Alessia Lazzaro

Ms. Alessia Lazzaro

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 October 2024 in IEEE Open Journal of the Industrial Electronics Society
Verified
Journal Article
Open access
ALAlessia Lazzaro
Alessia Lazzaro
DDDoriana Marilena D'ADdona
Doriana Marilena D'ADdona
Massimo MerendaMassimo Merenda
Massimo Merenda
Industrial and technological evolution has led to the identification of different techniques and strategies that can best adapt to the needs of Manufacturing Industry 4.0. As industrial production has become more automated, the need for more efficient maintenance strategies has increased. Today, among the possible, several applications demonstrate how the Predictive Maintenance (PdM) strategy is the best performing. In fact, PdM makes it possible to predict an impending failure with high accuracy in order to intervene before failure occurs. This work focuses on the application of PdM technique in order to predict the type of chips produced by a lathe through a Machine Learning algorithm. Moreover, being our application a delay-sensitive one, to drastically decrease the time delay in prediction, our solution proposes the combination of PdM with the Edge Computing paradigm. To simulate this paradigm, the chosen Machine Learning models were deployed on STM microcontrollers obtaining both high accuracy (98%) and an inference time in the order of milliseconds.

2 November 2025 in Electronics
Verified
Journal Article
Open access
PDF file
ZDZohra Dakhia
Zohra Dakhia
ALAlessia Lazzaro
Alessia Lazzaro
Mohamed Riad SEBTIMohamed Riad SEBTI
Mohamed Riad SEBTI
Mariateresa RussoMariateresa Russo
Mariateresa Russo
Massimo MerendaMassimo Merenda
Massimo Merenda
This study presents a novel federated learning (FL) methodology implemented directly on STM32-based microcontrollers (MCUs) for energy-efficient smart irrigation. To the best of our knowledge, this is the first work to demonstrate end-to-end FL training and aggregation on real STM32 MCU clients (STM32F722ZE), under realistic energy and memory constraints. Unlike most prior studies that rely on simulated clients or high-power edge devices, our framework deploys lightweight neural networks trained locally on MCUs and synchronized via message queuing telemetry transport (MQTT) communication. Using a smart agriculture (SA) dataset partitioned by soil type, 7 clients collaboratively trained a model over 3 federated rounds. Experimental results show that MCU clients achieved competitive accuracy (70–82%) compared to PC clients (80–85%) while consuming orders of magnitude less energy. Specifically, MCU inference required only 0.95 mJ per sample versus 60–70 mJ on PCs, and training consumed ∼70 mJ per epoch versus nearly 20 J. Latency remained modest, with MCU inference averaging 3.2 ms per sample compared to sub-millisecond execution on PCs, a negligible overhead in irrigation scenarios. The evaluation also considers the payoff between accuracy, energy consumption, and latency through the Energy Latency Accuracy Index (ELAI). This integrated perspective highlights the trade-offs inherent in deploying FL on heterogeneous devices and demonstrates the efficiency advantages of MCU-based training in energy-constrained smart irrigation settings.

29 May 2025 in IEEE Access
Verified
Journal Article
Open access
ALAlessia Lazzaro
Alessia Lazzaro
Massimo MerendaMassimo Merenda
Massimo Merenda
FPFilippo Giammaria Praticò
Filippo Giammaria Praticò
Tyre-pavement friction influences transport safety, particularly in autonomous vehicles (AVs), where real-time friction estimation impacts critical decisions (e.g., stopping, changing lanes, and deceleration rate). Despite this, there is a lack of methods to have low cost and continuous monitoring. Consequently, this study proposes a novel system to continuously provide on-board microprocessors with updated estimates of tyre-pavement friction as a function of pavement, tyre, meteorological, and driving conditions. To this end, the following main tasks were carried out: 1) In task 1, the problem was modelled, and high-level equations to satisfy were set up for both AVs and human-driven vehicles (HDVs). 2) In task 2, a framework was set up to solve friction-related issues considering critical scenarios involving unconnected vehicles. 3) In task 3, a solution to continuously estimate tyre-pavement friction for AVs and HDVs was set up, and the consequences for connected and cooperative vehicles were analysed and modelled. 4) In task 4, the following conclusions were derived: A) The system herein set up and implemented has outstanding characteristics in terms of small cost. B) It allows for quasi-continuous monitoring of pavement friction, and both long-term and short-term changes can be detected. C) Both points are crucial due to the limitations of common friction monitoring systems in achieving high sampling frequencies. Future research will focus on increasing the number of sets of braking manoeuvres and on the optimisation of the hardware and software of the system to add to the existing onboard system.