Politecnio die Bari - Catalogo di prodotti della Ricerca
Not a member yet
36616 research outputs found
Sort by
BRAINEX: A Systematic Framework for CNN Models Evaluation and XAI Methods Comparison in Brain Age Prediction
No full textThis study presents a systematic framework for evaluating convolutional neural networks (CNNs) in the context of brain age prediction, with a focus on interpretability through explainable AI (XAI) methods. Brain age prediction has advanced significantly using 2D and 3D CNNs, which provide high predictive accuracy. However, the complexity of these models creates challenges for clinical interpretation. Our framework not only assesses model performance but also provides deeper insights into various aspects of CNN models, including how the selected backgrounds can influence XAI values. By facilitating multisite data analysis, the framework helps identify the impact of site-specific characteristics on model behavior. The results underscore the importance of local explanations and highlight the need for careful interpretation when using population-level saliency maps
Enhancing Last-Mile Logistics: AI-Driven Fleet Optimization, Mixed Reality, and Large Language Model Assistants for Warehouse Operations
Full text linkDue to the rapid expansion of e-commerce and urbanization, Last-Mile Delivery (LMD) faces increasing challenges related to cost, timeliness, and sustainability. Artificial intelligence (AI) techniques are widely used to optimize fleet management, while augmented and mixed reality (AR/MR) technologies are being adopted to enhance warehouse operations. However, existing approaches often treat these aspects in isolation, missing opportunities for optimization and operational efficiency gains through improved information visibility across different roles in the logistics workforce. This work proposes the adoption of novel technological solutions integrated in an LMD framework that combines AI-based optimization of shipment allocation and vehicle route planning with a knowledge graph (KG)-driven decision support system. Additionally, the paper discusses the exploitation of relevant recent tools, including large language model (LLM)-powered conversational assistants for managers and operators and MR-based headset interfaces supporting warehouse operators by providing real-time data and enabling direct interaction with the system through virtual contextual UI elements. The framework prioritizes the customizability of AI algorithms and real-time information sharing between stakeholders. An experiment with a system prototype in the Apulia region is presented to evaluate the feasibility of the system in a realistic logistics scenario, highlighting its potential to enhance coordination and efficiency in LMD operations. The results suggest the usefulness of the approach while also identifying benefits and challenges in real-world applications
Angiolo Mazzoni in Colombia: un patrimonio di immagini: architettura di immagini - immagini di architettura
No full textLa storia, personale e professionale, dell’architetto Angiolo Mazzoni, noto per i suoi edifici pubblici ferroviari e postali disseminati in tutto il territorio italiano, dimentica spesso di raccontare il quindicennio da lui trascorso in Colombia, non attribuendo così il giusto valore ad una opera che, seppur perlopiù non realizzata, contribuisce
a ricostruire l’immagine poliedrica di una figura di spessore del Novecento come quella del Mazzoni.
La ricerca ha quindi voluto far emergere il portato dell’opera colombiana dell’architetto italiano, restituendo il giusto valore all’architettura rappresentata, ovvero riconoscendo al cospicuo corpus grafico lascito dell’architetto il valore di disegno teorico.
Il periodo colombiano ha, di fatto, rappresentato per il Mazzoni la sua personale “stagione dell’architettura di carta”, restituendoci probabilmente le sue più floride riflessioni, nonché, secondo una personale visione e lettura, le sue più personali interpretazioni del fatto architettonico.
L’indagine ha voluto sviluppare una attività di tipo interpretativa favorita dal disegno e dal rilievo, al fine di rendere, attraverso i mezzi più opportuni, non una mera restituzione pratica dell’architettura, quanto piuttosto la sua essenza spaziale anche nel suo statuto di disegno. L’analisi del materiale conservato tra Italia e Colombia ha, allora, permesso di ricostruire una geografia dei progetti colombiani, restituendo altresì l’estrema capacità dell’architetto di confrontarsi con i temi architettonici più disparati, a differenza di quanto si è abituati in relazione all’opera italiana.
Scrivere il racconto dei progetti e, più in generale, dell’esperienza colombiana è significato scrivere un racconto fatto di rappresentazioni grafiche, di immagini: i disegni. Quello che ne è risultato è stata una narrazione visuale che attraverso il riordino, la catalogazione e la selezione dei disegni autografi ha permesso di raccontare il quindicennio sudamericano, palesando al contempo l’assenza di architettura intesa quale fatto tangibile.
Lo studio condotto ha privilegiato l’indagine dello spazio, nella sua declinazione di spazio disegnato, intendendo il disegno quale tessuto di idee, ma anche quale luogo di sperimentalismo; un disegno teorico disvelatore dell’interiorità del suo autore, ma anche di una realtà autonoma, propria e caratteristica del disegno di architettura.
