Politecnio die Bari - Catalogo di prodotti della Ricerca
Not a member yet
36616 research outputs found
Sort by
Old but Sold? Innovation Through Tradition Strategy for Export and the Role of Family Involvement
No full textDespite the notable body of research, the family firm (FF) internationalization literature has overlooked the role of innovation strategies in explaining FFs' export performance. We focus on the innovation through tradition (ITT) strategy—specifically, the degree to which a firm leverages its firm-specific, mature (i.e., past) knowledge in the innovation search and recombination process. This strategy is particularly relevant for FFs as it can have ambivalent effects on export performance. On the one hand, it may ease the liability of foreignness; on the other, it could exacerbate it, creating tensions that firms must carefully navigate. Drawing from the socioemotional wealth perspective, we argue that FFs will be more prone to adopt an ITT strategy. However, the extent to which they will be able to reap the advantages or suffer the constraints of such a strategy for export depends on the particular type of family governance. Specifically, we contend that family managers will have a positive moderating effect on the relationship between the degree to which a firm leverages firm-specific mature knowledge and export intensity thanks to their direct involvement and operational control over innovation activities. Conversely, family owners lacking this direct involvement will have a negative moderating effect on this relationship. Our analyses, based on a global longitudinal sample of 134 listed firms in the automotive and pharma/biotech industries observed from 2008 to 2020, support our hypotheses. Our results contribute to the nexus of the FF internationalization and FF innovation literature streams, the ITT research in FFs, and the broader internationalization literature
Sideband Peak Count Technique for Monitoring Bond–Slip Behavior Between Reinforcement Systems and Masonry
No full textAbstract
The use of Fabric-Reinforced Cementitious Matrix (FRCM) composites has become a cornerstone in strengthening several types of structures, ranging from reinforced concrete structures to historical masonry constructions. It is crucial to experimentally assess the effectiveness of the application of the reinforcing layer. Indeed, the performance of FRCM reinforcing interventions highly depends on the bonding at the matrix–fiber and matrix–substrate interfaces. Therefore, the experimental characterization of FRCM bond behavior is essential for designing and ensuring the durability of reinforcement systems. To address this, we have proposed using an innovative nonlinear ultrasonic method: the Side-band Peak Count (SPC) technique. Specifically, the SPC approach has been applied to analyze ultrasonic test results obtained during Double-Lap Shear Tests (DLT) conducted on masonry specimens reinforced with B-FRCM (basalt fiber textile FRCM). We show that it is possible to correlate nonlinear ultrasonic parameters with the shear stress distribution at the reinforcement–masonry interface. Additionally, a relatively new nonlinear ultrasonic parameter, the SPC index, is shown to be effective in monitoring the evolution of the shear stress–slip relationship at this interface, a crucial aspect for understanding the mechanics of the reinforcement–substrate interaction. The nonlinear ultrasonic results have been compared with the results of DLT debonding tests to validate the proposed methodology. The effectiveness of the SPC technique is investigated and discussed. Finally, we have developed a robust numerical model to analyze the bond behavior between the reinforcement and masonry. The numerical model is valuable for both the design of experimental tests and the validation of the experimental results
Innovative interfaces for motor assessment
Get PDFMotor abilities may be reduced in different conditions, such as neuromotor diseases, the
physiological aging, or work-related musculoskeletal disorders. In the clinical realm, motor
assessment is useful to measure the severity level, thus supporting physicians’ decision for
diagnostic, prognostic, and rehabilitative purposes; on the other hand, an objective evaluation
of the motor performance could allow for recording the exertion perceived by the subject
while executing an industrial task.
However, the clinical scales may suffer from subjectivity, since they are observation
based and related to the specific background of different clinicians; the perceived exertion
is conventionally estimated by self-ratings, which may be biased by the user’s psychology.
Therefore, quantitative and objective measurement of motor abilities are needed to pursue
more generalizable outcomes in both clinical and occupational applications.
The purpose of this Ph.D. thesis is to illustrate the research works carried out during the
conceptualization, design, implementation, and validation of frameworks for the quantitative
assessment of motor capabilities by means of innovative interfaces based on serious game,
deep-learning methods, and robotic exoskeletons.
Serious games promote the engagement of the experimental subjects, thus keeping them
motivated during the execution of multiple repetitions of the experimental tasks. Deep
Learning models allow for the automatic recognition of motor patterns from raw data for a
variety of applications, including human activity recognition and pathological gait recognition.
