Internet of Things最新文献

In contrast to prior research that lacked a comprehensive analysis of a fundamental parameter, namely the average message delivery time (AMDT), this study significantly contributes to the academic community by offering cutting-edge insights into AMDT analysis for wireless healthcare monitoring in smart cities. Employing the innovative Narrowband Internet of Things (NB-IoT) technology, this article thoroughly investigates AMDT within wireless healthcare monitoring. The study focuses on analyzing AMDT during the transmission and reception of messages from base stations (BS) to remote healthcare data aggregators (HDAs) serving remote patients (RPs). Various vital parameters are considered, including modulation rate, average message length (AML), the average number of hops per path (H) required to reach remote patients, patient count, sub-carrier spacing (SS), and the modulation type within the communication range of the BS. This proposed assessment of health network performance offers a brief overview based on essential criteria, facilitating swift and precise evaluations of the efficacy of mobile health applications. The results obtained and the in-depth discussion of the outcomes of AMDT analysis provide valuable insights that can be harnessed to enhance the design and operation of wireless healthcare monitoring systems.

To minimize data traffic in industrial IoT applications, vibration-based condition monitoring should be conducted on sensors and without requiring machine-specific information. The proposed method enables blind estimation of vibration sources, eliminating the need for information about the monitored equipment or external measurements. Vibrations in rotating machinery primarily originate from two sources: dominant gear-related vibrations and low-energy signals associated with bearing faults. Both sources are distorted by the machine's transfer function before reaching the sensor. This method estimates both sources in two stages: first, the gear signal is isolated using a dilated CNN; second, the bearing fault signal is estimated using the squared log envelope of the residual. The effect of the transfer function is removed from both sources using a novel whitening-based deconvolution method (WBD). Both simulation and experimental results demonstrate the method's ability to detect bearing failures early without additional information. This study considers both local and distributed bearing faults, assuming the vibrations are recorded under stable operating conditions.

This paper presents our innovative approach in the field of Internet of Things (IoT). Our focus was on enhancing the research environment at Cheng Shiu University's laboratory through the creation of a dynamic network monitoring setup and lighting control system. Our aim was to simplify equipment management and promote energy efficiency. The core of our development is the Raspberry Pi 4B embedded Linux platform, which serves as a host for both a Message Queuing Telemetry Transport (MQTT) broker and a Node-RED server. Python programs, developed in PyCharm IDE, utilize multiple threads for efficient data exchange via Modbus TCP and MQTT protocols, facilitating seamless interaction between PLCs and MQTT broker servers. The Node-RED IoT development tool was instrumental in crafting a cross-platform Human-Machine Interface (HMI) web server with MQTT Publish/Subscribe capabilities, ensuring smooth communication with the MQTT broker server. By configuring the Apache HTTP server to direct to the Node-RED web directory, we can effectively monitor laboratory conditions and manage the lighting system. The experimental results validate the cost-effectiveness of our approach in transforming traditional control systems into a web-based supervisory control system, thereby enhancing the overall functionality of existing equipment. This project underscores our commitment to advancing IoT solutions in practical settings.

This paper presents a three-stage optimization mechanism designed to enhance Federated Learning (FL) in Industrial Internet of Things (IIoT) networks. Traditional FL optimizations, which typically focus on a single aspect, fall short in IIoT environments. Our approach integrates a multi-criteria enhancement: first, an Ensembled Client Selection Mechanism (ECSM) selects participants based on accuracy, reputation, and randomness. Second, Adaptive Distributed Client Training (ADCT) dynamically adjusts based on participant performance. Lastly, a Secure and Efficient Communication Channel (SECC), backed by blockchain, meets IIoT’s stringent security demands. The evaluation shows TSFed outperforms baseline methods, enhancing FL performance by increasing accuracy and F1-score. Importantly, TSFed improves the efficiency of achieving 80% accuracy on the MNIST dataset by 29.09% over baseline methods, showcasing significant gains in both security and efficiency. This mechanism also exhibits robustness against malicious attacks, setting a new benchmark for FL in IIoT environments.

As industries rapidly advance towards automation and intelligence, the Industrial Internet of Things (IIoT) plays a pivotal role in enhancing factory efficiency and production management. However, the advent of quantum computing introduces significant security challenges, particularly for data transmission and privacy protection, as its immense computational power can compromise the current cryptographic theories' safeguards. We introduce a novel lattice-based cryptographic authentication and key management scheme specifically designed for the IIoT within Wireless Sensor Networks (WSNs), focusing on establishing a secure industrial communication system capable of withstanding quantum threats. The proposed scheme utilizes Crystals-Kyber technology and is built around the Learning With Errors (LWE) problem, providing a robust foundation for security against both traditional and quantum threats. It features innovative methods to safeguard against password guessing, stolen verifier, and replay attacks, enhancing data confidentiality and system integrity. By integrating advanced cryptographic techniques resistant to quantum decryption, this research ensures the secure management and utilization of critical production data, supporting the ongoing transformation towards more intelligent manufacturing processes and sustaining global competitiveness for industries.

