PROPOSTE DI TESI

PROPOSTE DI TESI2019-01-16T15:05:23+00:00

La Fondazione LINKS, tra le azioni volte a formare personale qualificato nei diversi ambiti mette a disposizione un certo numero di proposte di tesi di laurea da svolgere presso i laboratori della Fondazione, sotto l’attenta supervisione dei ricercatori LINKS.

Se sei uno studente universitario giunto al termine del tuo percorso di studi e sei interessato a scoprire cosa può offrire la Fondazione per aiutarti ad intraprendere una carriera professionale nel mondo della ricerca, controlla qui sotto l’elenco delle proposte di tesi attualmente disponibili e manda la tua candidatura all’indirizzo e-mail che trovi specificato all’interno dell’annuncio. Sarai contattato dal nostro personale amministrativo per concordare un colloquio conoscitivo.

…allora cosa aspetti? Scegli la tesi che preferisci e contattaci!

Image processing tools for autonomous driving2019-01-30T14:23:57+00:00

Thesis Code: 19002
Thesis Type: Master Thesis for Telecommunication Engineering, Computer Engineering or related fields

Research Area: Multi-Layer Wireless Solutions

Requirements

  • Excellent software programming skills
  • Strong experience with Python/bash scripting and Linux environment
  • Strong experience with C/C++ programming languages
  • Basic knowledge of image processing
  • Basic knowledge of machine learning

Motivation
Accurate and time-efficient image processing tools are essential for autonomous driving vehicles. A timely detection of other vehicles, road users and obstacles can ensure to the autonomous vehicle the capability to perform safe road manoeuvres. Several cutting-edge tools are now being proposed by several research actors and companies. In the autonomous driving context, it is necessary to find the best trade-off between accuracy and time performance given the resources-constrained environment. A thorough evaluation is needed as well as a customization of the tools for the autonomous driving context.

Objective 
The aim of the thesis is to evaluate different image processing tools for finding the most suited for the automotive driving context. Customization of the selected tool is the final target of the thesis.
The first part of the thesis will be devoted to the analysis of cutting-edge image processing tools for selecting the most suitable one for the specific targeted scenario. The evaluation will be based on different performance criteria. In the second part of the thesis, the student will customize the selected image processing tool for enhancing its performance for the context of autonomous driving.
The student will have the possibility to work with real-data coming from the field in an informal cutting-edge research laboratory using the latest available technologies on these fields.

Contact: send a resume with attached the list of exams to daniele.brevi@linksfoundation.com specifying the thesis code and title.

Object tracking and trajectory prediction for safety enhancement of autonomous driving2019-01-30T14:21:24+00:00

Thesis Code: 19001
Thesis Type: Master Thesis for Telecommunication Engineering, Computer Engineering or related fields

Research Area: Multi-Layer Wireless Solutions

Requirements

  • Excellent software programming skills
  • Strong experience with Python/bash scripting and Linux environment
  • Strong experience with C/C++ programming languages
  • Basic knowledge of image processing
  • Basic knowledge of machine learning

Motivation
The knowledge of the surrounding environment is crucial for the connected and autonomous vehicles. These vehicles must timely know the position and the trajectories of other road users to perform safe road manoeuvres. If other road users cannot communicate such information, each vehicle has to rely on its own sensors to identify other cars, bicycles and pedestrians and to foresee their trajectories.  A significant support can be provided from the road-side infrastructure. In critical places, fixed sensors can continuously sense the surrounding environment to identify vehicles, pedestrians, other road users and obstacles and the infrastructure can communicate the gathered information to the connected vehicles.

Objective 
The aim of the thesis is to develop a framework for the identification of road users and for the prediction of their trajectories.
The first part of the thesis will be devoted to the analysis of state-of-art objects tracking methods. In the second part of the thesis, the student will develop a real object tracking system exploiting available cutting-edge image processing tool. Final step is the definition of trajectory prediction algorithm exploiting the gathered information.
The student will have the possibility to work with real-data coming from the field in an informal cutting-edge research laboratory using the latest available technologies on these fields.

Contact: send a resume with attached the list of exams to daniele.brevi@linksfoundation.com specifying the thesis code and title.

Deep Semantic Analysis of Public Procurement Contracts2019-01-16T15:00:42+00:00
Thesis Code: 18012

Thesis Type: Thesis in Computer Science, Data Engineering, Computer Engineering, Mathematical Engineering, Data Science

Research Area: Innovation Development

Requirements:
• Experience with Python and/or Java and/or Node.js
• Basic knowledge of modular development
• Beginner of (or willing to learn quickly) deep learning and natural language processing
• Curiosity-driven mindset.

Description
Public procurement contracts are a rich source of knowledge necessary for seizing the efforts in participating to public procurement calls. However, contracts are usually available in textual format making harder the task of extracting structured information automatically and being used in automated systems. This thesis will focus on extracting structured information from those documents such as specific dates, unique identifiers (VAT id, protocol numbers, telephone number), named entities (places, people, business entities, products). In this thesis the undergraduate will study and experiment with deep learning and natural language processing techniques that are the core of the Artificial Intelligence stack, by understanding the intrinsic semantics of document and identifying and linking pivotal information found in the text to an external database.

