital twin methodology, dynamic entities facilitate simulations based on various scenarios, allowing users to visualize, analyze, and predict the response of the urban system at the neighborhood level. This approach reveals antecedent and predictive patterns, trends, and correlations at the physical level of each city area, leading to improvements in urban functionality, resilience, and resident quality of life. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The sample report of a unique citizen in JSON format.</p> Full article ">Figure 2
<p>The distribution of the classes.</p> Full article ">Figure 3
<p>The performance of the trained classifiers.</p> Full article ">Figure 4
<p>Areas with higher counts are shown in darker shades of color.</p> Full article ">Figure 5
<p>The percentage of issues by region.</p> Full article ">Figure 6
<p>The total number of issues in each area over time.</p> Full article ">Figure 7
<p>Boxplots for the “Latitude”, “Longitude”, “Areas_int”, and “issue_int” columns.</p> Full article ">Figure 8
<p>This scatterplot for “Latitude” versus “Longitude” with “Areas_int” and “issue_int”.</p> Full article ">Figure 9
<p>Histograms of data.</p> Full article ">Figure 10
<p>The city of Patras in Greece.</p> Full article ">Figure 11
<p>The spread of the reported issues in the city in the last 5 years.</p> Full article ">Figure 12
<p>Issue clustering and information pop-up tag.</p> Full article ">Figure 13
<p>Filtering the frequency of issues and color clustering.</p> Full article ">Figure 14
<p>Issue probability for the next six months in each city neighborhood.</p> Full article ">Figure 15
<p>The digital twin app numbers the issues and colors the most frequent issue in region.</p> Full article ">Figure 16
<p>The digital twin app, an interactive service.</p> Full article ">Figure 17
<p>The digital twin app zooms to the lowest level of the neighborhood.</p> Full article ">Figure 18
<p>The digital twin app’s updated version includes layers, filters, and charts.</p> Full article ">Figure 19
<p>The digital twin app can display forecasts, issues, and area counts for each reported issue by area.</p> Full article ">Figure A1
<p>The digital twin app, update charts, issue clusters, and areas (filter issue: lighting).</p> Full article ">Figure A2
<p>The digital twin app, update charts, issue clusters, and areas (filter issue: lighting).</p> Full article ">Figure A3
<p>Digital twin app, update charts, issue clusters, and areas (filter issue: road-constructor).</p> Full article ">Figure A4
<p>The digital twin app, update charts, issue clusters, and areas (filter issue: green).</p> Full article ">Figure A5
<p>Digital twin app, update charts, issue clusters, and areas (filter issue: garbage).</p> Full article ">Figure A6
<p>The digital twin app, update charts, issue clusters, and areas (filter issue: plumbing).</p> Full article ">Figure A7
<p>Digital twin app, update charts, issue clusters, and areas (filter issue: protection policy).</p> Full article ">Figure A8
<p>Digital twin app, update charts, issue clusters, and areas (filter issue: environment).</p> Full article ">Figure A9
<p>Digital twin app, update charts, issue clusters, and areas (filter issue: parking).</p> Full article ">Figure A10
<p>Digital twin app, update charts, issue clusters, and areas (filter: years).</p> Full article ">Figure A11
<p>Digital twin app, update charts, issue clusters, and areas (filter years: 2023).</p> Full article ">Figure A12
<p>The digital twin app, update charts, issue clusters, and areas (filter: areas).</p> Full article ">Figure A13
<p>Digital twin app, update charts, issue clusters, and areas (filter areas: Skagiopoulio).</p> Full article ">Figure A14
<p>Digital twin app, update charts, issue clusters, and areas (filter areas: Skagiopoulio).</p> Full article ">Figure A15
<p>The digital twin app, update charts, issue clusters, and areas (filter: probabilities).</p> Full article ">Figure A16
<p>Digital twin app, update charts, issue clusters, and areas (filter probabilities: 95%).</p> Full article ">Figure A17
<p>Digital twin app, update charts, issue clusters, and areas (filter probabilities: 93%).</p> Full article ">Figure A18
<p>Digital twin app, update charts, issue clusters, and areas (filter probabilities: 88%).</p> Full article ">Figure A19
<p>Digital Twin app, update charts, issue clusters, and areas (two filters: area, probabilities).</p> Full article ">Figure A20
<p>The digital twin app, update charts, issue clusters, and areas (three filters: year, area, and probabilities).</p> Full article ">Figure A21
<p>The digital twin app, update charts, issue clusters, and areas (four filters: issue, year area, probabilities).</p> Full article ">
18 pages, 2156 KiB  
Article
Context-Aware Behavioral Tips to Improve Sleep Quality via Machine Learning and Large Language Models
by Erica Corda, Silvia M. Massa and Daniele Riboni
Future Internet 2024, 16(2), 46; https://doi.org/10.3390/fi16020046 - 30 Jan 2024
Cited by 2 | Viewed by 2275
Abstract
As several studies demonstrate, good sleep quality is essential for individuals’ well-being, as a lack of restoring sleep may disrupt different physical, mental, and social dimensions of health. For this reason, there is increasing interest in tools for the monitoring of sleep based [...] Read more.
