Construction and test of a cosmic ray telescope based on CMS Drift Tube chambers and data acquisition prototypes of the Phase-2 upgrade

In this thesis work, a cosmic-ray telescope was set up in the INFN laboratories in Bologna using smaller size replicas of CMS Drift Tubes chambers, called MiniDTs, to test and develop new electronics for the CMS Phase-2 upgrade. The MiniDTs were assembled in INFN National Laboratory in Legnaro, Italy. Scintillator tiles complete the telescope, providing a signal independent of the MiniDTs for offline analysis. The telescope readout is a test system for the CMS Phase-2 upgrade data acquisition design. The readout is based on the early prototype of a radiation-hard FPGA-based board developed for the High Luminosity LHC CMS upgrade, called On Board electronics for Drift Tubes. Once the set-up was operational, we developed an online monitor to display in real-time the most important observables to check the quality of the data acquisition. We performed an offline analysis of the collected data using a custom version of CMS software tools, which allowed us to estimate the time pedestal and drift velocity in each chamber, evaluate the efficiency of the different DT cells, and measure the space and time resolution of the telescope system.

The atmospheric stable boundary layer: Data analysis and comparison with similarity theories

The objective of this dissertation is to study the structure and behavior of the Atmospheric Boundary Layer (ABL) in stable conditions. This type of boundary layer is not completely well understood yet, although it is very important for many practical uses, from forecast modeling to atmospheric dispersion of pollutants. We analyzed data from the SABLES98 experiment (Stable Atmospheric Boundary Layer Experiment in Spain, 1998), and compared the behaviour of this data using Monin-Obukhov's similarity functions for wind speed and potential temperature. Analyzing the vertical profiles of various variables, in particular the thermal and momentum fluxes, we identified two main contrasting structures describing two different states of the SBL, a traditional and an upside-down boundary layer. We were able to determine the main features of these two states of the boundary layer in terms of vertical profiles of potential temperature and wind speed, turbulent kinetic energy and fluxes, studying the time series and vertical structure of the atmosphere for two separate nights in the dataset, taken as case studies. We also developed an original classification of the SBL, in order to separate the influence of mesoscale phenomena from turbulent behavior, using as parameters the wind speed and the gradient Richardson number. We then compared these two formulations, using the SABLES98 dataset, verifying their validity for different variables (wind speed and potential temperature, and their difference, at different heights) and with different stability parameters (zita or Rg). Despite these two classifications having completely different physical origins, we were able to find some common behavior, in particular under weak stability conditions.

Data analysis and simulations of VIRTIS venus spectra

VIRTIS, a bordo di Venus Express, è uno spettrometro in grado di operare da 0.25 a 5 µm. Nel periodo 2006-2011 ha ricavato un'enorme mole di dati ma a tutt'oggi le osservazioni al lembo sono poco utilizzate per lo studio delle nubi e delle hazes, specialmente di notte. Gli spettri al lembo a quote mesosferiche sono dominati dalla radianza proveniente dalle nubi e scatterata in direzione dello strumento dalle hazes. L'interpretazione degli spettri al lembo non può quindi prescindere dalla caratterizzazione dell'intera colonna atmosferica. L'obiettivo della tesi è di effettuare un’analisi statistica sulle osservazioni al nadir e proporre una metodologia per ricavare una caratterizzazione delle hazes combinando osservazioni al nadir e al lembo. La caratterizzazione delle nubi è avvenuta su un campione di oltre 3700 osservazioni al nadir. È stato creato un ampio dataset di spettri sintetici modificando, in un modello iniziale, vari parametri di nube quali composizione chimica, numero e dimensione delle particelle. Un processo di fit è stato applicato alle osservazioni per stabilire quale modello potesse descrivere gli spettri osservati. Si è poi effettuata una analisi statistica sui risultati del campione. Si è ricavata una concentrazione di acido solforico molto elevata nelle nubi basse, pari al 96% in massa, che si discosta dal valore generalmente utilizzato del 75%. Si sono poi integrati i risultati al nadir con uno studio mirato su poche osservazioni al lembo, selezionate in modo da intercettare nel punto di tangenza la colonna atmosferica osservata al nadir, per ricavare informazioni sulle hazes. I risultati di un modello Monte Carlo indicano che il numero e le dimensioni delle particelle previste dal modello base devono essere ridotte in maniera significativa. In particolare si osserva un abbassamento della quota massima delle hazes rispetto ad osservazioni diurne.