The thesis will be structured as follows:
• state-of-the-art analysis of text processing techniques
• problem formulation: objective function, data structures and resources to be used
• algorithm design and prototyping
• in-lab testing verification with real data and measurement of the performance of the approach.

The thesis will be co-tutored with Synapta Srl, a Spin-off of Politecnico di Torino. It will be an opportunity to work also with the Synapta team experimenting with real data. The undergraduate will benefit from being immersed in a existing start-up environment while applying scientific experimental practises learned in ISMB. At the end of the thesis, the undergraduate will be familiar with deep learning and natural language processing techniques, and she/he will acquire an understanding of the public-procurement domain. As additional benefit, she/he will proficiently use control version systems, continuous integration systems, remote deploying and monitoring techniques.

Contact: send a resume with attached the list of exams to giuseppe.rizzo@linksfoundation.com specifying the thesis code and title.

Intelligent Data Crawler of Unstructured Open Data2019-01-16T15:00:05+00:00
Thesis Code: 18011

Thesis Type: Thesis in Computer Science, Data Engineering, Computer Engineering, Mathematical Engineering, Data Science

Research Area: Innovation Development

Requirements:
• Experience with Python and/or Java and/or Node.js
• Basic knowledge of modular development
• Beginner of (or willing to learn quickly) machine learning and natural language processing
• Curiosity-driven mindset

Description
Italian public administration websites contain a lot of resources published as open data. However, administrations have multiple websites and each has its own semantic structure making harder to autonomous crawlers retrieving the necessary information. The aim of this thesis project is to develop an intelligent data crawler able to fetch specific types of resources across multiple file formats from selected sources. Relatively low precision and high recall is expected. The crawler should be able to detect relevant resources using state-of-the-art techniques based on machine learning and natural language processing techniques that are the core of the Artificial Intelligence stack.

The undergraduate will study and experiment with technologies for:
• extracting semantics from web resources
• understanding the content
• discriminating about the value of the retrieved content.

The thesis will be structured as follows:
• state-of-the-art analysis of information retrieval
• problem formulation: objective function, data structures and resources to be used
• algorithm design and prototyping
• in-lab testing verification with real data and measurement of the performance of the approach.

The thesis will be co-tutored with Synapta Srl, a Spin-off of Politecnico di Torino. It will be an opportunity to work also with the Synapta team experimenting with real data. The undergraduate will benefit from being immersed in a existing start-up environment while applying scientific experimental practises learned in ISMB. At the end of the thesis, the undergraduate will be familiar with machine learning and natural language processing techniques, and she/he will acquire an understanding of the public-procurement domain. As additional benefit, she/he will proficiently use control version systems, continuous integration systems, remote deploying and monitoring techniques.

Contact: send a resume with attached the list of exams to giuseppe.rizzo@linksfoundation.com specifying the thesis code and title.

Multi-Agent System Development2019-01-16T14:59:00+00:00
Thesis Code: 18010

Thesis Type: Master Thesis for Computer Science or Computer Engineering students or equivalent

Research Area: Pervasive Technologies

Required Skills:
• Good written English level
• Java, C++ or equivalent Object-Oriented language
• Version control system (GIT, SVN)
• Proactive mindset, problem-solving oriented
• Python or equivalent scripting languages will be a plus

Description
Motivation: One of the hottest topics of the last century is the environment preservation. Particularly, over the last years, the attention of our society has been focused, in the industrial context, on the need of reducing both waste production and toxic emissions. Internet-Of-Things (IOT) has become a well-established reality over the past years: by introducing very advanced devices and exposing heterogeneous service it has been used in various scenarios and areas of interest, i.e. the industry domain. IOT devices have already been adopted in factories, contributing to a more sustainable production. Such an interest adoption of IOT solutions has attracted ISMB’s interest, leading to the investigation on how Multi Agent Systems can be used in conjunction with IOT technologies to pursue the objectives previously described. A multi-agent system is a computerized system composed of multiple interacting intelligent agents, thought for resolving problems that would result difficult or impossible to solve for an individual agent. MAS can be used in the most disparate domains, ranging from Distributed Constraints Optimization (DCO) problems to coordination and delegation of computational tasks.

Objectives: This thesis will focus on the analysis and development of a Multi-Agent System (MAS) interacting with an IOT system, providing asynchronous data in a Fabric-Of-the-Future context (e.g. waste management collection and dispatching). After a state of the art survey of MAS, the student will build a system in either a collaborative or antagonist scenario, exploiting standard communication protocols (such as AMQP, MQTT, etc.). Furthermore, the student will analyze existing algorithms and interaction protocols (e.g. consensus algorithms, Vickrey auctions, etc.) in order to solve complex distributed problems such as searching the Pareto optimality in a virtual marketplace scenario.

Contact: send a resume with attached the list of exams to claudio.pastrone@linksfoundation.com or giuseppe.pacelli@linksfoundation.com specifying the thesis code and title.