As several studies demonstrate, good sleep quality is essential for individuals’ well-being, as a lack of restoring sleep may disrupt different physical, mental, and social dimensions of health. For this reason, there is increasing interest in tools for the monitoring of sleep based on personal sensors. However, there are currently few context-aware methods to help individuals to improve their sleep quality through behavior change tips. In order to tackle this challenge, in this paper, we propose a system that couples machine learning algorithms and large language models to forecast the next night’s sleep quality, and to provide context-aware behavior change tips to improve sleep. In order to encourage adherence and to increase trust, our system includes the use of large language models to describe the conditions that the machine learning algorithm finds harmful to sleep health, and to explain why the behavior change tips are generated as a consequence. We develop a prototype of our system, including a smartphone application, and perform experiments with a set of users. Results show that our system’s forecast is correlated to the actual sleep quality. Moreover, a preliminary user study suggests that the use of large language models in our system is useful in increasing trust and engagement. Full article
(This article belongs to the Special Issue Explainability, Reliability and Trust in Smart Internet of Things Healthcare Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>System architecture.</p> Full article ">Figure 2
<p>A sample of the random tree model. Considering the feature vector values, the decision tree follows a path from the root node to a leaf node that represents the model’s forecast. The <span class="html-italic">decision path</span> is the sequence of feature conditions evaluated along the branches from the root to the specific leaf node, determining the prediction for a particular feature vector. Consider, as an example, a feature vector including the following values: ‘Previous Day Sleep Duration’ = 11 h; ‘Duration of Noise in 20–22’ = 10 min; ‘Duration of Walking’ = 5 min; ‘Duration of Locked in 18–20’ = 40 min; ‘Duration of Dark in 22–24’ = 85 min; ‘Duration of Conversation in 18–20’ = 0 min. In this case, the decision path is composed of the nodes highlighted in green, and the random tree sleep quality forecast value is <math display="inline"><semantics> <mrow> <mn>3.81</mn> </mrow> </semantics></math>. In each node, <span class="html-italic">samples</span> represents the number of instances of the training set that match the conditions of the node, the <span class="html-italic">value</span> represents the most probable value, and <span class="html-italic">mse</span> represents the corresponding mean squared error. The listing in the box below the tree shows the notation that we use to represent the decision path. We use this notation to communicate the decision path to the LLM-based modules.</p> Full article ">Figure 3
<p>A sample of the prompt issued to the generative pre-trained Transformer.</p> Full article ">Figure 4
<p>The generative pre-trained Transformer response to the prompt shown in <a href="#futureinternet-16-00046-f003" class="html-fig">Figure 3</a>.</p> Full article ">Figure 5
<p>A different sample of the prompt issued to the generative pre-trained Transformer.</p> Full article ">Figure 6
<p>The generative pre-trained Transformer response to the prompt shown in <a href="#futureinternet-16-00046-f005" class="html-fig">Figure 5</a>.</p> Full article ">Figure 7
<p>LLM-based explanations of the use case illustrated in <a href="#futureinternet-16-00046-f005" class="html-fig">Figure 5</a> and <a href="#futureinternet-16-00046-f006" class="html-fig">Figure 6</a>.</p> Full article ">Figure 8
<p>Smartphone app prediction.</p> Full article ">Figure 9
<p>Smartphone app explanation.</p> Full article ">Figure 10
<p>Results of the questionnaire.</p> Full article ">
23 pages, 2549 KiB  
Article
Non-Profiled Unsupervised Horizontal Iterative Attack against Hardware Elliptic Curve Scalar Multiplication Using Machine Learning
by Marcin Aftowicz, Ievgen Kabin, Zoya Dyka and Peter Langendörfer
Future Internet 2024, 16(2), 45; https://doi.org/10.3390/fi16020045 - 29 Jan 2024
Cited by 2 | Viewed by 1564
Abstract
While IoT technology makes industries, cities, and homes smarter, it also opens the door to security risks. With the right equipment and physical access to the devices, the attacker can leverage side-channel information, like timing, power consumption, or electromagnetic emanation, to compromise cryptographic [...] Read more.