Study of the calibration method of pressure-velocity probes and its application in a field of progressive plane waves

This dissertation presents a calibration procedure for a pressure velocity probe. The dissertation is divided into four main chapters. The first chapter is divided into six main sections. In the firsts two, the wave equation in fluids and the velocity of sound in gases are calculated, the third section contains a general solution of the wave equation in the case of plane acoustic waves. Section four and five report the definition of the acoustic impedance and admittance, and the practical units the sound level is measured with, i.e. the decibel scale. Finally, the last section of the chapter is about the theory linked to the frequency analysis of a sound wave and includes the analysis of sound in bands and the discrete Fourier analysis, with the definition of some important functions. The second chapter describes different reference field calibration procedures that are used to calibrate the P-V probes, between them the progressive plane wave method, which is that has been used in this work. Finally, the last section of the chapter contains a description of the working principles of the two transducers that have been used, with a focus on the velocity one. The third chapter of the dissertation is devoted to the explanation of the calibration set up and the instruments used for the data acquisition and analysis. Since software routines were extremely important, this chapter includes a dedicated section on them and the proprietary routines most used are thoroughly explained. Finally, there is the description of the work that has been done, which is identified with three different phases, where the data acquired and the results obtained are presented. All the graphs and data reported were obtained through the Matlab® routine. As for the last chapter, it briefly presents all the work that has been done as well as an excursus on a new probe and on the way the procedure implemented in this dissertation could be applied in the case of a general field.

Computer vision in aseptic pharmaceutical manufacturing: a deep learning-based approach to vial inspection and pack formation during a lyophilizer loading cycle

Vision systems are powerful tools playing an increasingly important role in modern industry, to detect errors and maintain product standards. With the enlarged availability of affordable industrial cameras, computer vision algorithms have been increasingly applied in industrial manufacturing processes monitoring. Until a few years ago, industrial computer vision applications relied only on ad-hoc algorithms designed for the specific object and acquisition setup being monitored, with a strong focus on co-designing the acquisition and processing pipeline. Deep learning has overcome these limits providing greater flexibility and faster re-configuration. In this work, the process to be inspected consists in vials’ pack formation entering a freeze-dryer, which is a common scenario in pharmaceutical active ingredient packaging lines. To ensure that the machine produces proper packs, a vision system is installed at the entrance of the freeze-dryer to detect eventual anomalies with execution times compatible with the production specifications. Other constraints come from sterility and safety standards required in pharmaceutical manufacturing. This work presents an overview about the production line, with particular focus on the vision system designed, and about all trials conducted to obtain the final performance. Transfer learning, alleviating the requirement for a large number of training data, combined with data augmentation methods, consisting in the generation of synthetic images, were used to effectively increase the performances while reducing the cost of data acquisition and annotation. The proposed vision algorithm is composed by two main subtasks, designed respectively to vials counting and discrepancy detection. The first one was trained on more than 23k vials (about 300 images) and tested on 5k more (about 75 images), whereas 60 training images and 52 testing images were used for the second one.

Spacecraft Collision Probability Calculation from Conjunction Data Message and High Precision Orbit Propagator

The increasing number of Resident Space Objects (RSOs) is a threat to spaceflight operations. Conjunction Data Messages (CDMs) are sent to satellite operators to warn for possible future collision and their probabilities. The research project described herein pushed forward an algorithm that is able to update the collision probability directly on-board starting from CDMs and the state vector of the hosting satellite which is constantly updated thanks to an onboard GNSS receiver. A large set of methods for computing the collision probability was analyzed in order to find the best ones for this application. The selected algorithm was then tested to assess and improve its performance. Finally, parts of the algorithm and external software were implemented on a Raspberry Pi 3B+ board to demonstrate the compatibility of this approach with computational resources similar to those typically available onboard modern spacecraft.

Private sensors and crowdsourced rainfall data: accuracy and effectiveness for monitoring severe events and mapping pluvial flooding in urban areas

The increasing number of extreme rainfall events, combined with the high population density and the imperviousness of the land surface, makes urban areas particularly vulnerable to pluvial flooding. In order to design and manage cities to be able to deal with this issue, the reconstruction of weather phenomena is essential. Among the most interesting data sources which show great potential are the observational networks of private sensors managed by citizens (crowdsourcing). The number of these personal weather stations is consistently increasing, and the spatial distribution roughly follows population density. Precisely for this reason, they perfectly suit this detailed study on the modelling of pluvial flood in urban environments. The uncertainty associated with these measurements of precipitation is still a matter of research. In order to characterise the accuracy and precision of the crowdsourced data, we carried out exploratory data analyses. A comparison between Netatmo hourly precipitation amounts and observations of the same quantity from weather stations managed by national weather services is presented. The crowdsourced stations have very good skills in rain detection but tend to underestimate the reference value. In detail, the accuracy and precision of crowd- sourced data change as precipitation increases, improving the spread going to the extreme values. Then, the ability of this kind of observation to improve the prediction of pluvial flooding is tested. To this aim, the simplified raster-based inundation model incorporated in the Saferplaces web platform is used for simulating pluvial flooding. Different precipitation fields have been produced and tested as input in the model. Two different case studies are analysed over the most densely populated Norwegian city: Oslo. The crowdsourced weather station observations, bias-corrected (i.e. increased by 25%), showed very good skills in detecting flooded areas.