Multi-Agent System Algorithms2019-01-16T14:58:05+00:00
Thesis Code: 18009

Thesis Type: Master Thesis for Computer Science or Computer Engineering students or equivalent

Research Area: Pervasive Technologies

Required Skills:
• Good written English level
• Java, C++ or equivalent Object-Oriented language
• Version control system (GIT, SVN)
• Proactive mindset, problem-solving oriented
• Python or equivalent scripting languages will be a plus

Description
Motivation: one of the hottest topics of the last century is the environment preservation. Particularly, over the last years, the attention of our society has been focused, in the industrial context, on the need of reducing both waste production and toxic emissions. Internet-Of-Things (IOT) has become a well-established reality over the past years: by introducing very advanced devices and exposing heterogeneous service it has been used in various scenarios and areas of interest, i.e. the industry domain. IOT devices have already been adopted in factories, contributing to a more sustainable production. Such an interest adoption of IOT solutions has attracted ISMB’s interest, leading to the investigation on how Multi Agent Systems can be used in conjunction with IOT technologies to pursue the objectives previously described. A multi-agent system is a computerized system composed of multiple interacting intelligent agents, thought for resolving problems that would result difficult or impossible to solve for an individual agent. MAS can be used in the most disparate domains, ranging from Distributed Constraints Optimization (DCO) problems to coordination and delegation of computational tasks.

Objectives: This thesis will focus on the analysis and development of a Multi-Agent System (MAS) interacting with an IOT system, providing asynchronous data in a Fabric-Of-the-Future context (e.g. waste management collection and dispatching). After a state of the art survey of MAS, the student will build a system in either a collaborative or antagonist scenario, exploiting standard communication protocols (such as AMQP, MQTT, etc. ). Furthermore, the student will analyze existing algorithms and interaction protocols (e.g. consensus algorithms, Vickrey auctions, etc.) in order to solve complex distributed problems such as searching the Pareto optimality in a virtual marketplace scenario.

Contact: send a resume with attached the list of exams to claudio.pastrone@linksfoundation.com or giuseppe.pacelli@linksfoundation.com specifying the thesis code and title.

Multi-service Federation for IoT, Smart City Collective Awareness Platforms2019-01-16T14:57:23+00:00
Thesis Code: 18008

Thesis Type: Master Thesis for Computer Science or Computer Engineering students or equivalent

Research Area: Pervasive Technologies

Required Skills:
• Good written English level
• Java, C++ or equivalent Object-Oriented language
• Version control system (GIT, SVN)
• Proactive mindset, problem-solving oriented
• Python or equivalent scripting languages will be a plus

Description
Motivation: The number of mass-market application and services leveraging position information is continuously growing. The increasing availability of GNSS-enabled devices eases the growth of location-based solutions (LBS) also beyond personal applications. Furthermore, the concrete global business of the Internet of Things (IoT) is influencing almost all application domains, thanks to the massive amounts data generated by autonomous, connected objects all over the world. Exponential growth in the number of deployed IoT devices is foreseen to exceed the value of 4000 billion€ before 2025 – with a growing share of devices offering position information. Because interconnected devices collect and share large amounts of sensitive information, including personal position information, it is crucial to safeguard the privacy and security of users and the data exchanged throughout private and public networks.

Objectives: Study and implementation of state-of-the-art multi-service federation approaches to facilitate and foster interoperability of LBS with Internet of Things (IoT), Smart City, and Collective Awareness Platforms. Such integration aims to provide LBS outside the traditional bound of location-based applications and businesses. Special attention will be given to interoperability considering e-security aspects (ranging from secure certificate-based services to light-weight identity services such as the ones based on OpenID or OAuth 2.0). Federation methodologies will be based, again, on open standard and technologies, to ensure maximum outreach and facilitate business exploitation of the developed methodologies.

Contact: send a resume with attached the list of exams to claudio.pastrone@linksfoundation.com or giuseppe.pacelli@linksfoundation.com specifying the thesis code and title.

Infrastructure for device e-security and trust2019-01-16T14:56:35+00:00
Thesis Code: 18007

Thesis Type: Master Thesis for Computer Science or Computer Engineering students or equivalent

Research Area: Pervasive Technologies

Required Skills:
• Good written English level
• Java, C++ or equivalent Object-Oriented language
• Version control system (GIT, SVN)
• Proactive mindset, problem-solving oriented
• Python or equivalent scripting languages will be a plus

Description
Motivation: The number of mass-market application and services leveraging position information is continuously growing. The increasing availability of GNSS-enabled devices eases the growth of location-based solutions (LBS) also beyond personal applications. Furthermore, the concrete global business of the Internet of Things (IoT) is influencing almost all application domains, thanks to the massive amounts data generated by autonomous, connected objects all over the world. Exponential growth in the number of deployed IoT devices is foreseen to exceed the value of 4000 billion€ before 2025 – with a growing share of devices offering position information. Because interconnected devices collect and share large amounts of sensitive information, including personal position information, it is crucial to safeguard the privacy and security of users and the data exchanged throughout private and public networks.

Objectives: design and implement a cloud-based e-security infrastructure needed to provide end-to-end device e-security and trust. The thesis will start from a study of the state of the art and a subsequent implementation of security AAA (Authentication, Authorization, and Accounting) mechanisms for data and privacy protection from mobile devices up to cloud services. Special attention will be given to well-established techniques oriented to Public Key Infrastructures (PKIs) based on certificates’ exchange. Furthermore, open communication protocols will be preferred, able to natively provide data encryption mechanisms – such as Transport Layer Security (TLS) – and authentication techniques –such as Simple Authentication and Security Layer (SASL)-.

