Article
Computer Science, Information Systems
Ali Alzahrani, Theyazn H. H. Aldhyani
Summary: The Internet of Things (IoT) has become increasingly popular, promoting communication between smart devices to enhance process efficiency. The study proposes an intrusion detection system based on artificial intelligence algorithms, utilizing K-nearest neighbors (KNN), linear discriminant analysis (LDA), convolutional neural network (CNN), and convolutional long short-term memory neural network (CNN-LSTM) to identify MQTT protocol IoT intrusions. The deep learning algorithm achieved high precision in detecting intrusions, outperforming traditional machine learning methods like KNN and LDA.
Article
Computer Science, Information Systems
Axelle Hue, Gaurav Sharma, Jean-Michel Dricot
Summary: The integration of embedded sensors, actuators, and RFIDs in our surroundings, combined with rapid developments in high-speed wireless networks, is paving the road for the Internet-of-Things paradigm. However, privacy remains a major challenge for IoT, and this paper studies the design of a privacy-enhanced communication protocol for lightweight IoT devices.
Review
Chemistry, Analytical
Muhammad Almas Khan, Muazzam A. Khan, Sana Ullah Jan, Jawad Ahmad, Sajjad Shaukat Jamal, Awais Aziz Shah, Nikolaos Pitropakis, William J. Buchanan
Summary: This paper proposed a Deep Neural Network (DNN) for intrusion detection in MQTT-based protocol and compared its performance with traditional machine learning algorithms. Results showed that the DNN model achieved high accuracy in both binary and multi-label classification on different datasets.
Article
Computer Science, Artificial Intelligence
Angel Luis Munoz Castaneda, Jose Antonio Aveleira Mata, Hector Alaiz-Moreton
Summary: Nowadays, the cybersecurity of IoT environments is a major challenge. This paper aims to obtain a consistent dataset of network traffic in an IoT system based on the MQTT protocol and characterize different attacks using a hybrid feature selection algorithm. The dataset and algorithm provide an optimal characterization of the attacks in terms of accuracy and explaining their nature, confirming the dataset's consistency.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Chemistry, Analytical
Carlos D. Morales-Molina, Aldo Hernandez-Suarez, Gabriel Sanchez-Perez, Linda K. Toscano-Medina, Hector Perez-Meana, Jesus Olivares-Mercado, Jose Portillo-Portillo, Victor Sanchez, Luis Javier Garcia-Villalba
Summary: With the wide adoption of new data sharing technologies like IoT, intelligent security controls have become crucial to protect information assets. Traditional applications such as IDS, IPS, and SIEM are inadequate in handling novel security incidents like identity impersonation attacks. Therefore, a robust AI-based protection framework has been proposed to tackle these challenges by utilizing unsupervised pre-training techniques and deep feature engineering with DNN models.
Article
Computer Science, Theory & Methods
Wei Liu, Ruiming Wang, Xuyan Qi, Liehui Jiang, Jing Jing
Summary: Physical unclonable functions (PUFs) are promising solutions for low-cost device authentication. This study introduces a method that combines challenge-response pairs with side-channel information to model strong PUFs with complex structures, and experimental results show the effectiveness of this method.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2022)
Article
Computer Science, Information Systems
Hayette Zeghida, Mehdi Boulaiche, Ramdane Chikh
Summary: Nowadays, the IoT has enabled machines to communicate, collect data, and make decisions to improve daily human life. However, the vulnerability of IoT devices makes them susceptible to cyber-attacks. Traditional detection methods may not be efficient due to the complexity and variety of attacks. This paper proposes an ensemble learning-based intrusion detection model that utilizes a public dataset of MQTT attacks. The experiment results show that the proposed model increases detection accuracy and F1-score up to 95%, with MCC exceeding 90%.
INTERNATIONAL JOURNAL OF INFORMATION SECURITY
(2023)
Article
Chemistry, Analytical
Shaza Dawood Ahmed Rihan, Mohammed Anbar, Basim Ahmad Alabsi
Summary: This paper proposes an approach for detecting attacks on IoT networks using ensemble feature selection and deep learning models. The impact of the selected feature set on the performance of Deep Learning (DL) models is evaluated. The DL models achieved high detection accuracy, precision, recall, and F1 measure values.
Article
Engineering, Mechanical
Astrid Maritza Gonzalez-Zapata, Esteban Tlelo-Cuautle, Israel Cruz-Vega, Walter Daniel Leon-Salas
Summary: This paper demonstrates the use of artificial neurons to generate chaotic binary sequences on embedded systems, leveraging WiFi network connectivity to develop a lightweight cryptographic application. The randomness of the sequences is enhanced and evaluated using statistical tests, while security measures are implemented through different initialization conditions and testing methods for image encryption.