While IoT technology makes industries, cities, and homes smarter, it also opens the door to security risks. With the right equipment and physical access to the devices, the attacker can leverage side-channel information, like timing, power consumption, or electromagnetic emanation, to compromise cryptographic operations and extract the secret key. This work presents a side channel analysis of a cryptographic hardware accelerator for the Elliptic Curve Scalar Multiplication operation, implemented in a Field-Programmable Gate Array and as an Application-Specific Integrated Circuit. The presented framework consists of initial key extraction using a state-of-the-art statistical horizontal attack and is followed by regularized Artificial Neural Networks, which take, as input, the partially incorrect key guesses from the horizontal attack and correct them iteratively. The initial correctness of the horizontal attack, measured as the fraction of correctly extracted bits of the secret key, was improved from 75% to 98% by applying the iterative learning. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Depiction of (<b>a</b>) Feed-Forward Neural Network architecture and (<b>b</b>) autoencoder architecture.</p> Full article ">Figure 2
<p>General structure of our hardware accelerator.</p> Full article ">Figure 3
<p>Geometrical representation of optimal hyperplane separating negative examples (black squares) from positive examples (white triangles).</p> Full article ">Figure 4
<p>Three cases of model performance: (<b>a</b>) underfitting, (<b>b</b>) a good bias–variance trade-off, and (<b>c</b>) overfitting, given training set samples (red X).</p> Full article ">Figure 5
<p>The general structure of the unsupervised framework.</p> Full article ">Figure 6
<p>Trend of (<b>a</b>) the cumulative hamming distance of all key candidates and (<b>b</b>) the correctness of the best key candidate throughout iterations of the framework trained using the ANN implemented in scikit-learn with L2 regularization and λ = 10.</p> Full article ">Figure 7
<p>The correctness of each key candidate as an indication of leakage in a given clock cycle, for initial horizontal attack (blue) and last iteration of the scikit-learn implementation (orange) for (<b>a</b>) EMT of Design_seq, (<b>b</b>) EMT of Design_ultra, (<b>c</b>) PT of Design_seq, and (<b>d</b>) PT of Design_ultra.</p> Full article ">
14 pages, 2162 KiB  
Article
Computer Vision and Machine Learning-Based Predictive Analysis for Urban Agricultural Systems
by Arturs Kempelis, Inese Polaka, Andrejs Romanovs and Antons Patlins
Future Internet 2024, 16(2), 44; https://doi.org/10.3390/fi16020044 - 28 Jan 2024
Cited by 3 | Viewed by 2425
Abstract
Urban agriculture presents unique challenges, particularly in the context of microclimate monitoring, which is increasingly important in food production. This paper explores the application of convolutional neural networks (CNNs) to forecast key sensor measurements from thermal images within this context. This research focuses [...] Read more.
Urban agriculture presents unique challenges, particularly in the context of microclimate monitoring, which is increasingly important in food production. This paper explores the application of convolutional neural networks (CNNs) to forecast key sensor measurements from thermal images within this context. This research focuses on using thermal images to forecast sensor measurements of relative air humidity, soil moisture, and light intensity, which are integral to plant health and productivity in urban farming environments. The results indicate a higher accuracy in forecasting relative air humidity and soil moisture levels, with Mean Absolute Percentage Errors (MAPEs) within the range of 10–12%. These findings correlate with the strong dependency of these parameters on thermal patterns, which are effectively extracted by the CNNs. In contrast, the forecasting of light intensity proved to be more challenging, yielding lower accuracy. The reduced performance is likely due to the more complex and variable factors that affect light in urban environments. The insights gained from the higher predictive accuracy for relative air humidity and soil moisture may inform targeted interventions for urban farming practices, while the lower accuracy in light intensity forecasting highlights the need for further research into the integration of additional data sources or hybrid modeling approaches. The conclusion suggests that the integration of these technologies can significantly enhance the predictive maintenance of plant health, leading to more sustainable and efficient urban farming practices. However, the study also acknowledges the challenges in implementing these technologies in urban agricultural models. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Scheme of proposed model architecture for sensor measurement forecasting.</p> Full article ">Figure 2
<p>The process of model training.</p> Full article ">Figure 3
<p>Data gathering location and spatial characteristics: (<b>a</b>) Front view of the greenhouse and (<b>b</b>) top-down view of the greenhouse.</p> Full article ">Figure 4
<p>Different sensor correlations with thermal image pixels during all data gathering period. Correlation heatmap between pixel values and (<b>a</b>) air temperature sensor data; (<b>b</b>) air humidity sensor data; (<b>c</b>) soil water content sensor data; (<b>d</b>) light intensity sensor data.</p> Full article ">
17 pages, 851 KiB  
Article
A Spectral Gap-Based Topology Control Algorithm for Wireless Backhaul Networks
by Sergio Jesús González-Ambriz, Rolando Menchaca-Méndez, Sergio Alejandro Pinacho-Castellanos and Mario Eduardo Rivero-Ángeles 
Future Internet 2024, 16(2), 43; https://doi.org/10.3390/fi16020043 - 26 Jan 2024
Viewed by 2283
Abstract
This paper presents the spectral gap-based topology control algorithm (SGTC) for wireless backhaul networks, a novel approach that employs the Laplacian Spectral Gap (LSG) to find expander-like graphs that optimize the topology of the network in terms of robustness, diameter, energy cost, and [...] Read more.