Contact: send a resume with attached the list of exams to claudio.pastrone@linksfoundation.com or giuseppe.pacelli@linksfoundation.com specifying the thesis code and title.

Simulations for autonomous driving2019-01-16T14:55:50+00:00
Thesis Code: 18006

Thesis type: Master Thesis for Telecommunication Engineering, Computer Engineering or related fields

Research Area: Multi-Layer Wireless Solutions

Requirements:
• Good software programming skills
• Good experience with Python programming language
• Knowledge of deep learning, convolutional neural networks and simulations is a plus
• Some background in wireless communications will be a plus

Description:
Motivation: Autonomous driving is one of the most complex and interesting topics in the ICT landscape. One of the main challenges for AD is to collect enough data to test and validate autonomous algorithms. Moreover, many different sensors are involved in data collection to feed these algorithms. The behavior of these sensors should be modelled to simulate the real world perception of an AD car. Wireless communications are one of the most promising sensor that will permit to increase the safety of driving, thanks to low-delay direct communication between vehicles, often called V2V (Vehicle-to-vehicle).

Objective: At the beginning of the thesis, the student should analyze the panorama of open source projects aimed at the simulation, testing and validation of sensors and autonomous driving algorithms. The most promising ones should be carefully analyzed, so to understand their maturity level: the second part of the thesis work will be about some tests on such platforms, in particular, delving into the Deep Learning features. As an optional activity, depending on the effort spent on the main activities mentioned before, the student might develop a module to simulate vehicles communications and integrate it in the chosen platform.

Contact: send a resume with attached the list of exams to daniele.brevi@linksfoundation.com specifying the thesis code and title.

AI-based Technologies to Understand Clinical Notes2019-01-16T14:54:50+00:00
Thesis Code: 18003

Thesis Type: Thesis in Computer Science, Data Engineering, Computer Engineering, Mathematical Engineering, Data Science

Research Area: Innovation Development

Requirements
• Experience with Python and/or Java
• Basic knowledge of modular development
• Beginner of (or willing to learn quickly) machine learning
• Curiosity-driven mindset.

Description
The digital transformation that healthcare has undergone has encouraged the generation of a large quantity of digital clinical notes. Majority of those notes contain unstructured information which complicates the search, analysis, and the understanding of the content. The automated analysis and understanding of those notes is of now one of the biggest challenges in healthcare.

In this thesis the undergraduate will study and experiment with AI-based technologies for:
• extracting and classify key information such as adverse events from clinical notes written in natural language;
• generating a coherent and human-readable summary of a sequence of clinical notes.

The thesis will be structured as follows:
• state-of-the-art critical analysis in the field of artificial intelligence applied to healthcare;
• problem formulation: objective function, data structures and resources to be used;
• algorithm design and prototyping;
• in-lab testing verification with real data and measurement of the performance of the approach.

The thesis will be co-tutored with the Institute of Biomedical Engineering, University of Oxford. As the opportunity arises, there could be the possibility of doing this thesis abroad depending on the requirements and plans of the master you are enrolled in. The undergraduate will benefit from being immersed in a research environment. It is a unique setting to get into a research mindset with a strong push for innovation. At the end of the thesis, the undergraduate will be familiar with deep learning and semantic analysis, and he will acquire an understanding of the healthcare domain. During the project, he will be able to design and implement an intelligent system applied to real case studies. As additional benefit, she/he will use proficiently control version systems, continuous integration systems, remote deploying and monitoring techniques.

Contact: Send a resume with attached the list of exams to giuseppe.rizzo@linksfoundation.com specifying the thesis code and title.

Automated Scientific Content Generation Using Semantic Analysis and Deep Learning2019-01-16T14:54:03+00:00
Thesis Code: 18002

Thesis Type: Thesis in Computer Science, Data Engineering, Computer Engineering, Mathematical Engineering, Data Science

Research Area: Innovation Development

Requirements
• Experience with Python and/or Java
• Basic knowledge of modular development
• Beginner of (or willing to learn quickly) machine learning
• Curiosity-driven mindset.

Description
Automated assistants are now more than ever taking place in our daily life. Assistants are thus asked to generate content according to user’ inputs and contextual objectives. Let take the case of a scientist in his daily task of performing experiments, filling tables and reporting findings. Lots of his time is spent in transcribing findings that have been already elaborated and encoded in tables. The advancements achieved in artificial intelligence support scenarios of co-operation between an artificial intelligence-based assistant and a scientist when writing technical reports. The objective of this thesis will be thus researching and prototyping an intelligent system able to write science starting from tables. In this thesis the undergraduate will develop an AI-based system for writing scientific papers using both semantic analysis and deep learning. The system will be able to learn autonomously from pairs of tables and papers created as gold examples and generate from a newer table a report.

The thesis will be structured as follows:
• state-of-the-art critical analysis in the field of document generation using both semantic analysis and deep learning;
• problem formulation: objective function, data structures and resources to be used;
• algorithm design and prototyping;
• in-lab testing verification with real data and measurement of the goodness of the approach.