In questo senso nel ricco panorama di elaborati grafici redatti da Mazzoni si è letta una significatività rispetto al valore del disegno: il vasto patrimonio visuale è riuscito a restituire una architettura dell’immagine in grado di rendersi altresì palinsesto infinito di forme che attingono dalla memoria, nonché sistema disvelatore di un particolare modo di vedere e far vedere il fatto architettonico, seppur rimasto su carta
Linking lubricant oil contamination to pre-ignition events in hydrogen engines–The HyLube mechanism
Full text linkThe practical use of hydrogen as a fuel for Internal Combustion Engines (ICEs) poses unresolved challenges. However, many issues concerning combustion control and emissions do not originate from the fuel itself but rather from the inevitable interference of lubricant oil in the combustion process. This study investigates the catalytic action of lubricant oil to elucidate certain pre-ignition events in hydrogen ICEs that are not yet fully understood. In pursuit of this goal, a numerical tool is developed, which can also contribute to improving the design of future powertrains. In the first part of the work, a reduced chemical model named the “HyLube” reaction mechanism was developed for investigating hydrogen-lubricant interactions. A single surrogate species, i.e., n-hexadecane (n-C16H34), was selected to mimic the chemical properties of lubricant oil, based on prior research. Extensive validation using experimental data from the literature demonstrated the accuracy and reliability of the model. With 133 species and 2074 reactions, the kinetic mechanism is well-suited for integration into CFD simulations of engines. In the second part, using the “HyLube” mechanism, predictions were made to determine if lubricant oil contamination affects hydrogen ICE operation. Zero-dimensional simulations quantified the impact of adding the lubricant oil surrogate to H2/air mixtures on charge reactivity across a range of engine-like thermodynamic conditions, encompassing compression ratio and equivalence ratio variations. The findings indicated significant effects, especially at lower temperatures, suggesting potential hazards before combustion initiation
Statistical and Machine Learning-Based Imaging with Long Pulse Thermography for the Detection of Non-standardised Defects in CFRP Composites
No full textA new approach to the limit analysis of masonry structures as assemblies of rigid-plastic blocks with frictional sliding contacts
No full textDue to the lack of an associated flow in frictional sliding, the collapse load of an assembly of rigid-plastic blocks with frictional contacts depends on the loading history and cannot be studied in the context of standard limit analysis. In this work a theorem is developed to characterize the collapse states of these mechanical systems. In particular, the concept of kinematically admissible load multiplier is extended to these systems and it is established that if a load multiplier is both kinematically and statically admissible then it is a collapse multiplier. Based on this result, various plane masonry structures are examined, considering both Coulomb frictional sliding at blocks interfaces and plastic deformations. To take into account the lack of uniqueness of the load multiplier, its minimum and maximum values are determined. The numerical applications demonstrate the potential of this approach for assessing the collapse load of masonry-like structures
Design optimization and one-shot 3D printing of a bio-inspired soft robot capable of counter-flexural motion at low actuation pressures
No full textThis work presents the design, fabrication, and numerical analysis of a soft actuator with
counter-flexural motion fabricated in thermoplastic polyurethan (TPU), aiming at mimicking the
serpentine movement of snakes. Printing parameters were chosen to allow the one-shot 3D printing of the soft robot, ensuring its air tightness up to 200 [kPa]. The hyperelastic behavior of TPU was characterized by means of uniaxial tensile tests of samples produced with the abovementioned optimal printing settings. Non-linear finite element simulations were conducted to perform a
design optimization of the pneumatic chambers; the purpose is to minimize the actuation pressure needed to achieve a prescribed bending. The experimentally measured deflections were compared with numerical results, showing a very good agreement. Practical relapses are in the field of
medical devices where the proposed low-pressure actuated design will allow safer human-robot interactions
LUMEN: Leveraging Large Language Models for Dynamic Ontologies in Wind Energy Domain Analysis
No full textSoil-vegetation-atmosphere interaction for engineering applications: recent multi-scale and multi-disciplinary insights
No full textThe paper presents some of both recent and previous outcomes on soil-vegetation-atmosphere mechanisms and processes. Contributions from several authors across different scientific fields, each focusing on specific aspects of soil and plant behaviour under variable climatic conditions, highlight the urgency of multi- and inter-disciplinarily research aimed at the investigation, analysis, and modelling of soil-vegetation-atmosphere (SVA) interaction. The paper focuses on different aspects of such interaction, which plays the crucial role of a thermo-hydro-mechanical boundary condition acting at the ground surface and controlling the transient behaviour of every geotechnical system. Some specific remarks and novel aspects of such interaction, analysed at increasing levels of complexity, are depicted in the paper through a multi-disciplinary point of view and according to a multi-scale methodology