This fourth issue of the tenth volume of the Journal of Reliable Intelligent Environments collects eleven articles in various technical areas.

Automated phase-type distribution fitting via expectation maximization—by Marco Mialaret, Paulo Pereira, Antônio Sá Barreto, Thiago Pinheiro and Paulo Maciel—focuses on modeling complex data based on time and event duration. In particular, this paper introduces a strategy that leverages user-friendly tools, graphical adjustment features, and integration with existing tools to streamline the process of fitting phase-type (PH) distributions to empirical data.

Trainable Gaussian-based activation functions for sensor-based human activity recognition—by Javier Machacuay and Mario Quinde—proposes two Trainable Gaussian-based Activation Functions (AFs) for Multilayer Perceptron neural networks on sensor-based human activity recognition (HAR), namely, the Four-Parameter Activation Gaussian Radial Basis Function (T4GRBF) and the Weighted Gaussian Radial Basis Function (WGRBF). Such AFs have been validated concerning two distinct datasets.

Dependability analysis and disaster recovery measures in smart hospital systems—by Luiz Nelson Lima, Arthur Sabino, Vandirleya Barbosa, Leonel Feitosa, Carlos Brito, Jean Araujo and Francisco Airton Silva—focuses on availability and reliability measures for smart hospital infrastructures. The paper proposes a method for disaster analysis and recovery measures, which is based on Stochastic Petri Nets (SPN). The proposed method helps to identify the system’s most critical components, develop strategies to mitigate failures, and ensure system resilience.

A multilevel graph approach for IoT-based complex scenario management through situation awareness and semantic approaches—by Mario Casillo, Francesco Colace, Angelo Lorusso, Domenico Santaniello and Carmine Valentino—develops a methodology with predictive capabilities and context adaptability for managing complex scenarios. The methodology adopts an integrated set of graph-based approaches such as Ontologies, Context Dimension Trees, and Bayesian Networks. The methodology has been validated concerning two scenarios. The former is derived from open data related to a large-scale environment, such as the city of London, and the latter from a smaller environment, such as a smart home.

A formal model-based approach to design failure-aware Internet of Things architectures—by Imene Ben Hafaiedh, Amani Elaoud and Asma Maddouri—presents a model-based design approach for IoT architectures that uses formal models to analyze failure-related behaviors. The goal is twofold. On the one hand, it is to make an early reliability analysis in the development cycle before the implementation to reduce the cost associated with discovering and rectifying failures later in IoT architectures. On the other hand, it is to explore different design patterns and strategies without needing costly development, coding, and testing.

A survey on the contribution of ML & DL to the detection & prevention of botnet attacks—by Yassine EL Yamani, Youssef Baddi and Najib EL Kamoun—reports on current research on the use of Deep Learning for the detection and prevention of some kinds of cyber threats.

An efficient approach of epilepsy seizure alert system using IoT and machine learning—by Jagadeesh Basavaiah, Audre Arlene Anthony, S Mahadevaswamy and H. N Naveen Kumar—presents a system that exploits several types of sensors (eg, GPS, GSM modules, ECG sensors) and three distinct algorithms to detect and predict epilepsy.

Secure electronic monitoring of sex offenders—by Francesco Buccafurri, Vincenzo De Angelis, Maria Francesca Idone and Cecilia Labrini—focuses on controlling sex offenders via electronic monitoring to make the environment safer for potential victims. The authors analyze the existing approaches-based on RFID and GPS to detect the distance between offenders and victims—from the security perspective. Then, they overcome some drawbacks of current solutions by proposing a new GPS-based one.

Performance evaluation of a video surveillance system using stochastic petri nets for license plate detection on highways—by Carlos Brito, Vandirleya Barbosa, Luiz Nelson Lima, José Wanderlei Rocha, José Miquéias Araújo, Lucas Lopes, Paulo A. L. Rego, Michel Sales, Gustavo Callou, Iure Fé and Francisco Airton Silva—proposes a performance model based on stochastic Petri networks for the evaluation of video surveillance systems dedicated to the detection of license plates on highways.

ECC based certificateless aggregate signature scheme for healthcare wireless sensor networks—by Lalit Negi and Devender Kumar—defines a certificateless aggregate signature scheme (called E-CLAS) based on ECC for remote healthcare monitoring applications. The scheme generates an aggregate signature on healthcare information produced by a set of patients, which is verified by a designated doctor. The proposed schema has been formally analyzed using a random oracle model.

Enhancing trustworthiness in ML-based network intrusion detection with uncertainty quantification—by Jacopo Talpini, Fabio Sartori and Marco Savi—focuses on Machine Learning based Intrusion Detection Systems. The authors initially compared various ML-based methods for uncertainty quantification and open-set classification. Next, they developed a custom model based on Bayesian Neural Networks that showed a lower variance in the results over different scenarios, compared to the state-of-the-art.

We hope these articles stimulate the community to produce further improvements in these areas.