The undergraduate will benefit from being immersed in a research environment. It’s a unique setting to get into a research mindset with a strong push for innovation. At the end of the thesis the undergraduate will be familiar with semantic analysis and deep learning and be able to implement an intelligent system. As additional benefit, she/he will use proficiently control version systems, continuous integration systems, remote deploying and monitoring techniques.

Contact: Send a resume with attached the list of exams to giuseppe.rizzo@linksfoundation.com specifying the thesis code and title.

Deep Learning System to Characterize Scholars using Scientific Papers2019-01-16T14:53:12+00:00
Thesis Code: 18001

Thesis Type: Thesis in Computer Science, Data Engineering, Computer Engineering, Mathematical Engineering, Data Science

Research Area: Innovation Development

Requirements
• Experience with Python and/or Java
• Basic knowledge of modular development
• Beginner of (or willing to learn quickly) machine learning
• Curiosity-driven mindset.

Description
A critical pain of all organizations is managing competencies of their personnel due to both internal variations of topic characterizations (usually an employee acquires knowledge and evolves his professional competence spectrum) and external towards aligning new trends and market requirements. In this thesis the undergraduate will develop an intelligent system for compressing scientific papers into a list of topics in a lossy manner using both semantic analysis and deep learning. The system will be able to learn autonomously from a set of scientific papers authored by scholars and be able to characterize an unknown scholar starting from her/his set of scientific articles.

The thesis will be structured as follows:
• state-of-the-art critical analysis in the field of document summarization using both semantic analysis and deep learning;
• problem formulation: objective function, data structures and resources to be used;
• algorithm design and prototyping;
• in-lab testing verification with real data and measurement of the goodness of the approach.

The undergraduate will benefit from being immersed in a research environment. It’s a unique setting to get into a research mindset with a strong push for innovation. At the end of the thesis the undergraduate will be familiar with deep learning and semantic analysis and be able to implement an intelligent system. As additional benefit, she/he will use proficiently control version systems, continuous integration systems, remote deploying and monitoring techniques.

Contact: Send a resume with attached the list of exams to giuseppe.rizzo@linksfoundation.com specifying the thesis coda and title.

Virtualized Resources Orchestrators for Multi-access Edge Computing in 5G networks2019-01-16T14:52:16+00:00
Thesis Code: 17012

Thesis Type: Master Thesis for Telecommunication Engineering, Computer Engineering or related fields

Research Area: Multi-Layer Wireless Solutions

Requirements
• Excellent software programming skills
• Strong experience with Python/bash scripting and Linux environment
• Strong experience with C/C++ programming languages
• Basic knowledge of Software Define Networking (SDN) and Network Function Virtualization (NFV)
• Basic knowledge of mobile networks

Description
Motivation
The Multi-access Edge Computing (MEC) paradigm has been introduced in 5G mobile networks to achieve an ultra-low latency of less than 1 ms. Today the MEC approach is one of the most promising and studied field in telecommunication. The objective of MEC is to provide computational resources at the edge of the mobile networks, i.e. MEC servers. These servers are closer to the end-users reducing significantly the latency. The foreseen trend for deploying MEC servers is to abstract hardware resources providing a virtualized environment where applications can be managed in an efficient and dynamic approach. The development of Orchestrators in charge of managing the virtual environment is one the most important topics in the framework of the MEC.

Objective
The aim of thesis is to practically implement a MEC framework exploiting one among the Orchestrator tools that are currently under development in the research community. The first step of the thesis will be to survey the available MEC’s Orchestrators tools and to select the one estimated more mature and suited for further development. The second step of the thesis will be to identify the procedures to setup a MEC environment using the select Orchestrator tool. Finally, a specific use case will be implemented potentially improving the Orchestrator with new functionalities.

Contact: send a resume with attached the list of exams to daniele.brevi@linksfoundation.com specifying the thesis code and title.

Correlating audio and visual information for neonatal screening2019-01-16T14:50:14+00:00
Thesis Code: 17008

Thesis Type: Master Thesis for Electronic Engineering/Computer Science or equivalent

Research Area: Multi-Layer Wireless Solutions

Requirements
• Experience with MatLab and possibly OpenCV
• Good knowledge of image processing techniques
• Basic knowledge of audio processing

Description
Nowadays, the analysis of newborns’ level of wellness and reactivity is usually performed by operators which give a subjective evaluation of the baby. A lot of information can also be gathered by exploiting the movements and the sounds emitted by the babies and by correlating them.

The objective of this work is to develop image-processing based algorithms able to automatically analyze the wellness of a baby by tracking his movements, recognizing facial expressions and more. A framework will be proposed to study his/her physiological needs by processing the recorded audio and video.

Contact: send CV to marco.gavelli@linksfoundation.com specifying the thesis code and title.

Workload optimization through heterogeneous and low power accelerators targeting Cloud computing systems2019-01-16T14:49:35+00:00
Thesis Code: 17007

Thesis Type: Master Thesis for Computer Science, Computer Engineering

Research Area: Advanced Computing and Electromagnetics

Requirements
• MS students in Electronic Engineering/Computer Science
• Experience with main programming languages (C/C++, Python), basic knowledge of processor architectures.