Robotic exoskeletons can support humans in the execution of repetitive and exhausting motor
tasks, thus preventing the injuries connected with work-related musculoskeletal disorders.
The applications considered span from visuomotor adaptation to activity recognition and
power augmentation. Tasks under consideration concerned the locomotion on a treadmill
while controlling a virtual avatar, the execution of activities of daily living, as well as static
and dynamic lifting tasks that are typical of an industrial scenario.
Apaucity has been found in the different domains of the scientific literature to which
the works presented in this thesis belong. As regards visuomotor adaptation, a few works
implemented SGs to elicit sensorimotor learning in children during a walking task; therefore, more investigations are needed to perform a SG-aided evaluation of visuomotor adaptation
capabilities of people in developmental age during locomotion tasks. With regards to human
activity recognition, a minority of studies trained DL models with inertial data related to a
separate execution of human motor actions and tested them with data acquired during an
uninterrupted execution of the same activities; furthermore, there exist a few works exploiting
simulated gait disorders to train DL models for recognizing pathological gaits. In the field
of occupational exoskeletons, a gap has been found about the validation of such robotic
devices with motor tasks resembling those of an industrial scenario with both conventional
electromyographic measures and innovative methods based on graph theory.
Therefore, the technical contributions of this thesis include the conceptualization of
a locomotor task for the evaluation of visuomotor adaptation based on serious game; the
validation of a framework based on deep-learning for the recognition of human activities
executed in an uninterrupted sequence; the preliminary validation of a similar workflow
addressing the recognition of mimicked gait disorders; the validation of an occupational
exoskeleton assisting humans during industrial-like motor tasks by means of both traditional
electromyographic measures and innovative approaches based on muscle networks.
This thesis work is organized into two parts, each of which is divided in sections including
an introduction and the works belonging to the specific context. More in detail, Chapter 1
is focused on applications for clinical purposes, giving an introduction of the objective and
the technical contribution of the thesis in such context. Therefore, Section 1.2 describes the
contributions proposed in the context of visuomotor adaptation assessment based on serious
game, together with the related state-of-the-art. Sections 1.3 and 1.4 present the scientific
literature and the contributions proposed in the context of activity recognition, concerning
the classification of human motor actions performed continuously and pathological walking
patterns simulated by healthy subjects, respectively. On the other hand, Chapter 2 is focused
on applications for occupational purposes, giving an introduction of the objective and the
technical contribution of the thesis in such context. Hence, Sections 2.2 and 2.3 report the
state-of-art and the contributions proposed in the realm of the validation of occupational ex
oskeleton with conventional electromyographic metrics and functional connectivity analysis
based on muscle networks, respectively. Lastly, final remarks and considerations are drawn
in Chapter
A Matheuristic Approach for Delivery Planning and Dynamic Vehicle Routing in Logistics 4.0
No full textIn distribution logistics, the planning of vehicles’ routes and vehicles’ loads are traditionally managed separately, despite these activities are correlated. This often leads to various re-designs to make the routes and load plans compatible and applicable in practice. Moreover, the planned routes, which are static by definition, cannot always cope with unexpected events. Traffic congestion, vehicle failures, adverse meteorological conditions, and further undesired events can make the planned routes inapplicable and require vehicles’ re-routing. This results in lower service levels, undesired delays, and higher costs for logistics companies. With the aim of overcoming the above limitations, this work proposes a novel approach based on a matheuristic algorithm that jointly solves the problem of delivery planning and dynamic vehicle routing to automate the delivery process in a logistics 4.0 perspective. The presented algorithm includes two different phases: the static phase, which is executed offline and in advance with respect to the delivery day, and the dynamic phase, which is executed in real-time to cope with unexpected events during the delivery. For the first phase, a matheuristic approach is defined to efficiently solve the combined vehicle routing and loading problems. Differently, for the second phase, a genetic algorithm is proposed to re-route vehicles in real-time, considering both the redefinition in real-time of the nominal trip and/or of the sequence of the customers to be visited. The algorithm is tested both on a literature benchmark and on a real dataset provided by an Italian logistics company. The obtained results show that, on the one hand, the proposed algorithm can automatically provide feasible solutions that minimise travel costs, total travelled distance, and empty space on the vehicles; on the other hand, it can ensure in real-time effective re-routing solutions in case of unexpected events occurring