This paper presents the spectral gap-based topology control algorithm (SGTC) for wireless backhaul networks, a novel approach that employs the Laplacian Spectral Gap (LSG) to find expander-like graphs that optimize the topology of the network in terms of robustness, diameter, energy cost, and network entropy. The latter measures the network’s ability to promote seamless traffic offloading from the Macro Base Stations to smaller cells by providing a high diversity of shortest paths connecting all the stations. Given the practical constraints imposed by cellular technologies, the proposed algorithm uses simulated annealing to search for feasible network topologies with a large LSG. Then, it computes the Pareto front of the set of feasible solutions found during the annealing process when considering robustness, diameter, and entropy as objective functions. The algorithm’s result is the Pareto efficient solution that minimizes energy cost. A set of experimental results shows that by optimizing the LSG, the proposed algorithm simultaneously optimizes the set of desirable topological properties mentioned above. The results also revealed that generating networks with good spectral expansion is possible even under the restrictions imposed by current wireless technologies. This is a desirable feature because these networks have strong connectivity properties even if they do not have a large number of links. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Experimental workflow.</p> Full article ">Figure 2
<p>Network entropy <math display="inline"><semantics> <msub> <mi>S</mi> <mi>G</mi> </msub> </semantics></math> with increasing Laplacian Spectral Gap (LSG, <math display="inline"><semantics> <msub> <mi>λ</mi> <mn>2</mn> </msub> </semantics></math>).</p> Full article ">Figure 3
<p>Network diameter <math display="inline"><semantics> <msub> <mi>D</mi> <mi>G</mi> </msub> </semantics></math> with increasing Laplacian Spectral Gap (LSG, <math display="inline"><semantics> <msub> <mi>λ</mi> <mn>2</mn> </msub> </semantics></math>).</p> Full article ">Figure 4
<p>Network robustness with increasing Laplacian Spectral Gap (LSG, <math display="inline"><semantics> <msub> <mi>λ</mi> <mn>2</mn> </msub> </semantics></math>).</p> Full article ">Figure 5
<p>Network cost with increasing Laplacian Spectral Gap (LSG, <math display="inline"><semantics> <msub> <mi>λ</mi> <mn>2</mn> </msub> </semantics></math>).</p> Full article ">Figure 6
<p>Network entropy <math display="inline"><semantics> <msub> <mi>S</mi> <mi>G</mi> </msub> </semantics></math> with increasing network size.</p> Full article ">Figure 7
<p>Network diameter <math display="inline"><semantics> <msub> <mi>D</mi> <mi>G</mi> </msub> </semantics></math> with increasing network size.</p> Full article ">Figure 8
<p>Network robustness <math display="inline"><semantics> <msub> <mi>R</mi> <mi>G</mi> </msub> </semantics></math> with increasing network size.</p> Full article ">Figure 9
<p>Network cost <math display="inline"><semantics> <msub> <mi>C</mi> <mi>G</mi> </msub> </semantics></math> with increasing network size.</p> Full article ">
29 pages, 743 KiB  
Article
TinyML Algorithms for Big Data Management in Large-Scale IoT Systems
by Aristeidis Karras, Anastasios Giannaros, Christos Karras, Leonidas Theodorakopoulos, Constantinos S. Mammassis, George A. Krimpas and Spyros Sioutas
Future Internet 2024, 16(2), 42; https://doi.org/10.3390/fi16020042 - 25 Jan 2024
Cited by 7 | Viewed by 3474
Abstract
In the context of the Internet of Things (IoT), Tiny Machine Learning (TinyML) and Big Data, enhanced by Edge Artificial Intelligence, are essential for effectively managing the extensive data produced by numerous connected devices. Our study introduces a set of TinyML algorithms designed [...] Read more.
