Our paper " **MTL-Split: Multi-Task Learning for Edge Devices using Split Computing** " has been accepted at the **61st Design Automation Conference (DAC) 2024**. Split Computing (SC), where a Deep Neural Network (DNN) is intelligently split with a part of it deployed on an edge device and the rest on a remote server is emerging as a promising approach. It allows the power of DNNs to be leveraged for latency-sensitive applications that do not allow the entire DNN to be deployed remotely, while not having sufficient computation bandwidth available locally. In many such embedded scenarios, such as those in the automotive domain, computational resource constraints also necessitate Multi- Task Learning (MTL), where the same DNN is used for multiple inference tasks instead of having dedicated DNNs for each task, which would need more computing bandwidth. However, how to partition such a multi-tasking DNN to be deployed within a SC framework has not been sufficiently studied. This paper studies this problem, and MTL-Split, our novel proposed architecture, shows encouraging results on both synthetic and real-world data. The source code is available at github.com/intelligolabs/MTL-Split.

Our paper " **Neuro-Symbolic Empowered Denoising Diffusion Probabilistic Models for Real-Time Anomaly Detection in Industry 4.0: Wild-and-Crazy-Idea Paper**" has been accepted at the **Forum on Specification & Design Languages (FDL) 2023**. Industry 4.0 involves the integration of digital technologies, such as IoT, Big Data, and AI, into manufacturing and industrial processes to increase efficiency and productivity. As these technologies become more interconnected and interdependent, Industry 4.0 systems become more complex, which brings the difficulty of identifying and stopping anomalies that may cause disturbances in the manufacturing process. This paper aims to propose a diffusion-based model for real-time anomaly prediction in Industry 4.0 processes. Using a neuro-symbolic approach, we integrate industrial ontologies in the model, thereby adding formal knowledge on smart manufacturing. Finally, we propose a simple yet effective way of distilling diffusion models through Random Fourier Features for deployment on an embedded system for direct integration into the manufacturing process. To the best of our knowledge, this approach has never been explored before.

Our paper " **Verilog-A Implementation of Generic Defect Templates for Analog Fault Injection** " has been accepted at the **Great Lakes Symposium on VLSI (GLSVLSI) 2023**. With functional safety being increasingly important in the development of mixed-signal products for automotive applications, EDA solutions have appeared striving to help designers in the setup and execution of fault injection campaigns. Despite the ongoing work to standardize the definition of defect models and coverage calculation methods in the IEEE P2427 draft standard, there is a lack of a unified and portable method to define defect templates that can be used to inject in a systematic way defects in an analog circuit. Each of the existing EDA tool sets for fault injection proposes its own proprietary method to specify how defects should be defined and injected. The proposed paper describes a Verilog-A-based approach to coding defect templates, which through compliance with the Verilog-A standard, warrants portability across compatible simulators. The approach has been validated on the circuits from the Analogue Benchmark Circuits made available by the IEEE P2427 working group.

Our paper " **Thermal Digital Twin of a Multi-Domain System for Discovering Mechanical Faulty Behaviors**" has been accepted at the **IEEE 21st International Conference on Industrial Informatics (INDIN) 2023**. Constructing a holistic digital twin of a system composed of multiple physical domains is crucial for various tasks. In particular, when the simulation is extended with faults, it becomes a very important resource to achieve robust functional safety analysis. This article proposes a new methodology to build non-electrical fault models for the thermal domain. Such thermal faults are defined through an electrical circuit representing the thermal behavior of the system, known as the Cauer network, based on the physical analogies between the two domains. Including this thermal representation in a multi-domain system allows to simulate the interconnections between different physical domains, thus achieving a more realistic system behavior and evaluating the mutual impact of different domains (e.g., mechanical, electrical and thermal). The entire methodology is applied to a complex case of study implemented by using Verilog-AMS as a proof of concept.