Description
Cloud Computing and HPC infrastructures are rapidly evolving to embrace a large set of heterogeneous and low power computing devices. Such enormous variety of devices is necessary to improve energy efficiency of modern datacenters. On the other hand, it represents a challenge from the programming perspective. Moreover, this variety of processing elements make difficult to distribute workloads in an efficient manner.

The objective of the work is to start developing applications and benchmark kernels that take advantages from heterogeneous parallel (low power many-core) architectures (for example: FPGA, GPU…). During the work’s activity, the candidate will be focused on the development of parallel version of heuristic algorithms (e.g., genetic algorithms, PSO, ant colony, etc.) targeting the workload optimization of servers. The candidate will use the Parallella platform, a modern development board equipped with a many-core accelerator.

Contact: send CV to olivier.terzo@linksfoundation.com specifying the thesis code and title.

Implementation and performance of a standard LDM engine for future automotive applications2019-01-16T14:48:52+00:00
Thesis Code: 17004

Thesis Type: Thesis in Telecommunication/Electronic Engineering, Computer Science, Automotive Engineering

Research Area: Multi-Layer Wireless Solutions

Requirements:
• brief knowledge of 80211 wireless standards
• Basic knowledge of Linux Operating system and embedded platforms
• Advanced knowledge of C language
• Advanced software development skills
• Knowledge about database architectures is a plus

Description:
Motivation:
In the context of intelligent transport systems (ITS), a wide set of standards are being developed with the aim to allow ad-hoc communication between vehicles and allow vehicular applications to access vehicle information in real time. Currently, the standard common aggregation point of vehicle information, is the Local Dynamic Map (LDM). LDM is a database that contains a real time picture of the ITS environment surrounding the vehicle. From the application point of view, a set of applications (designed at ISMB) intended to provide information from vehicle to LDM and exploit information available on the LDM, are under continuous development in order to build a standard compliant prototype of ITS vehicle’s on board unit. Current ongoing developments of such LDM require specific design decisions in order to fulfill performance targets for future vehicles and being also compliant with the ITS standard.

Objective: The main goal of this thesis work is to explore different architectural possibilities and then implement a practical candidate LDM solution, which will work in our currently available embedded platforms; the student will also analyze LDM performance in terms of computational weight, response times, etc of the solutions proposed. The implementation will be standard-compliant and will be based on the current developments carried out by ISMB. The LDM will be tested using currently prototype automotive applications, developed by ISMB, on a real (or at least nearest to real) vehicle scenario. The student will have the opportunity to be in contact with European projects environments in the vehicular sector.

Contact: interested candidates send a resume with attached the list of exams/grades taken during the Bachelor of Science to daniele.brevi@linksfoundation.com specifying the thesis title.

Machine Learning Techniques for Industrial Applications2019-01-16T14:47:06+00:00
Thesis Code: 16021

Thesis Type: 6 months Master Thesis (Laurea Magistrale) for students of: Computer Engineering, Communications and ICT Engineering, Mathematical Engineering or equivalent.

Research Area: Pervasive Technologies

Requirements:
– Skills in algorithms development and programming
– Interest for industrial applications.

Description:
Motivation:
The proposed work originates from two facts:
• Predictive control in industry could produce significant economical and environmental advantages. However, the industrial processes are generally very complex to model and control.
• Machine Learning has been proved to deal with complex systems and large datasets very efficiently. It is able to model input-output systems with very few model constraints.
It is then natural to think to the implementation of Machine Learning techniques to control and optimize industrial processes. Even though very popular for a variety of technological services, Machine Learning has also a great potential in the industrial domain, which has not been fully investigated yet. The interest on this topic is growing rapidly both from the industrial and scientific sectors.

Objective:
The purpose of this thesis is to study and develop Machine Learning algorithms and implement them to tackle industrial problems. A theoretical in-depth analysis of Machine Learning techniques will be the starting point of the work. Afterwards, the implementation of suitable algorithms will be tested on industrial datasets, with the purpose of exploiting time series data to predict failures and/or classify low-quality products.

Contacts: send a resume specifying the thesis code and title to claudio.pastrone@linksfoundation.com

Open Data Platforms for the Industrial Internet of Things (IoT)2019-01-16T14:46:00+00:00
Thesis Code: 16019

Thesis Type: 5-6 months Master Thesis (Laurea Magistrale) for Computer Science, Computer Engineering students or related.

Research Area: Pervasive Technologies

Requirements:
General Knowledge of Database Technologies, Programming in Distributed Environments. Good design and programming skills in Java are required. Previous interest/experience in developing IoT applications and/or working with Big Data systems or Machine Learning libraries e.g. TensorFlow is a plus.

Description:
Motivation:
The Internet-of-Things (IoT) is a vision in which every physical object – enriched with communication capabilities – acquires an electronic identity and acts as a source of information. The rapid uptake of IoT technologies in different application domains is causing a tremendous increase in the amount of data being collected, stored and processed, above all in Industrial Manufacturing and Process Industries domain. In order to handle the increasing amount of data in sustainable fashion, many industrial players are deploying scalable data platforms based on open technologies, both using in-premises or cloud-oriented systems.