In the context of the Internet of Things (IoT), Tiny Machine Learning (TinyML) and Big Data, enhanced by Edge Artificial Intelligence, are essential for effectively managing the extensive data produced by numerous connected devices. Our study introduces a set of TinyML algorithms designed and developed to improve Big Data management in large-scale IoT systems. These algorithms, named TinyCleanEDF, EdgeClusterML, CompressEdgeML, CacheEdgeML, and TinyHybridSenseQ, operate together to enhance data processing, storage, and quality control in IoT networks, utilizing the capabilities of Edge AI. In particular, TinyCleanEDF applies federated learning for Edge-based data cleaning and anomaly detection. EdgeClusterML combines reinforcement learning with self-organizing maps for effective data clustering. CompressEdgeML uses neural networks for adaptive data compression. CacheEdgeML employs predictive analytics for smart data caching, and TinyHybridSenseQ concentrates on data quality evaluation and hybrid storage strategies. Our experimental evaluation of the proposed techniques includes executing all the algorithms in various numbers of Raspberry Pi devices ranging from one to ten. The experimental results are promising as we outperform similar methods across various evaluation metrics. Ultimately, we anticipate that the proposed algorithms offer a comprehensive and efficient approach to managing the complexities of IoT, Big Data, and Edge AI. Full article
(This article belongs to the Special Issue Internet of Things and Cyber-Physical Systems II)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed system architecture.</p> Full article ">Figure 2
<p>Performance evaluation of TinyCleanEDF.</p> Full article ">Figure 3
<p>Performance evaluation of EdgeClusterML.</p> Full article ">Figure 4
<p>Performance evaluation of CompressEdgeML.</p> Full article ">Figure 5
<p>Performance evaluation of CacheEdgeML.</p> Full article ">Figure 6
<p>Performance evaluation of TinyHybridSenseQ.</p> Full article ">Figure 7
<p>Cache hit rate comparison of CacheEdgeML with similar methods.</p> Full article ">Figure 8
<p>Compression efficiency of CompressEdgeML compared to similar method.</p> Full article ">Figure 9
<p>Compression speed of CompressEdgeML compared to similar method.</p> Full article ">
31 pages, 5283 KiB  
Article
Beyond Lexical Boundaries: LLM-Generated Text Detection for Romanian Digital Libraries
by Melania Nitu and Mihai Dascalu
Future Internet 2024, 16(2), 41; https://doi.org/10.3390/fi16020041 - 25 Jan 2024
Cited by 4 | Viewed by 2578
Abstract
Machine-generated content reshapes the landscape of digital information; hence, ensuring the authenticity of texts within digital libraries has become a paramount concern. This work introduces a corpus of approximately 60 k Romanian documents, including human-written samples as well as generated texts using six [...] Read more.
Machine-generated content reshapes the landscape of digital information; hence, ensuring the authenticity of texts within digital libraries has become a paramount concern. This work introduces a corpus of approximately 60 k Romanian documents, including human-written samples as well as generated texts using six distinct Large Language Models (LLMs) and three different generation methods. Our robust experimental dataset covers five domains, namely books, news, legal, medical, and scientific publications. The exploratory text analysis revealed differences between human-authored and artificially generated texts, exposing the intricacies of lexical diversity and textual complexity. Since Romanian is a less-resourced language requiring dedicated detectors on which out-of-the-box solutions do not work, this paper introduces two techniques for discerning machine-generated texts. The first method leverages a Transformer-based model to categorize texts as human or machine-generated, while the second method extracts and examines linguistic features, such as identifying the top textual complexity indices via Kruskal–Wallis mean rank and computes burstiness, which are further fed into a machine-learning model leveraging an extreme gradient-boosting decision tree. The methods show competitive performance, with the first technique’s results outperforming the second one in two out of five domains, reaching an F1 score of 0.96. Our study also includes a text similarity analysis between human-authored and artificially generated texts, coupled with a SHAP analysis to understand which linguistic features contribute more to the classifier’s decision. Full article
(This article belongs to the Special Issue Digital Analysis in Digital Humanities)
►▼ Show Figures

Figure 1

Figure 1
<p>Detection methods using RoBERT [<a href="#B58-futureinternet-16-00041" class="html-bibr">58</a>] and XGBoost [<a href="#B59-futureinternet-16-00041" class="html-bibr">59</a>] models.</p> Full article ">Figure 2
<p>Transformer-based model: confusion matrix for the books domain.</p> Full article ">Figure 3
<p>Classic ML-based model: SHAP summary plot for the books domain.</p> Full article ">Figure A1
<p>Transformer-based model: confusion matrices per domains.</p> Full article ">Figure A2
<p>Classic ML-based model: SHAP summary plot per domain.</p> Full article ">Figure A3
