Description:
ns3 is a simulation platform of networks. Millicar is a module for ns3 that allows simulating scenarios of vehicular communication. O-RAN is the new architecture of cellular networks. It is based on concepts of softwarization, disaggregation and flexibility to allow a level of control of all network aspects without precedent. The purpose of this thesis is to integrate millicar within an architecture O-RAN-based network control architecture. The final product will be the first existing simulation platform that integrates communiations vehicular and O-RAN.

Hide Description

Description:
Pierburg Pump Technology Italy Spa is focused on developing mechatronic products for automotive industry. The Torino Engineering Center develops electronic solutions serving automotive applications including coolant, oil, hydrogen electrical pumps, proportional coolant valves, and advanced development for e-mobility – in all ranges of voltage and power. The focus of the thesis will be on the motors used in the automotive industry for smart-actuators, e-water pumps, and e-oil pumps. The target of the thesis is to perform an analysis of the state of art in estimation techniques helpful in learning the electric parameters of E-motors in real-time. The analysis must lead to select the approaches most suitable for applications in the automotive industry, considering the limitations in computational complexity and sampling times. The comparison shall be conducted on a test bench, site in Turin, where it is possible to set the environmental conditions of different working points of the motor in a climate chamber. The mentioned approaches will be implemented on a rapid-prototyping platform and must be suitable for 3-phases brushless motor, dc-motor and stepper-motor with permanent magnet excitation. To complete the thesis the student will acquire abilities in rapid-prototyping embedded system, in NI instrumentations and testing tools used in the automotive R&D field. Required skills: • Capability to design data-driven models • Capability to formulate physical models of electromechanical systems • Capability to implement and validate numerical simulations of dynamic systems • Familiarity with Matlab-Simulink • Experience with system identification and dynamic modeling

Hide Description

Description:
Pierburg Pump Technology Italy Spa is focused on developing mechatronic products for the automotive industry. The Torino Engineering Center develops electronic solutions serving automotive applications including coolant, oil, hydrogen electrical pumps, proportional coolant valves, and advanced development for e-mobility – in all ranges of voltage and power. The thesis will focus on the motors used in the automotive industry for smart-actuators, e-water pumps, and e-oil pumps. Starting from the physical characteristics of an electric motor, the target of the thesis is to identify the best configuration for a virtual thermal sensor, able to observe the coils temperature. The virtual sensor shall have as inputs the currents measured by the shunt resistances and the voltages applied to the H-Bridge, available in the HW. The mentioned virtual sensor will be developed on a rapid-prototyping platform and should be suitable for 3-phase brushless, DC and stepper-motors, with permanent magnet excitation. The virtual sensor performance will be evaluate on a test bench, site in Turin, where it is possible to set the environmental conditions, through using a climate chamber, for each working point required for the system where the motor is involved. To complete the thesis the student will acquire abilities in rapid-prototyping embedded system, in NI instrumentations and testing tools used in the automotive R&D field. REquired skills: • Capability to design data driven models • Capability to formulate physical models of electromechanical systems • Capability to implement and validate numerical simulations of dynamic systems • Familiarity with Matlab-Simulink • Experience with system identification and dynamic modeling

Hide Description

Description:
There has been an increased interest in the utilization of unmanned aerial vehicles (UAVs) for different civilian, military, and commercial applications. In particular, the communications sector has been testing the usage of UAVs to either augment existing infrastructure or replace them in particular scenarios. This is achieved by mounting base stations on the UAVs, which are then flown to the desired location to provide coverage to users. With their advantages such as ease of deployment, lower costs than base station towers, and flexible location capabilities, these so-called aerial base stations (ABSs) are a promising technology for emergency response, e.g., in cases when the terrestrial communication infrastructure is inoperable due to extreme phenomena like typhoons or earthquakes. However, with these advantages to ABSs, there lies an open problem with regard to their optimal placement to maximize their coverage in terms of the user count and quality of service (QoS). There have been several proposed approaches to deal with the UAV placement problem. These approaches can be based on clustering, coverage circles optimization, or virtual forces on the UAV agents. However, most of these solutions assume known ground client (GC) locations, which is not realistic, especially in emergency situations when ground towers are unpowered and we do not have information on user locations to save bandwidth, and also due to privacy issues. However, we can have information on the sensed QoS of the connected users–whether individual QoS or the aggregate–which can be associated with the user coverage of the ABS. Due to the a priori unknown user locations, there is an intrinsic trade-off between the exploitation of existing data from previously-sensed locations, and exploration to different locations around the target region to acquire more information on user QoS, and in turn, coverage fitness. In this master thesis, you will investigate a black-box optimization approach to the ABS coverage problem, in the context of an ABS formation. Using the respective ABS locations and their sensed user QoS, the formation automatically decides on the best placement combinations and trajectories of the member ABSs, according to the trade-off between exploitation and exploration. Main tasks The student will perform the following: • Literature review on UAV usage for wireless network coverage applications • Study of a black-box optimization technique to be used for the problem • Formulation of a scheme for ABS formation placement for coverage, using real-time optimization • Development of a simulation testbed • Testing with time-invariant (but a priori unknown) client locations • Documentation of performed work Required skills: Highly skilled (or willing to learn) in programming in any language, but preferably Python and/or MATLAB/Simulink, Analytical thinking and applied mathematical skills, Willing to be trained on new concepts and skills