Our paper " **Analog Defect Injection and Fault Simulation Techniques: A Systematic Literature Review**" has been accepted at the **IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems**. Since the last century, the exponential growth of the semiconductor industry has led to the creation of tiny and complex integrated circuits, e.g., sensors, actuators, and smart power. Innovative techniques are needed to ensure the correct functionality of analog devices that are ubiquitous in every smart system. The ISO 26262 standard for functional safety in the automotive context specifies that fault injection is necessary to validate all electronic devices. For decades, standardization of defect modeling and injection mainly focused on digital circuits and, in a minor part, on analog ones. An initial attempt is being made with the IEEE P2427 draft standard that started to give a structured and formal organization to the analog testing field. Various methods have been proposed in the literature to speed up the fault simulation of the defect universe for an analog circuit. A more limited number of papers seek to reduce the overall simulation time by reducing the number of defects to be simulated. This literature survey describes the state- of-the-art of analog defect injection and the fault simulation methods. The survey is based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) methodological flow, allowing for a systematic and complete literature survey. Each selected paper has been categorized and presented to provide an overview of all the available approaches. In addition, the limitations of the various approaches are discussed by showing possible future directions.

Our paper titled " **Timing predictability for SOME/IP-based service-oriented automotive in-vehicle networks**" has been accepted to IEEE Design, Automation & Test in Europe Conference & Exhibition (DATE) 2023. In-vehicle network architectures are evolving from a typical signal-based client-server paradigm to a service-oriented one, introducing flexibility for software updates and upgrades. While signal-based networks are static by nature, service-oriented ones can more easily evolve during and after the design phase. As a result, service-oriented protocols are becoming more prominent in automotive in-vehicle networks. While applications like infotainment are less sensitive to delays, others like sensing and control have more stringent timing and reliability requirements. Hence, wider adoption of service-oriented protocols requires addressing the timing analysis and predictability of such protocols, which is more challenging than in their signal-oriented counterparts. In service-oriented architectures, the discovery phase defines how clients find their required services. The time required to complete the discovery phase is an important parameter since it determines the readiness of a sub-system or even the vehicle. In this paper, we develop a formal timing analysis of the discovery phase of SOME/IP, which is an emerging service-oriented protocol being considered for adoption by several automotive Original Equipment Manufacturers (OEMs) and suppliers.

Our paper " **Statistical Approach to Efficient and Deterministic Schedule Synthesis for Cyber-Physical Systems**" has been accepted at the **Automated Technology for Verification and Analysis (ATVA) 2023**. Correctness of controller implementations rely on real-time guarantees that all control tasks finish execution by their prescribed deadlines. However, with increased complexity and heterogeneity in hardware, the worst-case execution time estimates are becoming very conservative. Thus, for efficient usage of hardware resources, some control tasks might have to miss their deadlines. Recent work has shown that a system can still abide by its safety requirements even after missing some of its deadlines. This paper investigates an approach to synthesize a scheduler for control tasks that miss some deadlines without compromising its safety requirements. But given that the number of possible schedules increase combinatorially with the number of tasks involved, our scheduler synthesis uses an efficient automata representation to search for the appropriate schedule. We incorporate statistical verification techniques to construct this automaton and accelerate the search process. Statistical verification is advantageous compared to deterministic verification in the synthesis process in two ways: first, it enables us to synthesize schedules that would not be possible otherwise, and second, it drastically reduces the time taken to synthesize such a schedule. We demonstrate both these advantages through a case study with five controllers having different safety specifications, but sharing the same computational resource.

Our paper " **Multi-Domain Fault Models Covering the Analog Side of a Smart or Cyber-Physical System**" has been accepted at the **IEEE Transactions on Computers 2023**. Over the last decade, the industrial world has been involved in a massive revolution guided by the adoption of digital technologies. In this context, complex systems like cyber-physical systems play a fundamental role since they were designed and realized by composing heterogeneous components. The combined simulation of the behavioral models of these components allows to reproduce the nominal behavior of the real system. Similarly, a smart system is a device that integrates heterogeneous components but in a miniaturized form factor. The development of smart or cyber-physical systems, in combination with faulty behaviors modeled for the different physical domains composing the system, enables to support advanced functional safety assessment at the system level. A methodology to create and inject multi-domain fault models in the analog side of these systems has been proposed by exploiting the physical analogy between the electrical and mechanical domains to infer a new mechanical fault taxonomy. Thus, standard electrical fault models are injected into the electrical part, while the derived mechanical fault models are injected directly into the mechanical part. The entire flow has been applied to two case studies: a direct current motor connected with a gear train, and a three- axis accelerometer.