during delivery. Note to Practitioners—This work is motivated by the need for facilitating the operations of planning and routing deliveries in the external logistics sector. We propose an algorithm that automatically generates feasible routing and loading plans for a set of Transport Units (TUs) (i.e., the static phase), and then updates in real-time the nominal route in case of unexpected events (i.e., the dynamic phase). More specifically, the first phase of the algorithm takes as input the set of different clients, the list of products packed into bins (i.e., standard packing units) to be delivered to each client, and the set of transport units available for the deliveries, and provides as output the number and type of TUs to be used, the composition of the bins in each transport unit, and the corresponding route, while optimising the space occupation in each TU and the travel costs. The second phase, instead, takes as input the nominal routes computed in the first phase and, in case of unexpected events (e.g., accidents, slowdowns, etc.) affecting one or more routes, it re-routes the involved trucks guaranteeing the maximum efficiency in regards to travel cost, travel time, and quality of service. The adoption of this algorithm by logistic companies supports the automation of the delivery process and drastically improves the efficiency of logistic operations, with particular regard to the number of used TUs, costs, safety of goods, and customers’ satisfaction
Control of hyperbolic and parabolic equations on networks and singular limits
No full textWe study the controllability properties of transport equations and of parabolic equations with vanishing diffusivity posed on a tree-shaped net-work. Using a control localized on the exterior nodes, we obtain a null-controllability result for both systems. The hyperbolic proof relies on the method of characteristics; while the parabolic one on duality arguments and Carleman inequalities. In particular, we estimate the cost of the null-controllability of advection-diffusion equations with diffusivity ε > 0 and study its asymptotic behavior when ε → 0+. More specifically, we show that the cost of null-controllability decays exponentially for a time sufficiently large and ex-plodes for short times. The core of the proof consists in proving an observabil-ity estimate keeping track of the viscosity parameter by relying on a suitable Carleman inequality
Time-dependent modelling of short-term variability in the TeV-blazar VER J0521+211 during the major flare in 2020
No full textSelf-repairing graphite protective layer has been discovered as a suitable protective layer in blast furnace (BF) hearth in recent years. In the current study, actual samples of self-repairing graphite protective layer taken from a commercial BF were analyzed in detail. The results revealed that the hot face of graphite protective layer exhibits a distinct white graphite luster, with large areas of graphite adhering to the surface. Along the direction of its formation, the sample displays a striped pattern with alternating layers. The graphite is strip-shaped, it is relatively coarse and unevenly distributed. The coarse graphite runs in the same direction, unlike graphite in molten iron which has no fixed direction in a chaotic state. The formation process of selfrepairing graphite protective layer can be concluded, graphite precipitates at the interface through heterogeneous nucleation. Crystal nuclei often preferentially adhere to the surface of these impurities to form, owing to the fact that the nucleation energy of heterogeneous nucleation is lower than that of homogeneous nucleation. Titanium is discovered during the observation of microscopic morphology of graphite protective layer, graphite protective layer is more robust due to the strengthening effect of titanium. Titanium strengthening mechanism of self-repairing graphite protective layer is summarized, the strengthening mechanism can be divided into four steps. TiC particles are dispersed around graphite, which reduces the difficulty of the orientation of flake graphite growth. The presence of TiC increases the growth rate of crystals. The four steps are cyclically performed, so the self-repairing graphite protective layer can precipitate layer by layer through titanium strengthening mechanism, which serves to protect the carbon brick in BF hearth
Event-driven control of hybrid systems using Batches Petri nets: Application to high throughput manufacturing systems
No full textMulti-modal temporal action segmentation for manufacturing scenarios
No full textIndustrial robots have become prevalent in manufacturing due to their advantages of accuracy, speed, and reduced operator fatigue. Nevertheless, human operators play a crucial role in primary production lines. This study focuses on the temporal segmentation of human actions, aiming to identify the physical and cognitive
behavior of operators working alongside collaborative robots. While existing literature explores temporal action segmentation datasets, there is a lack of evaluation for manufacturing tasks. This work assesses six state-of-the art action segmentation models using the Human Action Multi-Modal Monitoring in Manufacturing (HA4M)