NONLINEAR DYNAMICS
(2021)
Article
Chemistry, Multidisciplinary
Sunoh Choi, Jaehyuk Cho
Summary: This study analyzed malicious attacks on Internet of Things systems, proposed a new feature extraction method, and classified the attacks using a Seq2Seq model. The experimental results showed an accuracy of 99.97%, which is 7.33% higher than previously reported methods.
APPLIED SCIENCES-BASEL
(2022)
Article
Chemistry, Analytical
Kevin Carvalho, Jorge Granjal
Summary: IoT applications are increasingly integrated into society, but privacy remains a crucial concern. Privacy by design approaches are necessary to ensure data security in IoT applications. Proposals to enhance user privacy using blockchain and other technologies have been proven effective in addressing security and privacy concerns.
Article
Computer Science, Artificial Intelligence
R. Mohan Das, U. Arun Kumar, S. Gopinath, V. Gomathy, N. A. Natraj, N. K. Anushkannan, Adhavan Balashanmugham
Summary: In the innovative concept of the Social Internet of Things (IoT), the combination of IoT and social platforms allows devices to interact with each other. However, customers have concerns about privacy and information security. To improve trust and privacy in IoT, categorization of safe nodes and identification of fraudulent nodes need to be addressed. This research proposes a novel Multi-hop Convolutional Neural Network with an attention mechanism (MH-CNN-AM) to identify and separate hostile nodes from the network, aiming to enhance the performance and security of IoT.
Article
Chemistry, Analytical
Oezlem Seker, Goekhan Dalkilic, Umut Can cabuk
Summary: This article introduces a mutual authentication and role-based authorization scheme (MARAS) for lightweight Internet of things applications, which brings mutual authentication and authorization to the network using dynamic access tokens, HMAC-based HOTP, AES, hash chains, and a trusted server running OAuth2.0 with MQTT. The overhead of MARAS for publish messages is 49 bytes and for connect messages is 127 bytes. Overall data traffic with MARAS remains lower than without it, and the round-trip times for connect messages are delayed very minimally. The delay for publish messages depends on the size and frequency of the published information, but is upper bounded by 163% of the network defaults. The comparison with similar works shows that MARAS offers better computational performance by offloading computationally intensive operations to the broker side.
Article
Chemistry, Multidisciplinary
Milutin Radonjic, Sanja Vujnovic, Aleksandra Krstic, Zarko Zecevic
Summary: This featured application proposes a system that uses acoustic signals to estimate the condition of rotating machines for more reliable preventive maintenance. With the advancement of Industrial Internet of Things solutions and artificial intelligence algorithms, intelligent maintenance systems have outperformed traditional approaches. The proposed system, involving a mobile and inexpensive IoT device and an algorithm combining wavelet transform and neural networks, has been tested in a real industrial setting with high accuracy.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Min-Huang Ho, Ming-Yi Lai, Yung-Tien Liu
Summary: In this paper, a cloud platform based on the DDS communication protocol is implemented to acquire data from various sensors and display them in real time on a webpage. The platform also allows for recording the motion displacement of the machine tool for further performance examination. The proposed platform enables legacy machines to have sensing and communication abilities, making the development of smart machines feasible for future Industry 4.0 applications.
Article
Computer Science, Cybernetics
Maria Teresa Garcia-Ordas, Hector Alaiz-Moreton, Jose-Luis Casteleiro-Roca, Esteban Jove, Jose Alberto Benitez-Andrades, Isaias Garcia-Rodriguez, Hector Quintian, Jose Luis Calvo-Rolle
Summary: This study compares the performance of four clustering techniques and aims to achieve strong hybrid models in supervised learning tasks. A real dataset from a bio-climatic house in a wind farm in Galicia, Spain is collected. The authors utilize different clustering methods followed by a regression technique to predict the output temperature of a thermal solar generation system. Two possible solutions are implemented to evaluate the quality of each clustering method, including unsupervised learning metrics and common error measurements for regression algorithms such as Multi Layer Perceptron.
CYBERNETICS AND SYSTEMS
(2023)
Article
Chemistry, Analytical
Manuel Castejon-Limas, Laura Fernandez-Robles, Hector Alaiz-Moreton, Jaime Cifuentes-Rodriguez, Camino Fernandez-Llamas
Summary: This paper introduces the PipeGraph software library, an open-source python toolbox for easing the creation of machine learning models. PipeGraph extends the concept of scikit-learn's Pipeline by using a graph structure, allowing for diverse operations and compatibility with the GridSearchCV function.
Article
Computer Science, Information Systems
Esteban Jove, Jose Aveleira-Mata, Hector Alaiz-Moreton, Jose-Luis Casteleiro-Roca, David Yeregui Marcos del Blanco, Francisco Zayas-Gato, Hector Quintian, Jose Luis Calvo-Rolle
Summary: This article presents the implementation of an Intrusion Detection System (IDS) based on the deployment of different one-class classifiers to prevent attacks over the Internet of Things (IoT) protocol Message Queuing Telemetry Transport (MQTT). The utilization of real data sets has allowed the algorithms to show remarkable performance in detecting attacks.