Our newest work, a collaboration between the University of Verona, Italy, and the University of Stavanger, Norway, " _ ** _SHPIA 2.0: An easily scalable, low-cost, multi-purpose Smart Home Platform for Intelligent Applications_**_ ," has been accepted to Springer Nature Computer Science! 🎉 _**More information below**_ 🧵 Sensors, electronic devices, and smart systems have invaded the market and our daily lives. As a result, their utility in smart home contexts to improve the quality of life, especially for the elderly and people with special needs, is getting stronger and stronger. Therefore, many systems based on smart applications and intelligent devices have been developed, for example, to monitor people's environmental contexts, help in daily life activities, and analyze their health status. However, most existing solutions have drawbacks related to accessibility and usability. They tend to be expensive and lack generality and interoperability. These solutions are not easily scalable and are typically designed for specific constrained scenarios. This paper tackles such drawbacks by presenting SHPIA 2.0, an easily scalable, low-cost, multipurpose smart home platform for intelligent applications. It leverages low-cost Bluetooth Low Energy (BLE) devices featuring both BLE- connected and BLE broadcast modes, to transform common objects of daily life into smart objects. Moreover, SHPIA 2.0 allows the collection and automatic labeling of different data types to provide indoor monitoring and assistance. Specifically, SHPIA 2.0 is designed to be adaptable to different home-based application scenarios, including human activity recognition, coaching systems, and occupancy detection and counting. The SHPIA platform is open source and freely available to the scientific community, fostering collaboration and innovation.

Our newest work, a collaboration between the University of Limerick, Ireland, and the University of Stavanger, Norway "CNN-based Human Activity Recognition on Edge Computing Devices" has been accepted at IEEE International Conference on Omni-layer Intelligent Systems #IEEECOINS23! 🎉 A big thank goes to Amandeep Singh, and Prof. Tiziana Margaria for their effort and collaborative spirit. More information below 🧵. Human Activity Recognition (HAR) is a widely known research area that involves wearable devices integrating inertial and/or physiological sensors to classify human actions and status across various application domains, such as healthcare, sports, industry, and entertainment. However, executing HAR algorithms on remote devices or the cloud can lead to several issues, such as latency, bandwidth requirements, energy consumption, privacy risks, and limited offline capabilities, limiting its applicability to various fields that require a quasi-real-time response. On the other hand, performing HAR on edge devices close to the data source can offer several advantages, such as reduced latency, low bandwidth usage, energy efficiency, privacy, and offline capabilities. Therefore, transitioning towards Edge HAR can be a more effective and versatile solution that can overcome the challenges associated with traditional HAR techniques. This paper presents a novel approach to HAR computation on edge devices, utilizing a Convolutional Neural Network (CNN) Deep Learning approach, and compares its performance with cloud-based HAR computation. The paper is accompanied by a self-collected dataset including nine different daily activities from 12 users and new algorithms specifically designed for edge computing. The experiments on the self-collected dataset demonstrate that the proposed edge computing model achieves high accuracy of over 92%, high Precision, Recall, and F1-score. Furthermore, the model exhibits significantly reduced latency, with only 117 ms, and utilizes minimal memory, with a peak of 18.8 Kb RAM and 956 Kb Flash memory. #IEEECOINS23 #edgecomputing #health #collaboration #university #research

Our newest work, a collaboration between the University of Limerick, Ireland, and the University of Turin, Italy " _ ** _Experiences from the first delivery of a new Immersive Software Engineering course: mathematical foundations and data analytics_**_ " has been accepted to IEEE Computer Society Signature Conference on Computers, Software, and Applications #COMPSAC23! 🎉 #ImmersiveSoftwareEngineering #ISE #mathematicalfoundations #dataanalytics