<p>Text similarity metrics for books domain.</p> Full article ">Figure A4
<p>Text similarity metrics for news domain.</p> Full article ">Figure A5
<p>Text similarity metrics for medical domain.</p> Full article ">Figure A6
<p>Text similarity metrics for Legal domain.</p> Full article ">Figure A7
<p>Text similarity metrics for RoCHI domain.</p> Full article ">Figure A8
<p>Pairwise Wilcoxon test by model per domain.</p> Full article ">Figure A9
<p>Top three RBIs for books domain.</p> Full article ">Figure A10
<p>Top three RBIs for legal domain.</p> Full article ">Figure A11
<p>Top three RBIs for medical domain.</p> Full article ">Figure A12
<p>Top 3 RBIs for news domain.</p> Full article ">Figure A13
<p>Top three RBIs for RoCHI domain.</p> Full article ">
57 pages, 2070 KiB  
Review
A Holistic Analysis of Internet of Things (IoT) Security: Principles, Practices, and New Perspectives
by Mahmud Hossain, Golam Kayas, Ragib Hasan, Anthony Skjellum, Shahid Noor and S. M. Riazul Islam
Future Internet 2024, 16(2), 40; https://doi.org/10.3390/fi16020040 - 24 Jan 2024
Cited by 9 | Viewed by 8161
Abstract
Driven by the rapid escalation of its utilization, as well as ramping commercialization, Internet of Things (IoT) devices increasingly face security threats. Apart from denial of service, privacy, and safety concerns, compromised devices can be used as enablers for committing a variety of [...] Read more.
Driven by the rapid escalation of its utilization, as well as ramping commercialization, Internet of Things (IoT) devices increasingly face security threats. Apart from denial of service, privacy, and safety concerns, compromised devices can be used as enablers for committing a variety of crime and e-crime. Despite ongoing research and study, there remains a significant gap in the thorough analysis of security challenges, feasible solutions, and open secure problems for IoT. To bridge this gap, we provide a comprehensive overview of the state of the art in IoT security with a critical investigation-based approach. This includes a detailed analysis of vulnerabilities in IoT-based systems and potential attacks. We present a holistic review of the security properties required to be adopted by IoT devices, applications, and services to mitigate IoT vulnerabilities and, thus, successful attacks. Moreover, we identify challenges to the design of security protocols for IoT systems in which constituent devices vary markedly in capability (such as storage, computation speed, hardware architecture, and communication interfaces). Next, we review existing research and feasible solutions for IoT security. We highlight a set of open problems not yet addressed among existing security solutions. We provide a set of new perspectives for future research on such issues including secure service discovery, on-device credential security, and network anomaly detection. We also provide directions for designing a forensic investigation framework for IoT infrastructures to inspect relevant criminal cases, execute a cyber forensic process, and determine the facts about a given incident. This framework offers a means to better capture information on successful attacks as part of a feedback mechanism to thwart future vulnerabilities and threats. This systematic holistic review will both inform on current challenges in IoT security and ideally motivate their future resolution. Full article
(This article belongs to the Special Issue Cyber Security in the New "Edge Computing + IoT" World)
►▼ Show Figures

Figure 1

Figure 1
<p>Security issues with commercialized IoT devices [<a href="#B9-futureinternet-16-00040" class="html-bibr">9</a>].</p> Full article ">Figure 2
<p>The organization and structure of this paper.</p> Full article ">Figure 3
<p>The operational model of an IoT system.</p> Full article ">Figure 4
<p>Communication interfaces in an IoT system.</p> Full article ">Figure 5
<p>External attack.</p> Full article ">Figure 6
<p>Internal attack.</p> Full article ">Figure 7
<p>Man-in-the-Middle attack.</p> Full article ">Figure 8
<p>Message modification.</p> Full article ">Figure 9
<p>Message fabrication.</p> Full article ">Figure 10
<p>Replay attack.</p> Full article ">Figure 11
<p>User credential compromise.</p> Full article ">Figure 12
<p>Software/hardware compromise threat model for Internet of Vehicles. RSU = Roadside Unit.</p> Full article ">Figure 13
<p>A taxonomy of IoT attacks.</p> Full article ">Figure 14
<p>Complexity parameters for security solutions.</p> Full article ">Figure 15
<p>Classification of the existing security schemes.</p> Full article ">Figure 16
<p>Internet stack vs. IoT stack [<a href="#B147-futureinternet-16-00040" class="html-bibr">147</a>].</p> Full article ">Figure 17
<p>Device hierarchy.</p> Full article ">Figure 18
<p>HIP-Based Exchange (HIP-BEX).</p> Full article ">Figure 19
<p>RBAC vs. CapBAC model.</p> Full article ">Figure 20
<p>Access-control approach. Policy Decision Point (PDP) is a service that makes authorization decisions. PDP is either offered by an external entity or embedded into an IoT device.</p> Full article ">Figure 21
<p>A model for secure service discovery.justification=justified.</p> Full article ">Figure 21 Cont.