In February 2024, the Council and the European Parliament (EP) agreed on the Artificial Intelligence Regulation (usually known as AI Act, AIA) .² This regulation evaluates AI applications to ensure they are used ethically and responsibly, promoting the development of safe and lawful AI across the EU's single market. It establishes a comprehensive legal framework with a risk-based approach, aiming to achieve a balance between protecting the health, safety, and fundamental rights of European citizens and ensuring that the growing AI industry in Europe remains competitive and continues to innovate. The AIA also includes governance mechanisms oriented towards achieving effective implementation throughout the EU. For this purpose, a European Artificial Intelligence Office has already been established. In accordance with the provisions of the forthcoming AIA, it will establish a European Artificial Intelligence Board, an advisory forum, and a scientific panel. Furthermore, it will be set up at the national level the so-called national competent authorities. In this way, a single European governance system for AI is emerging, inspired by collaborative governance, which is essential for achieving fair and effective implementation of AI regulations across the EU. The main objective of this text is to critically examine the governance system established by the AIA. Using the contents of the current version of the AIA (April 2024), this analysis delves into the mechanisms and structures designed to implement AI across the EU. As a conclusion, it offers a critical perspective on the collaborative governance, highlighting its strengths and potential areas for improvement.

The availability of large amounts of data generated by a growing number of Internet of Things devices disseminated in everyday life objects joined with the ability to use data from various sources, to gain insights into human behaviors and subsequently influence them (Internet of Behaviors), is opening the way to novel applications and paradigms for their development. Designing and engineering these applications is still a challenge due to the lack of systematic approaches and reference architectures. In this paper, we introduce the Behavior-enabled IoT paradigm to help architects and engineers design complex IoT applications based on the integration of cyber–physical systems with behavior-enabled models. We instantiate the approach by designing the architecture of a platform for governing a renewable energy community to provide an optimal trade-off between quality of service (i.e., maximizing shared load) and quality of experience (i.e., ensuring a satisfying comfort to community people) through the production of day-ahead scheduling based on expressed people preferences. The architecture is designed following a specific methodology for developing BeT systems and the obtained platform has been tested in a simulated environment. The results show the potential benefits for the community in terms of revenues, quality of experience, and quality of services. BeT-driven strategies, such as day-ahead scheduling for user device activation, provide a cost-effective way to match consumption to renewable energy production and prove to be an alternative to battery storage support systems or even an improvement.

This paper presents the design of a robust IoT (Internet of Things) system developed for Smart Beaches Applications that allows the planning and smart management of these areas. The contributions of this development were to implement low-cost and maintained-free hardware devices with wireless data transmission based on Sigfox LPWAN (Low Power Wide Area Network) technology, evaluation of the application, and overcome the aggressive environment where nodes were placed. Nodes have a low power consumption, which allows to be supplied by batteries, and these batteries are recharged by an energy harvesting module with solar panels. These nodes include different sensors for monitoring physical parameters such as: temperature, relative humidity, ultraviolet index, etc. Sensors’ data are transmitted via the network base stations to an IoT platform and from there to an application server. A database has been designed to save the data and make it available to tourists via a mobile application. In this way, the proposed IoT system achieves real-time monitoring of the existing environmental conditions at beaches and provides tourists with information on the environmental situation of the beach. The node developed presents an average current and power consumption of 1.88 mA and 6 mW, respectively. This consumption allows an autonomous operation without solar panels for around 10 days with a full charged 500 mA battery. Conducted experiments with real deployments at beaches in the Mediterranean coast of Spain have proved that, in case of operating the nodes with solar panels and protecting them adequately, they can run for several years without maintenance thanks to a ruggedized electronic.

In this study, we address the challenge of online malware detection for IoT devices. We propose a method that monitors malware behavior, extracts dynamic features, and converts them into sparse binary images for analysis. The primary problem is to identify the most effective approach among clustering, probabilistic, and deep learning methods for analyzing this unique image dataset. We extract dynamic features from the monitored malware behavior, transforming them into binary images, which are then subjected to three different analysis methods. The clustering, probabilistic, and deep learning approaches are compared and evaluated in terms of various metrics. Our study contributes insights into the performance of various online malware detection approaches for IoT devices. We demonstrate that deep learning outperforms other methods, achieving the best results in seven out of eight metrics. The results of our analysis reveal that the deep learning approach exhibits the highest accuracy in seven of the eight evaluated metrics. We found that the lattice-based approach consistently returns the maximum maliciousness level, which can be instrumental in label flipping scenarios.

Ultra-wideband technology has become increasingly prevalent in localization systems, particularly with the emergence of multi-antenna systems capable of estimating both distance and angle of arrival (AoA) for incoming signals. However, most scientific research analyzes the accuracy of AoA in one specific environment for which the estimator is trained. In this paper, we analyze the performance of various AoA estimation algorithms, such as deep convolutional neural networks (DCNN), multiple signal classification (MUSIC), and phase difference of arrival (PDoA), in unseen environments. Prior work already demonstrated the superior performance of ML for AoA estimation compared to PDoA or MUSIC. We show that MUSIC, PDoA and ML solutions suffer from degradation in unseen environments at the 90th percentile of error, with ML-based AoA estimation degrading by about 14 degrees in unseen environments compared to 4 degrees for PDoA. We demonstrate that while PDoA more effectively corrects AoA at the median level in unseen environments, ML-based methods excel at correcting higher-percentile AoA errors, including outliers. Finally, we propose a novel framework to further improve the correction of AoA outliers for DCNN-based AoA estimators using data augmentation and transfer learning, resulting in a median angular error of only 5 degrees in unseen environments, even considering a field of view up to 90 degrees.