Our newest work, a collaboration between the University of Verona, Italy, and the University of Stavanger, Norway, " _ ** _Noninvasive Monitoring of Alzheimer 's Patients through WiFi Channel State Information_**_," has been accepted to IEEE International Workshop on Advances in Sensors and Interfaces #IWASI23! 🎉 More information below 🧵 The design of noninvasive systems for monitoring people's activities is becoming of central interest in recent years. Such systems are essential for those affected by diseases that modify their cognitive status and are not collaborative in using wearable or interactive systems (e.g., mobile apps to communicate). This is particularly true regarding neurodegenerative diseases that involve memory loss, cognitive decline, communication difficulties, behavioral changes, loss of independence, and physical complications. In response to the need of healthcare structures and caregivers to monitor this category of people during their in-home daily life, this paper proposes a non-intrusive system capable of detecting whether or not a person is in his/her room and if he/she is lying on the bed. Checking these conditions is of utmost importance, in particular, during the night, to support the monitoring activity of caregivers and social-health operators taking care of people with Dementia and Alzheimer's disease. The proposed system exploits WiFi's Channel State Information (CSI) gathered by common access points installed in the room. CSI data are then used to train a Convolutional Neural Network (CNN), and a fine-tuning technique is applied to increase the generalization capabilities of the CNN model on new environments. In our experimental analysis, we trained the CNN model by collecting CSI data in four different rooms from two subjects performing three distinct activities. Promising results (accuracy >97.5 %) in recognizing the target activities have been achieved.

Our newest work, a collaboration between the University of Verona, Italy, TU Chemnitz, and the University of Stavanger, Norway, " _ _ **A low-cost Wireless Body Area Network for Human Activity Recognition in Healthy Life and Medical Applications**__ ," has been accepted to IEEE Transactions on Emerging Topics in Computing **#IEEETETC** !🎉 More information below Moved by the necessity, also related to the ongoing COVID-19 pandemic, of the design of innovative solutions in the context of digital health, and digital medicine, Wireless Body Area Networks (WBANs) are more and more emerging as a central system for the implementation of solutions for well-being and healthcare. In fact, by elaborating the data collected by a WBAN, advanced classification models can accurately extract health-related parameters, thus allowing, as examples, the implementations of applications for fitness tracking, monitoring of vital signs, diagnosis, and analysis of the evolution of diseases, and, in general, monitoring of human activities and behaviors. Unfortunately, commercially available WBANs present some technological and economic drawbacks from the point of view, respectively, of data fusion and labeling and the cost of the adopted devices. To overcome existing issues, in this paper, we present the architecture of a low-cost WBAN, which is built upon accessible off-the- shelf wearable devices and an Android application. Then, we report its technical evaluation concerning resource consumption. Finally, we demonstrate its versatility and accuracy in medical and well-being applications. The designed WBAN will be made freely available to the community.

I'm glad to announce that the European Commission has awarded **Enrico Fraccaroli** a three-year Marie Skłodowska-Curie Postdoctoral Fellowship (Global Fellowship). Marie Skłodowska-Curie Actions (MSCA) are a set of major research fellowships created by the European Union/European Commission to support research in the European Research Area (ERA). MSCA fellowships are among Europe's most competitive and prestigious awards, aimed at supporting the best and most promising scientists. The fellowship will allow us to develop the " _STRATEgic GUide to Smart manufacturing_ " (STRATEGUS) project. During the three-year project, Enrico will spend 24 months at the **University of North Carolina at Chapel Hill** , and 12 months at the **University of Verona**. The STRATEGUS project aims at providing understandable techniques and software architectures for guiding industrials towards the proper integration of new technologies in their existing production plants. Website: https://strategus-he2022.github.io/

Our paper " **Modeling Cyber-Physical Production Systems with SystemC-AMS**" has been accepted to **IEEE Transactions on Computers 2022**. The heterogeneous nature of SystemC-AMS makes it a perfect candidate solution to support Cyber-Physical Production Systems (CPPSs), i.e., systems that are characterized by a tight interaction of the cyber part with the surrounding physical world and with manufacturing production processes. Nonetheless, the support for the modeling of physical and mechanical dynamics typical of production machinery goes far beyond the initial application scenario of SystemC-AMS, thus limiting its effectiveness and adoption in the production and manufacturing context. This paper starts with an analysis of the current adoption of SystemC-AMS to highlight the open points that still limit its effectiveness, with the goal of pinpointing current issues and to propose solutions that could improve its effectiveness, and make SystemC-AMS an essential resource also in the new Industry 4.0 scenario.