Special issue on embedded real-time applications

As computing platforms are becoming more powerful and energy efficient, embedded real-time systems are spreading in emerging application domains, from small autonomous robots to unmanned aerial vehicles, medical wearable devices, and intelligent sensors and actuators for the Internet of Things. Many of such systems are required to interact with the surrounding environment, reacting to events within stringent deadlines, also guaranteeing security and safety features. In addition, the heavy use of machine learning algorithms for perception and control tasks is increasing the complexity of the software architecture, which is often organized in modular components with different levels of criticality. One common way to manage such a software complexity is to partition the computational resources available on the platform into a set of execution domains coordinated by a hypervisor, which encapsulates each software component into a virtual machine, while guaranteeing safety, security, and predictability properties. The four papers collected in this special issue address some of the crucial topics highlighted above, presenting them in different application contexts. The articles are extended versions of papers carefully selected over twenty-four works presented at the 7th Italian Workshop on Embedded Systems (IWES 2022), held at Politecnico di Bari, Bari, Italy, on September 22–23 of 2022, and went through a rigorous and independent review process, being evaluated by at least three expert reviewers in the subject area of each article. The article titled “Evaluating Virtualization for Fog Monitoring of Real-time Applications in Mixed-Criticality Systems” by Marcello Cinque, Luigi De Simone, Nicola Mazzocca, Daniele Ottaviano, and Francesco Vitale presents a model-based system development process to design, deploy, and evaluate mixed criticality systems on multiprocessors system-on-chips. The authors consider the use of an embedded hypervisor to virtualize resources and meet real-time requirements. The article shows the applicability of the proposed approach using industry-relevant case studies in the domain of nuclear fusion applications. The article titled “Time-Sensitive Autonomous Architectures” by Donato Ferraro, Luca Palazzi, Federico Gavioli, Michele Guzzinati, Andrea Bernardi, Benjamin Rouxel, Paolo Burgio, and Marco Solieri presents a software architecture to enable time-sensitive networks in autonomous and software-defined vehicles. The architecture is based on a hypervisor providing strong isolation and virtual access to time-sensitive networks for virtual machines in the context of an autonomous car controlled by two Xilinx accelerators and a multiport TSN switch. The paper discusses the engineering challenges and the performance evaluation of the project demonstrator. The article titled “Supporting AI-Powered Real-Time Cyber-Physical Systems on Heterogeneous Platforms via Hypervisor Technology”, by Edoardo Cittadini, Mauro Marinoni, Alessandro Biondi, Giorgiomaria Cicero, and Giorgio Buttazzo presents a multi-domain hypervisor-based architecture that leverages heterogeneous platforms and virtualization technologies to support AI-powered applications consisting of modules with mixed criticalities and safety requirements. In particular, deep learning algorithms run in a high-performance domain under the Linux operating system, whereas control and monitoring functions run in a safety-critical domain under a real-time operating system. The proposed approach is validated on a drone capable of tracking moving targets using a deep neural network accelerated on the FGPA available on the platform. The article titled “A Real-Time Vital Control Module to Increase Capabilities of Railway Control Systems in Highly Automated Train Operations”, by Arturo Amendola, Mario Barbareschi, Salvatore De Simone, Giovanni Mezzina, Alberto Moriconi, Cataldo Luciano Saragaglia, Diana Serra, and Daniela De Venuto, proposes a hardware/software vital control module architecture able to expand the control capabilities of existing train control systems. The module includes a board managed by a reliable and safe hard real-time operating system developed to be compliant with related safety standards. The application logic has been developed with a model-based approach and implemented on a Xilinx Ultrascale+ platform. We would like to thank all the authors for submitting their excellent work to this special issue and the reviewers for providing their valuable feedback. We believe this special issue provides a snapshot of up-to-date applications of real-time systems and we hope it fosters further research along this line.