Applying machine learning: a multi-role perspective

Machine (and deep) learning technologies are more and more present in several fields. It is undeniable that many aspects of our society are empowered by such technologies: web searches, content filtering on social networks, recommendations on e-commerce websites, mobile applications, etc., in addition to academic research. Moreover, mobile devices and internet sites, e.g., social networks, support the collection and sharing of information in real time. The pervasive deployment of the aforementioned technological instruments, both hardware and software, has led to the production of huge amounts of data. Such data has become more and more unmanageable, posing challenges to conventional computing platforms, and paving the way to the development and widespread use of the machine and deep learning. Nevertheless, machine learning is not only a technology. Given a task, machine learning is a way of proceeding (a way of thinking), and as such can be approached from different perspectives (points of view). This, in particular, will be the focus of this research. The entire work concentrates on machine learning, starting from different sources of data, e.g., signals and images, applied to different domains, e.g., Sport Science and Social History, and analyzed from different perspectives: from a non-data scientist point of view through tools and platforms; setting a problem stage from scratch; implementing an effective application for classification tasks; improving user interface experience through Data Visualization and eXtended Reality. In essence, not only in a quantitative task, not only in a scientific environment, and not only from a data-scientist perspective, machine (and deep) learning can do the difference.

On the motion planning & control of nonlinear robotic systems

In the last decades, we saw a soaring interest in autonomous robots boosted not only by academia and industry, but also by the ever in- creasing demand from civil users. As a matter of fact, autonomous robots are fast spreading in all aspects of human life, we can see them clean houses, navigate through city traffic, or harvest fruits and vegetables. Almost all commercial drones already exhibit unprecedented and sophisticated skills which makes them suitable for these applications, such as obstacle avoidance, simultaneous localisation and mapping, path planning, visual-inertial odometry, and object tracking. The major limitations of such robotic platforms lie in the limited payload that can carry, in their costs, and in the limited autonomy due to finite battery capability. For this reason researchers start to develop new algorithms able to run even on resource constrained platforms both in terms of computation capabilities and limited types of endowed sensors, focusing especially on very cheap sensors and hardware. The possibility to use a limited number of sensors allowed to scale a lot the UAVs size, while the implementation of new efficient algorithms, performing the same task in lower time, allows for lower autonomy. However, the developed robots are not mature enough to completely operate autonomously without human supervision due to still too big dimensions (especially for aerial vehicles), which make these platforms unsafe for humans, and the high probability of numerical, and decision, errors that robots may make. In this perspective, this thesis aims to review and improve the current state-of-the-art solutions for autonomous navigation from a purely practical point of view. In particular, we deeply focused on the problems of robot control, trajectory planning, environments exploration, and obstacle avoidance.

Identification of sunflower genotypes suitable for organic agriculture and development of analytical tools for their traceability

Thanks to the development and combination of molecular markers for the genetic traceability of sunflower varieties and a gas chromatographic method for the determination of the FAs composition of sunflower oil, it was possible to implement an experimental method for the verification of both the traceability and the variety of organic sunflower marketed by Agricola Grains S.p.A. The experimental activity focused on two objectives: the implementation of molecular markers for the routine control of raw material deliveries for oil extraction and the improvement and validation of a gas chromatographic method for the determination of the FAs composition of sunflower oil. With regard to variety verification and traceability, the marker systems evaluated were the following: SSR markers (12) arranged in two multiplex sets and SCAR markers for the verification of cytoplasmic male sterility (Pet1) and fertility. In addition, two objectives were pursued in order to enable a routine application in the industrial field: the development of a suitable protocol for DNA extraction from single seeds and the implementation of a semi-automatic capillary electrophoresis system for the analysis of marker fragments. The development and validation of a new GC/FID analytical method for the determination of fatty acids (FAME) in sunflower achenes to improve the quality and efficiency of the analytical flow in the control of raw and refined materials entering the Agricola Grains S.p.A. production chain. The analytical performances being validated by the newly implemented method are: linearity of response, limit of quantification, specificity, precision, intra-laboratory precision, robustness, BIAS. These parameters are used to compare the newly developed method with the one considered as reference - Commission Regulation No. 2568/91 and Commission Implementing Regulation No. 2015/1833. Using the combination of the analytical methods mentioned above, the documentary traceability of the product can be confirmed experimentally, providing relevant information for subsequent marketing.