<p>A model for secure service discovery.justification=justified.</p> Full article ">Figure 22
<p>An adversary profiles the movement of a target by tracking the static identifier of the target. RSU = Roadside Unit. VANET = Vehicular Ad Hoc Network.</p> Full article ">Figure 23
<p>An adversary tracks the identifiers of target devices for disrupting communications. T2T = Things to Things, T2C = Things to Clouds.</p> Full article ">Figure 24
<p>An overview of the Data Transparency Service.</p> Full article ">Figure 25
<p>Application data security during protocol translation.</p> Full article ">Figure 26
<p>Secure software update scheme.</p> Full article ">Figure 27
<p>A PUF-based scheme for storing credentials securely.</p> Full article ">Figure 28
<p>An adaptive anomaly-detection system. NMS = Network Monitoring System. ML = Machine Learning.</p> Full article ">Figure 29
<p>A Blockchain-based forensic framework.</p> Full article ">
23 pages, 958 KiB  
Systematic Review
Volumetric Techniques for Product Routing and Loading Optimisation in Industry 4.0: A Review
by Ricardo Lopes, Marcello Trovati and Ella Pereira
Future Internet 2024, 16(2), 39; https://doi.org/10.3390/fi16020039 - 24 Jan 2024
Viewed by 1834
Abstract
Industry 4.0 has become a crucial part in the majority of processes, components, and related modelling, as well as predictive tools that allow a more efficient, automated and sustainable approach to industry. The availability of large quantities of data, and the advances in [...] Read more.
Industry 4.0 has become a crucial part in the majority of processes, components, and related modelling, as well as predictive tools that allow a more efficient, automated and sustainable approach to industry. The availability of large quantities of data, and the advances in IoT, AI, and data-driven frameworks, have led to an enhanced data gathering, assessment, and extraction of actionable information, resulting in a better decision-making process. Product picking and its subsequent packing is an important area, and has drawn increasing attention for the research community. However, depending of the context, some of the related approaches tend to be either highly mathematical, or applied to a specific context. This article aims to provide a survey on the main methods, techniques, and frameworks relevant to product packing and to highlight the main properties and features that should be further investigated to ensure a more efficient and optimised approach. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>PRISMA workflow diagram.</p> Full article ">Figure 2
<p>A summary of the algorithmic approaches and the packing problems they address.</p> Full article ">Figure 3
<p>A summary of the technological solutions and the packing problems they address.</p> Full article ">Figure 4
<p>The degree distribution of the network extracted, as described in <a href="#sec6-futureinternet-16-00039" class="html-sec">Section 6</a>. Note that is appears to exhibit some level of scale-free topology, which is supported by the overall behaviour of textual information [<a href="#B53-futureinternet-16-00039" class="html-bibr">53</a>,<a href="#B54-futureinternet-16-00039" class="html-bibr">54</a>,<a href="#B55-futureinternet-16-00039" class="html-bibr">55</a>].</p> Full article ">
14 pages, 1090 KiB  
Article
Refined Semi-Supervised Modulation Classification: Integrating Consistency Regularization and Pseudo-Labeling Techniques
by Min Ma, Shanrong Liu, Shufei Wang and Shengnan Shi
Future Internet 2024, 16(2), 38; https://doi.org/10.3390/fi16020038 - 23 Jan 2024
Viewed by 2111
Abstract
Automatic modulation classification (AMC) plays a crucial role in wireless communication by identifying the modulation scheme of received signals, bridging signal reception and demodulation. Its main challenge lies in performing accurate signal processing without prior information. While deep learning has been applied to [...] Read more.