Objective:
The student(s) will be involved in analyzing, designing and deploying a reference platform for real-time processing of IoT data built upon open-source technologies and/or cloud-based platforms, responding to requirements derived from real industrial scenario. Selected student(s) will be involved in the development of a proof-of-concept Machine Learning application working over large, real-time data flows. The proposed Machine Learning application will be deployed and evaluated using the proposed platform.

Note: this thesis project can also be potentially developed by a team of 2 or 3 students, each focusing on separate aspects of the problem (Algorithms, Data Platforms, etc.). Students will be jointly tutored by researchers of the ACE and PerT Areas.

Contacts: send a resume specifying the thesis code and title to claudio.pastrone@linksfoundation.com

Development of a secure communication framework for smart factory monitoring2019-01-16T14:44:44+00:00
Thesis Code: 16018

Thesis Type: 6 months Master Thesis (Laurea Magistrale) for Computer Science or Computer Engineering students.

Research Area: Pervasive Technologies

Requirements:
• Good design and programming skills in Java. Some knowledge of OSGi is a preference.

Description:
Motivation:
Internet-of-Things (IoT) technologies are expected to become a key instrument to extract information from the physical world and integrate it with existing IT infrastructures. Particularly, in the smart industry scenario, also called Industry 4.0, the Internet-of-Things paradigm is going to introduce several innovations: cyber-physical systems monitor physical processes and the physical devices are “abstracted” in virtual entity, which can be more easily interconnected, despite their differences in terms of hardware and software. Indeed, one of the main innovation of the Internet of Things in the smart-factory scenario, is the possibility to enable real time communication machine-to-machine and machine-to-humans in a standard, secure, scalable and reliable way.

Assets:
• VIRTUS is an IoT-oriented middleware created and maintained by ISMB which provides a scalable, agile, event-driven, network independent tool to manage a large-scale network of heterogeneous cooperating objects leveraging on open and standard instruments
• Clayster Exchange is a solution based on open standards, which allows applying provisioning, delegation of trust and device discovery in Internet of Things devices and applications.

Objective:
The purpose of this thesis is to design and develop a framework able to perform continuous monitoring of plant-wide resources, based on VIRTUS and the Clayster provisioning server. The secure and reliable communication infrastructure will be ensured between all levels: from shopfloor to management level; in such a way that, for example, the communication-related malfunctions are properly detected and reported to control systems so that “failure awareness”, “graceful degradation” or “selfhealing” approaches can be employed. ISMB will be the external technical advisors, with support from Clayster (http://www.clayster.com/).

Contacts: send a resume with attached the list of exams taken during the Bachelor of Science to claudio.pastrone@linksfoundation.com specifying the thesis code and title.

Development of an adaptation layer for monitoring of smart factory devices2019-01-16T14:43:36+00:00
Thesis Code: 16017

Thesis Type: 6 months Master Thesis (Laurea Magistrale) for Computer Science or Computer Engineering students or equivalent.

Research Area: Pervasive Technologies

Requirements:
• Good design and programming skills in Java. Some knowledge of OSGi is a preference.

Descriptions:
Motivation:
Internet-of-Things (IoT) technologies are expected to become a key instrument to extract information from the physical world and integrate it with existing IT infrastructures. Specifically, in the Industry 4.0 scenario, the IoT paradigm is going to be used to abstract physical devices in Virtual Devices, in order to enable them to communicate to each other and with the human operators in standard ways, despite their differences in terms of hardware and software.

Assets:
VIRTUS is an IoT-oriented middleware created and maintained by ISMB which provides a scalable, agile, event-driven, network independent tool to manage a large-scale network of heterogeneous cooperating objects leveraging on open and standard instruments.

Objective:
The purpose of this thesis is to design and develop an abstraction layer, based on VIRTUS, to allow monitoring and control of devices in a smart-industry scenario. The focus of the thesis will be on the development of a driver based on the OPC-UA standard (https://opcfoundation.org/about/opc-technologies/opc-ua/), to interact with PLCs, in a standard way. Furthermore, the thesis will design and develop a set of tools to ease the deployment and tuning of drivers, like the OPC one, used to communicate with physical factory devices.

Contacts: send a resume with attached the list of exams taken during the Bachelor of Science to claudio.pastrone@linksfoundation.com specifying the thesis code and title.

Traffic data as an enabler of urban mobility analysis2019-01-16T14:34:26+00:00
Thesis Code: 16011

Thesis Type: Master of Science in Computer Engineering (ICT) or Telecommunication Engineering

Research Area: Smart City

Requirements:
• Interest in urban mobility
• Good programming skills
• Experience in data analysis
• Experience with maps (OpenStreetMaps, Google Maps, etc.) is a plus.

Description:
Traffic data allows scientists to have a direct observation of urban mobility. For the sake of example, traffic data enables the analysis of the impact of connected vehicles on the communication network. This work wants to gather and to process traffic data (GPS traces) with the final objective of designing a new tool that enable a more realistic analysis of urban mobility than the state-of-the-art traffic simulation.

Contacts: Send your CV to andrea.vesco@linksfoundation.com specifying the thesis code and title.