Active control of sound radiated by vibrating surfaces on a McLaren supercar

Sound radiators based on forced vibrations of plates are becoming widely employed, mainly for active sound enhancement and noise cancelling systems, both in music and automotive environment. Active sound enhancement solutions based on electromagnetic shakers hence find increasing interest. Mostly diffused applications deal with active noise control (ANC) and active vibration control systems for improving the acoustic experience inside or outside the vehicle. This requires investigating vibrational and, consequently, vibro-acoustic characteristics of vehicles. Therefore, simulation and processing methods capable of reducing the calculation time and providing high-accuracy results, are strongly demanded. In this work, an ideal case study on rectangular plates in fully clamped conditions preceded a real case analysis on vehicle panels. The sound radiation generated by a vibrating flat or shallow surface can be calculated by means of Rayleigh’s integral. The analytical solution of the problem is here calculated implementing the equations in MATLAB. Then, the results are compared with a numerical model developed in COMSOL Multiphysics, employing Finite Element Method (FEM). A very good matching between analytical and numerical solutions is shown, thus the cross validation of the two methods is achieved. The shift to the real case study, on a McLaren super car, led to the development of a mixed analytical-numerical method. Optimum results were obtained with mini shakers excitement, showing good matching of the recorded SPL with the calculated one over all the selected frequency band. In addition, a set of directivity measurements of the hood were realized, to start studying the spatiality of sound, which is fundamental to active noise control systems.

Data surveillance for infectious diseases: models, complex networks and machine learning

In this thesis, we investigate the role of applied physics in epidemiological surveillance through the application of mathematical models, network science and machine learning. The spread of a communicable disease depends on many biological, social, and health factors. The large masses of data available make it possible, on the one hand, to monitor the evolution and spread of pathogenic organisms; on the other hand, to study the behavior of people, their opinions and habits. Presented here are three lines of research in which an attempt was made to solve real epidemiological problems through data analysis and the use of statistical and mathematical models. In Chapter 1, we applied language-inspired Deep Learning models to transform influenza protein sequences into vectors encoding their information content. We then attempted to reconstruct the antigenic properties of different viral strains using regression models and to identify the mutations responsible for vaccine escape. In Chapter 2, we constructed a compartmental model to describe the spread of a bacterium within a hospital ward. The model was informed and validated on time series of clinical measurements, and a sensitivity analysis was used to assess the impact of different control measures. Finally (Chapter 3) we reconstructed the network of retweets among COVID-19 themed Twitter users in the early months of the SARS-CoV-2 pandemic. By means of community detection algorithms and centrality measures, we characterized users’ attention shifts in the network, showing that scientific communities, initially the most retweeted, lost influence over time to national political communities. In the Conclusion, we highlighted the importance of the work done in light of the main contemporary challenges for epidemiological surveillance. In particular, we present reflections on the importance of nowcasting and forecasting, the relationship between data and scientific research, and the need to unite the different scales of epidemiological surveillance.

Numerical and physical modelling of axially loaded piles for offshore wind turbines

The scope of the thesis is to broaden the knowledge about axially loaded pipe piles, that can play as foundations for offshore wind turbines based on jacket structures. The goal of the work was pursued by interpreting experimental data on large-scale model piles and by developing numerical tools for the prediction of their monotonic response to tensile and compressive loads to failure. The availability of experimental results on large scale model piles produced in two different campaigns at Fraunhofer IWES (Hannover, Germany) represented the reference for the whole work. Data from CPTs, blow counts during installation and load-displacement curves allowed to develop considerations on the experimental results and comparison with empirical methods from literature, such as CPT-based methods and Load Transfer methods. The understanding of soil-structure interaction mechanisms has been involved in the study in order to better assess the mechanical response of the sand with the scope to help in developing predictive tools of the experiments. A lack of information on the response of Rohsand 3152 when in contact with steel was highlighted, so the necessity of better assessing its response was fulfilled with a comprehensive campaign of interface shear test. It was found how the response of the sand to ultimate conditions evolve with the roughness of the steel, which is a precious information to take account of when attempting the prediction of a pile capacity. Parallel to this topic, the work has developed a numerical modelling procedure that was validated on the available large-scale model piles at IWES. The modelling strategy is intended to build a FE model whose mechanical properties of the sand come from an interpretation of commonly available geotechnical tests. The results of the FE model were compared with other predictive tools currently used in the engineering practice.