dataset, where subjects assemble an industrial object in realistic manufacturing scenarios. By employing Cross Subject and Cross-Location approaches, the study not only demonstrates the effectiveness of these models in industrial settings but also introduces a new benchmark for evaluating generalization across different subjects and locations. The evaluation further includes new videos in simulated industrial locations, assessed with both fully and semi-supervised learning approaches. The findings reveal that the Multi-Stage Temporal Convolutional Network ++ (MS-TCN++) and the Action Segmentation Transformer (ASFormer) architectures
exhibit high performance in supervised and semi-supervised learning settings, also using new data, particularly when trained with Skeletal features, advancing the capabilities of temporal action segmentation in real-world manufacturing environments. This research lays the foundation for addressing video activity understanding challenges in manufacturing and presents opportunities for future investigations
Performance comparison of machine learning algorithms for the estimation of blood pressure using photoplethysmography
Get PDFThis paper deals with an in-depth performance analysis on the estimation of Systolic Blood Pressure (SBP) and Diastolic Blood Pressure (DBP) by using features from the photoplethysmography (PPG) signal enhanced using the Maximal Overlap Discrete Wavelet Transform (MODWT), to train many machine learning (ML) regression models, including eXtreme Gradient Boost (XGBoost). The impact of different features selections methods, ML methods and training set sizes has been analyzed. One result is on the achievable improvements using features extracted from MODWT enhanced PPG signals. The most significant features have been selected using three different algorithms, namely RReliefF, Minimum Redundancy Maximum Relevance (MRMR) and Correlation based Feature Selection (CFS). This comparison has been critical to underline that the exploitation of the new features allows to improve SBP and DBP estimation. Moreover, the authors have trained several ML algorithms to provide a comparison of their accuracy and training time, showing the Pareto frontier. RReliefF and MRMR selections algorithms, and several ML algorithms such as XGBoost, Gaussian Process Regression (GPR) and Ensemble stood out for their performance, with a different compromise between prediction error and training time. In addition, a further result has been obtained by varying the dimension of the dataset to understand the impact on Root Mean Square Error (RMSE) for models that have shown better performance, giving an empirical relationship on achievable RMSE as a function of training set size. From that relationship it has been extrapolated an upper boundary of the set size over which no further RMSE improvements are expected
Extension of the dynamic Thickened Flame model for partially-premixed multi-fuel multi-injection combustion and application to an ammonia–hydrogen swirled flame
No full textAn extension of the widely-used Thickened Flame model for Large Eddy Simulations (TFLES) is proposed to take into account multi-fuel multi-injection combustion processes. Indeed, in such systems the local variations of the fuel composition and the local evolution of the equivalence ratio issued by differential diffusion effects inferred by the potential different nature of the used fuels need to be addressed for a proper use of the standard TFLES model. To do so, the extended model relies on a description of the differentiated fuel injections mixing that is computed from a transported mixture fraction tracing the spatial evolution of each fuel stream. This allows to both incorporate local fuel composition inhomogeneities into the combustion model and a proper parameterization of the flame sensor or turbulent combustion model. The proposed modeling is then used to predict the ammonia–air swirling flame stabilized by multiple hydrogen injection holes and operated at Cardiff University. To perform this specific simulations, a dedicated and novel analytically reduced chemical kinetics model for NH3-H2-N2/air combustion is also derived and validated at gas turbine operating conditions and for multiple ammonia–hydrogen binary fuel blends as well as ternary fuel blends derived from ammonia decomposition. The results obtained by the use of the novel Multi-Fuel TFLES model (MF-TFLES) are compared against the conventional TFLES predictions and assessed via OH* chemiluminescence and NO Planar Laser Induced Fluorescence (NO-PLIF) experimental data. As shown, the proposed modeling improves the flame shape and structure prediction by assuring the correct local application of the artificial flame thickening coherently, taking into consideration the multi-fuel complex mixing process, a feature that the standard TFLES model cannot consider hindering the quality of the prediction. Novelty and significance statement The novelty of this research can be summarized in two statements: 1. Extension of the widely used TFLES turbulent combustion model to consider flames where differential diffusion is present, as well as, partially-premixed multi-fuel multi-injection problems. 2. A novel NH3-H2-N2 analytically reduced chemistry suited for reactive LES. This work is significant because it allows to simulate unconventional burner setups which are being explored for decarbonized fuels, such as NH3 and the highly diffusive H2, in an effort to reduce the impact of power generation on climate change. Furthermore, high-fidelity modeling, such as the approach presented in this study, will allow the scientific community to understand the pollutant production of decarbonized fuels in the gas turbines context, thereby contributing to the development of adapted technological solutions towards carbon-free power generation solutions