Deep learning: vision and perspectives2019-01-16T14:32:51+00:00
Thesis Code: 16009
Research Area: Multi-Layer Wireless Solutions

Requirements:
• MS students in Electronic Engineering/Computer Science or equivalent
• Good attitude towards the understanding of emerging trends
• Basic knowledge of network protocols and tools for data analysis

Description:
Deep learning is an emerging branch of machine learning based on the study of algorithms that attempt to model behaviors (e.g. actions of a user) to predict, analyze and propose optimal strategies to solve problems in an automated way e.g. by adopting neural networks. The objective of this work is to carefully study the literature and the state-of-the-art on this topic together with the tools and framework actually proposed in this field (e.g. Google TensorFlow, Caffe , Torch…)

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

Image processing for the automotive industry2019-01-16T14:32:05+00:00
Thesis Code: 16008
Research Area: Multi-Layer Wireless Solutions

Requirements:
● MS students in Electronic Engineering/Computer Science or equivalent
● Experience with main programming languages (C/C++)
● Good knowledge of image processing techniques, network protocols

Description:
The objective of the thesis will be to study and implement innovative and effective tools to help car drivers. The state-of-the-art in this area already includes applications such as road signs and lane detectors. New and promising applications will be studied and tested in real scenarios (finding free parking slots, interacting with neighbors…).

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

Emerging standards in multimedia communication2019-01-16T14:31:05+00:00
Thesis Code: 16007
Research Area: Multi-Layer Wireless Solutions

Requirements:
● MS students in Electronic Engineering/Computer Science or equivalent
● Experience with main programming languages (C/C++)
● Good knowledge of image processing techniques, network protocols

Description:
Acronyms like H.265, VP9, MMT, DASH, WebRTC, UPNP 2.0 are becoming familiar in these years as new and promising standards for video encoding and multimedia transmission. The objective of the work will be defined more in details when the thesis will be carried out according to the most attractive technologies at the moment and the student’s interests. As an example, tests using the software verification model of the new H.265/HEVC video codec in different scenarios can be performed and also its extensions for future applications can be analyzed.

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

On-board live H.264/AVC HW video encoding2019-01-16T14:29:29+00:00
Thesis Code: 16006
Research Area: Multi-Layer Wireless Solutions

Requirements:
• MS students in Electronic Engineering/Computer Science or equivalent
• Experience with main programming languages (C/C++)
• Basic knowledge of video processing and transmission techniques

Description:
Unmanned Aerial Systems equipped with video cameras usually require a significant bandwidth to transmit in real-time data to the ground for live analysis and mission control. State-of-the-art mobile infrastructures (3G, 4G…) nowadays offer a very time-variant channel to transmit such information so the adaptation of the video stream is a desired feature. The objective of this thesis is to study a framework able to perform an on-board live H.264/AVC video encoding by means of dedicated HW boards. Different configurations and solutions will be analyzed to gather information about performance and complexity of such a scheme. Furthermore, on-demand adaptive streaming will be studied and implemented in order to adapt the data rate to the actual channel state.

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

Concealment algorithms for aerial video analysis for surveillance and prevention2019-01-16T14:28:45+00:00
Thesis Code: 16005
Research Area: Multi-Layer Wireless Solutions

Requirements:
● MS students in Electronic Engineering/Computer Science or equivalent
● Experience with main programming languages (C/C++)
● Good knowledge of image processing techniques

Description:
Unmanned Aerial Systems equipped with HD cameras and a plethora of sensors are becoming popular in recent years. A wide range of applications is then expected to gain attention by exploiting image processing algorithms. The objective of this work is to discuss and implement innovative features for video concealment in case any losses occur during data transmission to the base station.

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

MPEG CDVS for biometrical analysis2019-01-16T14:26:31+00:00
Thesis Code: 16004
Research Area: Multi-Layer Wireless Solutions

Requirements:
● MS students in Electronic Engineering/Computer Science or equivalent
● Experience with main programming languages (C/C++)
● Good knowledge of image processing techniques

Description:
The new MPEG Compact Descriptor for Visual Search (CDVS) standard allows to significantly compress the key-points of an image in order to allow the processor to deal with much less information thus efficiently perform data analysis and matching. Deriving the key-point for an image allows comparison, storage and retrieval in a faster way in image-based applications where speed and efficiency are required features. The objective of this work is to explore how this new standard can improve biometric analysis. The student will study the framework of MPEG CDVS to understand how the keypoints for biometric images are generated. As a possible application, images of fingerprints and/or retina will be analyzed and an architecture encompassing fingerprint matching will be developed.

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title

Study of body subtle motion for biomedical applications2019-01-16T14:03:01+00:00
Thesis Code: 16002
Research Area: Multi-Layer Wireless Solutions

Requirements:
• MS students in Electronic Engineering/Computer Science or equivalent
• Experience with MatLab
• Good knowledge of image processing techniques

Description:
Image processing can dramatically help the recognition of a user’s vital sign (heart rate, respiratory rate…) by processing the so-called micro-movement of the body. This subtle motion can be study by acquiring video data thus deriving significant information for clinical assessment without the need of a contact with the user. The objective of this work is to study the already implemented algorithms in this emerging field, and modify them in order to fit them to heterogeneous scenarios and test the performance.

Contacts: Send CV to daniele.brevi@linksfoundation.com specifying the thesis code and title