Development of multicomponent models for the prediction of lubricant oil-fuel dilution in internal combustion engines

The current environmental crisis is forcing the automotive industry to face tough challenges for the Internal Combustion Engines development in order to reduce the emissions of pollutants and Greenhouse gases. In this context, in the last decades, the main technological solutions adopted by the manufacturers have been the direct injection and the engine downsizing, which led to the rising of new concerns related to the fuel-cylinder walls physical interaction. The fuel spray possibly impacts the cylinder liner wall, which is wetted by the lubricant oil thus causing the derating of the lubricant properties, increasing the oil consumption, and contaminating the lubricant oil in the crankcase. Also, concerning hydrogen fuelled internal combustion engines, it is likely that the high near-wall temperature, which is typical of the hydrogen flame, results in the evaporation of a portion of the lubricant oil, increasing its consumption. With regards on the innovative combustion systems and their control strategies, optical accessible engines are fundamental tools for experimental investigations on such combustion systems. Though, due to the optical measurement line, optical engines suffer from a high level of blow-by, which must be accounted for. In light of the above, this thesis work aims to develop numerical methodologies with the aim to build useful tools for supporting the design of modern engines. In particular, a one-dimensional modelling of the lubricant oil-fuel dilution and oil evaporation has been performed and coupled with an optimization algorithm to achieve a lubricant oil surrogate. Then, a quasi-dimensional blow-by model has been developed and validated against experimental data. Such model, has been coupled with CFD 3D simulations and directly implemented in CFD 3D. Finally, CFD 3D simulations coupled with the VOF method have been performed in order to validate a methodology for studying the impact of a liquid droplet on a solid surface.

Novel methods of state estimation and parameter identification for high-performance motorcycles

This research activity aims at providing a reliable estimation of particular state variables or parameters concerning the dynamics and performance optimization of a MotoGP-class motorcycle, integrating the classical model-based approach with new methodologies involving artificial intelligence. The first topic of the research focuses on the estimation of the thermal behavior of the MotoGP carbon braking system. Numerical tools are developed to assess the instantaneous surface temperature distribution in the motorcycle's front brake discs. Within this application other important brake parameters are identified using Kalman filters, such as the disc convection coefficient and the power distribution in the disc-pads contact region. Subsequently, a physical model of the brake is built to estimate the instantaneous braking torque. However, the results obtained with this approach are highly limited by the knowledge of the friction coefficient (μ) between the disc rotor and the pads. Since the value of μ is a highly nonlinear function of many variables (namely temperature, pressure and angular velocity of the disc), an analytical model for the friction coefficient estimation appears impractical to establish. To overcome this challenge, an innovative hybrid solution is implemented, combining the benefit of artificial intelligence (AI) with classical model-based approach. Indeed, the disc temperature estimated through the thermal model previously implemented is processed by a machine learning algorithm that outputs the actual value of the friction coefficient thus improving the braking torque computation performed by the physical model of the brake. Finally, the last topic of this research activity regards the development of an AI algorithm to estimate the current sideslip angle of the motorcycle's front tire. While a single-track motorcycle kinematic model and IMU accelerometer signals theoretically enable sideslip calculation, the presence of accelerometer noise leads to a significant drift over time. To address this issue, a long short-term memory (LSTM) network is implemented.