Article
Environmental Sciences
Leticia Sanchez, Nelida Fernandez, Angela P. Calle, Valentina Ladera, Ines Casado, Enrique Bayon, Isaias Garcia, Ana M. Sahagun
Summary: Breast cancer has significant impact on public health, being the most frequent malignant tumor and leading cause of cancer death in women. This study explores the use of focus groups by nurses to gather information on the emotional state of breast cancer women and develop coping strategies. The findings show that focus groups allow nurses to evaluate the expression of emotions, collect and share information, helping survivors cope with the stress related to their illness more easily.
INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH
(2022)
Article
Computer Science, Artificial Intelligence
David Escudero Garcia, Noemi DeCastro-Garcia, Angel Luis Munoz Castaneda
Summary: In this research, the performance of transfer learning techniques for malware detection is evaluated over different time horizons and learning settings. Experiments are conducted on unbalanced data with different file types to address additional challenges in malware detection. The goal is to determine if transfer learning can help solve the concept drift problem and build models that can detect new malware.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Alvaro Michelena, Jose Aveleira-Mata, Esteban Jove, Martin Bayon-Gutierrez, Paulo Novais, Oscar Fontenla Romero, Jose Luis Calvo-Rolle, Hector Alaiz-Moreton
Summary: This paper presents a novel method to develop an intrusion detection system (IDS) to detect man-in-the-middle attacks. The method uses intelligent models based on feature extraction and supervised classification techniques. Experimental results show excellent performance in detecting MQTT attacks.
Article
Computer Science, Artificial Intelligence
Ivan de-Paz-Centeno, Maria Teresa Garcia-Ordas, Oscar Garcia-Olalla, Hector Alaiz-Moreton
Summary: The accurate measurement of energy production in photovoltaic power stations is a significant issue, especially when similar measurements are lacking or predictive models are difficult to generate. This paper proposes an artificial neural network-based solution that can effectively impute missing data. The model is well-suited for environments with missing values ranging from 30% to 70%, particularly outperforming other methods when the severity of missing values reaches 50%.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Engineering, Electrical & Electronic
Natalia Prieto-Fernandez, Sergio Fernandez-Blanco, Alvaro Fernandez-Blanco, Jose Alberto Benitez-Andrades, Francisco Carro-De-Lorenzo, Carmen Benavides
Summary: This article presents an efficient feature extraction approach, called weighted conformal LiDAR-mapping (WCLM), for mapping structured environments. The proposed methodology uses conformal Möbius transformation to extract polygonal profiles and propagate uncertainties from raw measurement data. Experimental validation using 2-D data obtained from a low-cost LiDAR range finder demonstrates a significant improvement in computational efficiency compared to other SLAM approaches.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Jose Alberto Benitez-Andrades, Alvaro Gonzalez-Jimenez, Alvaro Lopez-Brea, Carmen Benavides, Jose Aveleira-Mata, Jose-Manuel Alija-Perez, Maria Teresa Garcia-Ordas
Summary: This paper presents a solution based on deep learning models and transfer learning to detect racist and xenophobic messages in Spanish. Promising results were achieved using a dataset from Twitter and BERT models.
METADATA AND SEMANTIC RESEARCH, MTSR 2021
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Martin Bayon-Gutierrez, Jose Alberto Benitez-Andrades, Sergio Rubio-Martin, Jose Aveleira-Mata, Hector Alaiz-Moreton, Maria Teresa Garcia-Ordas
Summary: This paper presents an architecture and results of a roadway detection system that utilizes both camera and LiDAR data for segmenting the road surface from a Bird's-eye view. The combined use of camera and LiDAR data is discussed, along with example images and the development of a neural model. The proposed method performs well on the Kitti Road dataset compared to other state-of-the-art methods. The paper also introduces future research possibilities and discusses the benefits of using the full LiDAR FOV for road detection.
HYBRID ARTIFICIAL INTELLIGENT SYSTEMS, HAIS 2022
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Esteban Jove, Jose Aveleira-Mata, Hector Alaiz-Moreton, Jose-Luis Casteleiro-Roca, David Yeregui Marcos del Blanco, Francisco Zayas-Gato, Hector Quintian, Jose Luis Calvo-Rolle
Summary: The significant advance in smart devices connected to the Internet has led to the development of Internet of Things technology. However, this success has also brought the need for implementing Intrusion Detection Systems to address potential attacks. This research focuses on intrusion detection in a network using the Message Queuing Telemetry Transport protocol, and has implemented various one-class classifiers from a real dataset, achieving good performance in detecting intrusion attacks.
SUSTAINABLE SMART CITIES AND TERRITORIES
(2022)
