Dynamic Human Robot Interaction Framework Using Deep Learning and Robot Operating System (ROS): a practical approach

Trying to explain to a robot what to do is a difficult undertaking, and only specific types of people have been able to do so far, such as programmers or operators who have learned how to use controllers to communicate with a robot. My internship's goal was to create and develop a framework that would make that easier. The system uses deep learning techniques to recognize a set of hand gestures, both static and dynamic. Then, based on the gesture, it sends a command to a robot. To be as generic as feasible, the communication is implemented using Robot Operating System (ROS). Furthermore, users can add new recognizable gestures and link them to new robot actions; a finite state automaton enforces the users' input verification and correct action sequence. Finally, the users can create and utilize a macro to describe a sequence of actions performable by a robot.

Implementation of a non ground meta interpreter for defeasible logic

Human reasoning is a fascinating and complex cognitive process that can be applied in different research areas such as philosophy, psychology, laws and financial. Unfortunately, developing supporting software (to those different areas) able to cope such as complex reasoning it’s difficult and requires a suitable logic abstract formalism. In this thesis we aim to develop a program, that has the job to evaluate a theory (a set of rules) w.r.t. a Goal, and provide some results such as “The Goal is derivable from the KB5 (of the theory)”. In order to achieve this goal we need to analyse different logics and choose the one that best meets our needs. In logic, usually, we try to determine if a given conclusion is logically implied by a set of assumptions T (theory). However, when we deal with programming logic we need an efficient algorithm in order to find such implications. In this work we use a logic rather similar to human logic. Indeed, human reasoning requires an extension of the first order logic able to reach a conclusion depending on not definitely true6 premises belonging to a incomplete set of knowledge. Thus, we implemented a defeasible logic7 framework able to manipulate defeasible rules. Defeasible logic is a non-monotonic logic designed for efficient defeasible reasoning by Nute (see Chapter 2). Those kind of applications are useful in laws area especially if they offer an implementation of an argumentation framework that provides a formal modelling of game. Roughly speaking, let the theory is the set of laws, a keyclaim is the conclusion that one of the party wants to prove (and the other one wants to defeat) and adding dynamic assertion of rules, namely, facts putted forward by the parties, then, we can play an argumentative challenge between two players and decide if the conclusion is provable or not depending on the different strategies performed by the players. Implementing a game model requires one more meta-interpreter able to evaluate the defeasible logic framework; indeed, according to Göedel theorem (see on page 127), we cannot evaluate the meaning of a language using the tools provided by the language itself, but we need a meta-language able to manipulate the object language8. Thus, rather than a simple meta-interpreter, we propose a Meta-level containing different Meta-evaluators. The former has been explained above, the second one is needed to perform the game model, and the last one will be used to change game execution and tree derivation strategies.

Rheological Characterization of Bitumen and Asphalt including the effect of Added-recycled Material

This thesis evaluates the rheological behaviour of asphalt mixtures and the corresponding extracted binders from the mixtures containing different amounts of Reclaimed Asphalt (RA). Generally, the use of RA is limited to certain amounts. The study materials are Stone Mastic Asphalts including a control sample with 0% RA, and other samples with RA rates of 30%, 60% and 100%. Another set of studied mixtures are Asphalt Concretes (AC) types with again a control mix having 0% RA rate and the other mixtures designs containing 30%, 60% and 90% of reclaimed asphalt which also contain additives. In addition to the bitumen samples extracted from asphalt mixes, there are bitumen samples directly extracted from the original RA. To characterize the viscoelastic behaviour of the binders, Dynamic Shear Rheometer (DSR) tests were conducted on bitumen specimens. The resulting influence of the RA content in the bituminous binders are illustrated through master curves, black diagrams and Cole-Cole plots with regressing these experimental data by the application of the analogical 2S2P1D and the analytical CA model. The advantage of the CA model is in its limited number of parameters and thus is a simple model to use. The 2S2P1D model is an analogical rheological model for the prediction of the linear viscoelastic properties of both asphalt binders and mixtures. In order to study the influence of RA on mixtures, the Indirect Tensile Test (ITT) has been conducted. The master curves of different mixture samples are evaluated by regressing the test data points to a sigmoidal function and subsequently by comparing the master curves, the influence of RA materials is studied. The thesis also focusses on the applicability and also differences of CA model and 2S2P1D model for bitumen samples and the sigmoid function for the mixtures and presents the influence of the RA rate on the investigated model parameters.

Studies of CMS data access patterns with machine learning techniques

This thesis presents a study of the Grid data access patterns in distributed analysis in the CMS experiment at the LHC accelerator. This study ranges from the deep analysis of the historical patterns of access to the most relevant data types in CMS, to the exploitation of a supervised Machine Learning classification system to set-up a machinery able to eventually predict future data access patterns - i.e. the so-called dataset “popularity” of the CMS datasets on the Grid - with focus on specific data types. All the CMS workflows run on the Worldwide LHC Computing Grid (WCG) computing centers (Tiers), and in particular the distributed analysis systems sustains hundreds of users and applications submitted every day. These applications (or “jobs”) access different data types hosted on disk storage systems at a large set of WLCG Tiers. The detailed study of how this data is accessed, in terms of data types, hosting Tiers, and different time periods, allows to gain precious insight on storage occupancy over time and different access patterns, and ultimately to extract suggested actions based on this information (e.g. targetted disk clean-up and/or data replication). In this sense, the application of Machine Learning techniques allows to learn from past data and to gain predictability potential for the future CMS data access patterns. Chapter 1 provides an introduction to High Energy Physics at the LHC. Chapter 2 describes the CMS Computing Model, with special focus on the data management sector, also discussing the concept of dataset popularity. Chapter 3 describes the study of CMS data access patterns with different depth levels. Chapter 4 offers a brief introduction to basic machine learning concepts and gives an introduction to its application in CMS and discuss the results obtained by using this approach in the context of this thesis.

Development of a relationship between physical parameters of tsunami waves and tsunami intensity based on macroscopic effects and its application to the Italian tsunami database

Tsunamis are rare events. However, their impact can be devastating and it may extend to large geographical areas. For low-probability high-impact events like tsunamis, it is crucial to implement all possible actions to mitigate the risk. The tsunami hazard assessment is the result of a scientific process that integrates traditional geological methods, numerical modelling and the analysis of tsunami sources and historical records. For this reason, analysing past events and understanding how they interacted with the land is the only way to inform tsunami source and propagation models, and quantitatively test forecast models like hazard analyses. The primary objective of this thesis is to establish an explicit relationship between the macroscopic intensity, derived from historical descriptions, and the quantitative physical parameters measuring tsunami waves. This is done first by defining an approximate estimation method based on a simplified 1D physical onshore propagation model to convert the available observations into one reference physical metric. Wave height at the coast was chosen as the reference due to its stability and independence of inland effects. This method was then implemented for a set of well-known past events to build a homogeneous dataset with both macroseismic intensity and wave height. By performing an orthogonal regression, a direct and invertible empirical relationship could be established between the two parameters, accounting for their relevant uncertainties. The target relationship is extensively tested and finally applied to the Italian Tsunami Effect Database (ITED), providing a homogeneous estimation of the wave height for all existing tsunami observations in Italy. This provides the opportunity for meaningful comparison for models and simulations, as well as quantitatively testing tsunami hazard models for the Italian coasts and informing tsunami risk management initiatives.