On the Democratization of AI: Bridging the Gap with AutoML

Riding the wave of recent groundbreaking achievements, artificial intelligence (AI) is currently the buzzword on everybody’s lips and, allowing algorithms to learn from historical data, Machine Learning (ML) emerged as its pinnacle. The multitude of algorithms, each with unique strengths and weaknesses, highlights the absence of a universal solution and poses a challenging optimization problem. In response, automated machine learning (AutoML) navigates vast search spaces within minimal time constraints. By lowering entry barriers, AutoML emerged as promising the democratization of AI, yet facing some challenges. In data-centric AI, the discipline of systematically engineering data used to build an AI system, the challenge of configuring data pipelines is rather simple. We devise a methodology for building effective data pre-processing pipelines in supervised learning as well as a data-centric AutoML solution for unsupervised learning. In human-centric AI, many current AutoML tools were not built around the user but rather around algorithmic ideas, raising ethical and social bias concerns. We contribute by deploying AutoML tools aiming at complementing, instead of replacing, human intelligence. In particular, we provide solutions for single-objective and multi-objective optimization and showcase the challenges and potential of novel interfaces featuring large language models. Finally, there are application areas that rely on numerical simulators, often related to earth observations, they tend to be particularly high-impact and address important challenges such as climate change and crop life cycles. We commit to coupling these physical simulators with (Auto)ML solutions towards a physics-aware AI. Specifically, in precision farming, we design a smart irrigation platform that: allows real-time monitoring of soil moisture, predicts future moisture values, and estimates water demand to schedule the irrigation.

LQG control policy design against agent identity privacy problem

This thesis project studies the agent identity privacy problem in the scalar linear quadratic Gaussian (LQG) control system. For the agent identity privacy problem in the LQG control, privacy models and privacy measures have to be established first. It depends on a trajectory of correlated data rather than a single observation. I propose here privacy models and the corresponding privacy measures by taking into account the two characteristics. The agent identity is a binary hypothesis: Agent A or Agent B. An eavesdropper is assumed to make a hypothesis testing on the agent identity based on the intercepted environment state sequence. The privacy risk is measured by the Kullback-Leibler divergence between the probability distributions of state sequences under two hypotheses. By taking into account both the accumulative control reward and privacy risk, an optimization problem of the policy of Agent B is formulated. The optimal deterministic privacy-preserving LQG policy of Agent B is a linear mapping. A sufficient condition is given to guarantee that the optimal deterministic privacy-preserving policy is time-invariant in the asymptotic regime. An independent Gaussian random variable cannot improve the performance of Agent B. The numerical experiments justify the theoretic results and illustrate the reward-privacy trade-off. Based on the privacy model and the LQG control model, I have formulated the mathematical problems for the agent identity privacy problem in LQG. The formulated problems address the two design objectives: to maximize the control reward and to minimize the privacy risk. I have conducted theoretic analysis on the LQG control policy in the agent identity privacy problem and the trade-off between the control reward and the privacy risk.Finally, the theoretic results are justified by numerical experiments. From the numerical results, I expected to have some interesting observations and insights, which are explained in the last chapter.

Integration of ROS for the control of industrial robotic systems in a proprietary environment

The paper deals with the integration of ROS, in the proprietary environment of the Marchesini Group company, for the control of industrial robotic systems. The basic tools of this open-source software are deeply studied to model a full proprietary Pick and Place manipulator inside it, and to develop custom ROS nodes to calculate trajectories; speaking of which, the URDF format is the standard to represent robots in ROS and the motion planning framework MoveIt offers user-friendly high-level methods. The communication between ROS and the Marchesini control architecture is established using the OPC UA standard; the tasks computed are transmitted offline to the PLC, supervisor controller of the physical robot, because the performances of the protocol don’t allow any kind of active control by ROS. Once the data are completely stored at the Marchesini side, the industrial PC makes the real robot execute a trajectory computed by MoveIt, so that it replicates the behaviour of the simulated manipulator in Rviz. Multiple experiments are performed to evaluate in detail the potential of ROS in the planning of movements for the company proprietary robots. The project ends with a small study regarding the use of ROS as a simulation platform. First, it is necessary to understand how a robotic application of the company can be reproduced in the Gazebo real world simulator. Then, a ROS node extracts information and examines the simulated robot behaviour, through the subscription to specific topics.

Methods for monitoring Machine Learning models

The following thesis aims to investigate the issues concerning the maintenance of a Machine Learning model over time, both about the versioning of the model itself and the data on which it is trained and about data monitoring tools and their distribution. The themes of Data Drift and Concept Drift were then explored and the performance of some of the most popular techniques in the field of Anomaly detection, such as VAE, PCA, and Monte Carlo Dropout, were evaluated.