Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Luca Davoli has a total of 186 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Performance Assessment of DL for Network Intrusion Detection on a Constrained IoT Device

    by Armin Mazinani, Daniele Antonucci, Luca Davoli, Gianluigi Ferrari

    This work investigates the deployment of Deep Learning (DL) models for network intrusion detection on resource-constrained IoT devices, using the public CICIoT2023 dataset. In particular, we consider the following DL models: Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), Temporal Convolutional Network (TCN), Multi-Layer Perceptron (MLP). Bayesian optimization is employed to fine-tune the models’ hyperparameters and ensure reliable performance evaluation across both binary (2-class) and multi-class (8-class, 34-class) intrusion detection. Then, the computational complexity of each DL model is analyzed—in terms of the number of Multiply–ACCumulate operations (MACCs), RAM usage, and inference time—through the STMicroelectronics Cube.AI Analyzer tool, with models being deployed on an STM32H7S78-DK board. To assess the practical deployability of the considered DL models, a trade-off score (balancing classification accuracy and computational efficiency) is introduced: according to this score, our experimental results indicate that MLP and TCN outperform the other models. Furthermore, Post-Training Quantization (PTQ) to 8-bit integer precision is applied, allowing the model size to be reduced by more than 90% with negligible performance degradation. This demonstrates the effectiveness of quantization in optimizing DL models for real-world deployment on resource-constrained IoT devices.


  • Open Access

    Sustainable Water Management in an Evolving Tomato Cultivation Testbed

    by Giulia Oddi, Laura Belli, Luca Davoli, Luca Preite, Martina Galaverni, Margherita Rodolfi, Giuseppe Vignali, Tommaso Ganino, Gianluigi Ferrari

    Nowadays traditional agricultural irrigation practices present significant challenges for both sustainability and productivity, resulting in excessive water consumption—70% of the world’s water consumption is dedicated to agricultural irrigation—and inefficient resource management. The adoption of innovative solutions and technologies to support smart agriculture applications allows to overcome these problems and improve the productivity in terms of crop yield, fruits quality, efficiency, resource management, and waste reduction. Guided by these goals, this work examines the activities carried out within the NextGenerationEU-funded AGRITECH project in an evolving tomato cultivation testbed. The evaluation has been conducted for two consecutive years, namely, 2023 and 2024, with the deployment of Internet of Things devices and the support of the Agriware cloud platform. In particular, the automated irrigation system deployed during the second year of experimentation allowed for more precise and efficient irrigation scheduling with respect to that of the first year (manually managed). In fact, $\mathbf{0.13}$ m$^{\mathbf{3}}$ of water per square mater have been saved, while the water use efficiency improved by 22.2%, considering the global yield, thanks to the revised automated system. The experimental results demonstrate the feasibility of on-field automated irrigation processes, allowing to reduce the water consumption, increasing the accuracy, and improving the efficiency, if compared to manually managed irrigations. Overall, the experimental findings clearly highlight that a fully automated decision support system could enhance agricultural practices by enabling more precise resource management and by reducing water waste and farmers’ workload.


  • Open Access

    DARE to wear: digital health promotion and disease prevention using wearable devices

    by Ilaria D’AScanio, Serena Moscato, Jose Albites-Sanabria, Maria Giulia Bacalini, Marianna Bellafiore, Laura Belli, Lisa Berti, Junayed Mohammad Bhuyan, Elisa Cainelli, Giacomo Cappon, et al.

    Global healthcare systems are under increasing strain to ensure long-term sustainability, necessitating a fundamental transition towards proactive, preventive healthcare strategies. Wearable sensor technologies are uniquely positioned to facilitate this paradigm shift by enabling continuous physiological and behavioral monitoring, facilitating early risk stratification, and enabling timely interventions. DigitAl lifelong pREvention (DARE) is a nationwide initiative conceived in Italy to develop and evaluate novel digital tools and services aimed at fostering innovative health promotion and disease prevention pathways, with wearable devices identified as a critical enabling technology. Herein, we present a comprehensive overview of 16 pilot investigations designed in DARE to rigorously assess the utility of wearable devices in the continuous monitoring of physical activity, joint mobility, sleep architecture, heart rate variability, nutrition, and glucose homeostasis across diverse health domains, including healthy aging, chronic disease prevention, and lifestyle modification programs targeting over 20000 participants. We meticulously detail the methodological characteristics of these study protocols, explicitly outlining the technical specifications, functional capabilities, and inherent limitations of the 18 distinct wearable devices (sourced from 12 different manufacturers) employed across these clinical trials. The breadth of applications and the heterogeneity of the target populations underscore the significant potential of wearable devices as an integral component of future digital health prevention strategies.