Data analysis and hydrological modelling of rainfall and discharge extremes along the Tione River: implications for the design of water structures

There are many natural events that can negatively affect the urban ecosystem, but weather-climate variations are certainly among the most significant. The history of settlements has been characterized by extreme events like earthquakes and floods, which repeat themselves at different times, causing extensive damage to the built heritage on a structural and urban scale. Changes in climate also alter various climatic subsystems, changing rainfall regimes and hydrological cycles, increasing the frequency and intensity of extreme precipitation events (heavy rainfall).  From an hydrological risk perspective, it is crucial to understand future events that could occur and their magnitude in order to design safer infrastructures. Unfortunately, it is not easy to understand future scenarios as the complexity of climate is enormous.  For this thesis, precipitation and discharge extremes were primarily used as data sources. It is important to underline that the two data sets are not separated: changes in rainfall regime, due to climate change, could significantly affect overflows into receiving water bodies. It is imperative that we understand and model climate change effects on water structures to support the development of adaptation strategies.   The main purpose of this thesis is to search for suitable water structures for a road located along the Tione River. Therefore, through the analysis of the area from a hydrological point of view, we aim to guarantee the safety of the infrastructure over time.   The observations made have the purpose to underline how models such as a stochastic one can improve the quality of an analysis for design purposes, and influence choices.

Multiple pathogen detection using nanoplasmonic sensing

Opportunistic diseases caused by Human Immunodeficiency Virus (HIV) and Hepatitis B Virus (HBV) is an omnipresent global challenge. In order to manage these epidemics, we need to have low cost and easily deployable platforms at the point-of-care in high congestions regions like airports and public transit systems. In this dissertation we present our findings in using Localized Surface Plasmon Resonance (LSPR)-based detection of pathogens and other clinically relevant applications using microfluidic platforms at the point-of-care setting in resource constrained environment. The work presented here adopts the novel technique of LSPR to multiplex a lab-on-a-chip device capable of quantitatively detecting various types of intact viruses and its various subtypes, based on the principle of a change in wavelength occurring when metal nano-particle surface is modified with a specific surface chemistry allowing the binding of a desired pathogen to a specific antibody. We demonstrate the ability to detect and quantify subtype A, B, C, D, E, G and panel HIV with a specificity of down to 100 copies/mL using both whole blood sample and HIV-patient blood sample discarded from clinics. These results were compared against the gold standard Reverse Transcriptase Polymerase Chain Reaction (RT-qPCR). This microfluidic device has a total evaluation time for the assays of about 70 minutes, where 60 minutes is needed for the capture and 10 minutes for data acquisition and processing. This LOC platform eliminates the need for any sample preparation before processing. This platform is highly multiplexable as the same surface chemistry can be adapted to capture and detect several other pathogens like dengue virus, E. coli, M. Tuberculosis, etc.

Synthetic generation of amyloid PET images by non-linear dimensionality reduction inversion

Privacy issues and data scarcity in PET field call for efficient methods to expand datasets via synthetic generation of new data that cannot be traced back to real patients and that are also realistic. In this thesis, machine learning techniques were applied to 1001 amyloid-beta PET images, which had undergone a diagnosis of Alzheimer’s disease: the evaluations were 540 positive, 457 negative and 4 unknown. Isomap algorithm was used as a manifold learning method to reduce the dimensions of the PET dataset; a numerical scale-free interpolation method was applied to invert the dimensionality reduction map. The interpolant was tested on the PET images via LOOCV, where the removed images were compared with the reconstructed ones with the mean SSIM index (MSSIM = 0.76 ± 0.06). The effectiveness of this measure is questioned, since it indicated slightly higher performance for a method of comparison using PCA (MSSIM = 0.79 ± 0.06), which gave clearly poor quality reconstructed images with respect to those recovered by the numerical inverse mapping. Ten synthetic PET images were generated and, after having been mixed with ten originals, were sent to a team of clinicians for the visual assessment of their realism; no significant agreements were found either between clinicians and the true image labels or among the clinicians, meaning that original and synthetic images were indistinguishable. The future perspective of this thesis points to the improvement of the amyloid-beta PET research field by increasing available data, overcoming the constraints of data acquisition and privacy issues. Potential improvements can be achieved via refinements of the manifold learning and the inverse mapping stages during the PET image analysis, by exploring different combinations in the choice of algorithm parameters and by applying other non-linear dimensionality reduction algorithms. A final prospect of this work is the search for new methods to assess image reconstruction quality.