Automatic modulation classification (AMC) plays a crucial role in wireless communication by identifying the modulation scheme of received signals, bridging signal reception and demodulation. Its main challenge lies in performing accurate signal processing without prior information. While deep learning has been applied to AMC, its effectiveness largely depends on the availability of labeled samples. To address the scarcity of labeled data, we introduce a novel semi-supervised AMC approach combining consistency regularization and pseudo-labeling. This method capitalizes on the inherent data distribution of unlabeled data to supplement the limited labeled data. Our approach involves a dual-component objective function for model training: one part focuses on the loss from labeled data, while the other addresses the regularized loss for unlabeled data, enhanced through two distinct levels of data augmentation. These combined losses concurrently refine the model parameters. Our method demonstrates superior performance over established benchmark algorithms, such as decision trees (DTs), support vector machines (SVMs), pi-models, and virtual adversarial training (VAT). It exhibits a marked improvement in the recognition accuracy, particularly when the proportion of labeled samples is as low as 1–4%. Full article
(This article belongs to the Special Issue Key Intelligent Technologies for Wireless Communications and Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>A semi-supervised learning framework for signal recognition.</p> Full article ">Figure 2
<p>Semi-supervised-learning-based AMC system model.</p> Full article ">Figure 3
<p>Data preprocessing.</p> Full article ">Figure 4
<p>The structure of the proposed semi-supervised AMC method.</p> Full article ">Figure 5
<p>Accuracy comparison of different methods.</p> Full article ">
15 pages, 481 KiB  
Article
DDPG-MPCC: An Experience Driven Multipath Performance Oriented Congestion Control
by Shiva Raj Pokhrel, Jonathan Kua, Deol Satish, Sebnem Ozer, Jeff Howe and Anwar Walid
Future Internet 2024, 16(2), 37; https://doi.org/10.3390/fi16020037 - 23 Jan 2024
Cited by 3 | Viewed by 2196
Abstract
We introduce a novel multipath data transport approach at the transport layer referred to as ‘Deep Deterministic Policy Gradient for Multipath Performance-oriented Congestion Control’ (DDPG-MPCC), which leverages deep reinforcement learning to enhance congestion management in multipath networks. Our method combines DDPG [...] Read more.
We introduce a novel multipath data transport approach at the transport layer referred to as ‘Deep Deterministic Policy Gradient for Multipath Performance-oriented Congestion Control’ (DDPG-MPCC), which leverages deep reinforcement learning to enhance congestion management in multipath networks. Our method combines DDPG with online convex optimization to optimize fairness and performance in simultaneously challenging multipath internet congestion control scenarios. Through experiments by developing kernel implementation, we show how DDPG-MPCC performs compared to the state-of-the-art solutions. Full article
(This article belongs to the Section Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>Abstract view of the interrelationship between DDPG-MPCC Modules.</p> Full article ">Figure 2
<p>Three scenarios to quantify the performance of MPCC across varying bandwidth: (<b>a</b>) MPTCP connection vs. singlepath connection; (<b>b</b>) MPTCP connection vs. two singlepath connections; (<b>c</b>) Two MPTCP connections contending with each other in both paths.</p> Full article ">Figure 3
<p>Cloulab Testbed Setup with Two paths.</p> Full article ">Figure 4
<p>Throughput comparison of one of the shared paths of MPCC with the state-of-the-art MPTCP and TCP protocols over the variation in the size of bottleneck buffer in the path (network scenario a).</p> Full article ">Figure 5
<p>Throughput comparison of MPCC with the state-of-the-art MPTCP and TCP protocols over the variation in the percentage of path loss (network scenario b).</p> Full article ">Figure 6
<p>Comparison of the evolution of the average sending rates of MPCC [<a href="#B5-futureinternet-16-00037" class="html-bibr">5</a>] and DDPG-MPCC on the number of MI (network scenario c, <a href="#futureinternet-16-00037-f002" class="html-fig">Figure 2</a>).</p> Full article ">Figure 7
<p>Evolution of the average utilities of MPCC [<a href="#B5-futureinternet-16-00037" class="html-bibr">5</a>] and DDPG-MPCC over the number of MI (network scenario c, compared to <a href="#futureinternet-16-00037-f006" class="html-fig">Figure 6</a>).</p> Full article ">Figure 8
<p>Evolution of the average reward (<math display="inline"><semantics> <msub> <mi>R</mi> <mi>t</mi> </msub> </semantics></math>) of the trained DDPG over the number of MIs, while DDPG-MPCC and MPCC for utility maximization as demonstrated in <a href="#futureinternet-16-00037-f006" class="html-fig">Figure 6</a>.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-31
Previous Issue
Next Issue
Back to TopTop