  • Open Access

    Air Quality Prediction via Embedded ML/DL and Quantized Models

    by Armin Mazinani, Daniele Antonucci, Danilo Pietro Pau, Luca Davoli, Gianluigi Ferrari

    In the people’s everyday life, one of the most significant (and unfortunately well-known) environmental problems with substantial impact and effects is air pollution. To this end, extensive research (e.g., conducted by the World Health Organization, WHO) has uncovered a non-negligible correlation between hazardous environments and airborne Particulate Matters (PMs), emphasizing the importance of monitoring pollution to mitigate its adverse effects. In fact, since PMs are microscopic particles of solid or liquid materials suspended in the atmosphere and have negative effects on both weather and precipitations, it has been highlighted that the diffusion of diseases through the inhalation or absorption of contaminants is a concerning outcome of polluted environments. Hence, focusing on the critical need to monitor air quality in common (livable and work) environments (e.g., airports, public transports, or any closed environment) using affordable, compact IoT devices capable of collecting and processing environmental data, as well as forecasting future air quality trends, in this paper we present an experimental evaluation of several Machine Learning (ML), Deep Learning (DL), and quantized DL models designed to accurately predict air pollution (in terms of PM2.5 concentration). Furthermore, we conduct a comparative analysis of state-of-the-art DL models, unveiling the trade-offs associated with each ML/DL model, and discussing their practical applications and potential deployment on tiny IoT devices for on-site computing. Our results show that Convolutional Neural Network-Bidirectional Gated Recurrent Unit (CNN-BiGRU) is the best performing model. According to the obtained results, using 8-bit quantization on CNN-BiGRU allows to achieve a 66% model size’s reduction with only a 1%drop in average prediction accuracy, making it a balanced and efficient choice for several applications.


  • Open Access

    Prediction of hop cone ripening through Internet of Things (IoT) and Machine Learning (ML) technologies

    by Giulia Oddi, Martina Galaverni, Laura Belli, Margherita Rodolfi, Luca Davoli, Gianluigi Ferrari, Tommaso Ganino


  • Open Access

    Artificial Intelligence for Wireless Communications: The InSecTT Perspective

    by Ramiro Samano Robles, Gowhar Javanmardi, Christoph Pilz, Przemyslaw Kwapisiewicz, Mateusz Rzymowski, Lukasz Kulas, Luca Davoli, Laura Belli, Gianluigi Ferrari, Bernd-Ludwig Wenning, et al.

    This paper presents an overview of how Artificial Intelligence (AI) and edge technology have been used to improve wireless connectivity in multiple industrial Use Cases (UCs) of the EU project “Intelligent Secure Trustable Things” (InSecTT). We present a brief introduction of the InSecTT framework for crossdomain architecture design, which targets UCs assisted by reusable and/or interoperable technical Building Blocks (BBs). These BBs constitute the construction bricks containing AI and supporting components that were used to build different UCs. The framework consists of multiple stages based on the processing of UC/BB requirements (RQs). These stages include: i) collection, ii) harmonization, iii) refinement, iv) classification, v) architecture alignment, and vi) functionality modeling of RQs. The most relevant results of these stages are discussed here, with emphasis on the need for a refined granularity of technical components with common functionalities named sub-building blocks (SBBs), where collaboration and cross-domain reusability were optimized. The design process shed light on how AI and SBBs were implemented across different layers and entities of our reference architecture for the Internet-of-Things (IoT), including the interfaces used for information exchange. This detailed interface analysis is expected to reveal issues such as bottlenecks, constraints, vulnerabilities, scalability problems, security threats, etc. This will, in turn, contribute to identifying design gaps in AI-enabled IoT systems. The paper summarizes the SBBs related to wireless connectivity, including a general description, implementation issues, comparison of results, adopted interfaces, and conclusions across domains.

  • Open Access

    Smart agriculture dataset in a tomato cultivation under different irrigation regimes

    by Laura Belli, Luca Davoli, Giulia Oddi, Luca Preite, Martina Galaverni, Tommaso Ganino, Gianluigi Ferrari


  • Open Access

    Deep Learning Algorithms for Cryptocurrency Price Prediction: A Comparative Analysis

    by Armin Mazinani, Luca Davoli, Gianluigi Ferrari

    Over the past years, cryptocurrencies have experienced a surge in popularity within the financial markets. As of today, besides being considered for investment purposes, they also serve as a widely accepted form of currency for everyday transactions. Due to the intricate characteristics of financial markets and their dependence on various factors to determine the prices of stocks and assets, the ability to predict such prices is crucial to make investment choices, especially in terms of cryptocurrencies. In this work, a comparative analysis on the suitability of Deep Learning (DL) algorithms (effective for time series forecasting) in predicting the price of three cryptocurrencies (namely Bitcoin, BTC; Ethereum, ETH; and Ripple, XRP) is assessed in terms of both short-term and long-term prediction accuracy. The results, evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and coefficient of determination (denoted as \(R^{2}\) ), reveal that: Transformer is generally more effective for short-term forecasts and also performs well for long-term predictions; Convolutional Neural Network-Recurrent Neural Network (CNN-RNN) demonstrates the lowest complexity in terms of number of Multiply and ACcumulate (MAC) operations; SimpleRNN has the fewest parameters and the smallest FLASH memory requirement. Overall, CNN-Gated Recurrent Unit (CNN-GRU) provides the best joint accuracy-complexity for predicting BTC and ETH prices, whereas CNN-RNN yields superior results for XRP price prediction.


  • Cardiac Autonomic Responsivity to Car Driving in a Real Context

    by Sara Gambetta, Veronica Mattioli, Luca Davoli, Laura Belli, Luca Carnevali, Riccardo Raheli, Gianluigi Ferrari, Andrea Sgoifo

    Driving a vehicle is a complex behavior requiring an optimal psychophysiological state in order to safely accomplish the task. The identification of human factors influencing driving behavior could enable the development of technologies for monitoring drivers and improving their comfort. This study aims to monitor the driver’s cardiac autonomic activity during real road driving, assuming psychophysiological variations depending on the driving context (e.g., ring road vs. urban driving) and the induction of psychosocial stress during urban driving activity. Moreover, we investigated the extent to which perceived stress, anxiety symptoms, habitual driving behavior as well as sex, and driving experience, influenced cardiac autonomic responses while driving. Heart rate (HR) and its variability (HRV) were assessed during a real driving task in a sample of thirty-eight drivers, including twenty males and eighteen females. Drivers’ psychometric characteristics were collected using questionnaires. HRV analysis revealed a significant overall autonomic activation while driving, independent of the exposure to external stressors. Neither sex nor driving experience seemed to affect cardiac autonomic response to driving. A significant positive correlation emerged between anxiety-stress symptoms and aberrant driving behavior. In summary, our results suggest that the overall driving task produced a notable impact on cardiac autonomic neural modulation. Understanding the factors that influence driving performance and modulate the resulting physiological response could provide a springboard for practical applications, such as the development of human-vehicle interaction monitoring systems for optimal psychophysiological arousal while driving.

  • Open Access

    RouMBLE: A Sink-Oriented Routing Protocol for BLE Mesh Networks

    by Luca Davoli, Massimo Moreni, Gianluigi Ferrari

    In Internet of Things (IoT)-like contexts, there is often the need to leverage traffic routing mechanisms among heterogeneous devices, especially when classical (and well-known) addressing paradigms cannot be adopted or supported by constrained IoT devices deployed on the field (e.g., due to memory footprint, internal limitations, etc.). This is even more true (and necessary) when nodes interact in unstructured networks (e.g., mesh-like) lacking a specific topology (e.g., exploiting flooding approaches to transfer information) and external “smart” devices should be allowed to interact with these networks. To this end, in this paper a multi-sink routing protocol, denoted as Routing on Mesh Bluetooth Low Energy (), is proposed. Our implementation relies on BLE advertisement channels and allows sink nodes to control topology formation and data collection (with both unicast and broadcast communications), with nodes identified with compressed addresses. A relevant experimental application to environmental lighting management is presented.