Hide Description

Description:
In the automotive industry, the most common battery technology uses lithium, whose autonomy is currently in the order of a few hundred kilometers. The reliability and performance of these batteries are influenced by the management of the charging and discharging phases. Continuous and accurate monitoring of the battery state takes on significant importance to extend the battery lifetime, effectively plan the trip route and charging stops, and optimize the energy flow management of HEVs. The main parameters to be assessed for correct battery monitoring are the state of charge (SOC), and the state of health (SOH). These two states cannot be directly measured since the technology is unavailable. Many approaches are proposed in the literature to estimate these parameters. The objective of this thesis is to perform a deep literature analysis of the state of the art on the main industrial approaches and to develop a hybrid physical-data-driven approach design to estimate the SOC and SOH of batteries in automatic applications. Required skills: • Capability to design data-driven models • Capability to formulate physical modeling approach • Capability to implement and validate numerical simulations of dynamic systems • Solid background in estimation Desirable: • Familiarity with Matlab-Simulink • Experience with system identification and dynamic modeling • Basic knowledge of process control and modeling of chemical processes

Hide Description

Description:
The aim of the thesis is to develop a digital platform, based on wireless sensors, for the analysis of movement in manufacturing environments. The system will integrate surface electromyographic devices and inertial measurement units, wirelessly connected to a central unit. The platform must allow estimating the musculoskeletal workload during specific assembling tasks in a typical manufacturing setup, synchronizing the readings from the different sets of sensors, and processing the row data in “real-time” to provide accurate estimates of the biomechanical effort. The system will be used to compare the behavior and workload of operators in assembling tasks when using a passive exoskeleton with respect to a baseline situation without any mechanical support for the activity. Required skills: • Solid background in signal processing • Capability to manipulate and analyze biomedical signals related to musculoskeletal workload • Capability to analyze and correlate inertial measurements with human biomechanical models • Capability to implement software solutions for signal acquisition and processing • Familiarity with APIs for wireless sensors

Hide Description

Description:
The work is in cooperation with Hitachi Energy.

Hide Description

Description:
The emission of GreenHouse Gases (GHG) and other particulate in the atmosphere is an ever growing concern of all government bodies around the world. Mapping with more detail sources of GHG emission and doing surveys to collect experimental data is key to understanding the problem in more depth. The key objective of this thesis is developing a low weight GHG measurement platform based on Raspberry Pi or Arduino, to be mounted on a small UAV for data collection campaigns. The work will span from sensor selection by surveying datasheets of commercially available solutions, to hardware implementation and writing the software for performing the experimental data acquisitions. Several challenges will have to be tackled from constraints due to weight limitations to synchronization of the captured data with UAV flight path.

Hide Description

Description:
Context There has been an increased interest in the utilization of unmanned aerial vehicles (UAVs) for different civilian, military, and commercial applications. In particular, the communications sector has been testing the usage of UAVs to either augment existing infrastructure, or to replace them in particular scenarios. This is achieved by mounting base stations on the UAVs, which are then flown to the desired location to provide coverage to users. With their advantages such as ease of deployment, lower costs than base station towers, and flexible location capabilities, these so-called aerial base stations (ABSs) are a promising technology for emergency response, e.g., in cases when the terrestrial communication infrastructure is inoperable due to extreme phenomena like typhoons or earthquakes. However, with these advantages to ABSs, there lies an open problem with regards to their optimal placement as to maximize their coverage in terms of the users count and quality of service (QoS). There have been several proposed approaches to deal with the UAV placement problem. These approaches can be based on clustering, coverage circles optimization, or the use of virtual forces on the UAV agents. However, most of these solutions assume known ground client (GC) locations, which is not realistic especially in emergency situations when ground towers are unpowered and we do not have information on user locations to save bandwidth, and due to privacy issues. However, we can have information on the sensed QoS of the connected users–whether individual QoS or the aggregate–which can be associated with the user coverage of the ABS. Due to the a priori unknown user locations, there is an intrinsic trade-off between exploitation of existing data from previously-sensed locations, and exploration to different locations around the target region to acquire more information on user QoS, and in turn, coverage fitness. In this master thesis, you will be investigating a black-box optimization approach to the ABS coverage problem, in the context of an ABS formation. Using the respective ABS locations and their sensed user QoS, the formation automatically decides on the best placement combinations and trajectories of the member ABSs, according to the trade-off between exploitation and exploration. Tasks The master student will perform the following - Literature review on UAV usage for wireless network coverage applications - Study of a black-box optimization technique to be used for the problem - Formulation of a scheme for ABS formation placement for coverage, using real-time optimization - Development of a simulation testbed - Testing with time-invariant (but a priori unknown) client locations - Documentation of performed work Expected profile Currently-enrolled master student in Computer Engineering or Automation Engineering Highly skilled (or willing to learn) in programming in any language, but preferably Python and/or MATLAB/Simulink Analytical thinking and applied mathematical skills Willing to be trained on new concepts and skills Initial references Viet, P. Q., & Romero, D. (2022). Aerial Base Station Placement: A Tutorial Introduction. IEEE Communications Magazine, 60(5), 44–49. https://doi.org/10.1109/MCOM.001.2100861 Cao, X., Yang, P., Alzenad, M., Xi, X., Wu, D., & Yanikomeroglu, H. (2018). Airborne communication networks: A survey. IEEE Journal on Selected Areas in Communications, 36(9), 1907–1926. https://doi.org/10.1109/JSAC.2018.2864423 Mozaffari, M., Saad, W., Bennis, M., Nam, Y. H., & Debbah, M. (2019). A Tutorial on UAVs for Wireless Networks: Applications, Challenges, and Open Problems. IEEE Communications Surveys and Tutorials, 21(3), 2334–2360. https://doi.org/10.1109/COMST.2019.2902862 Alzenad, M., El-Keyi, A., & Yanikomeroglu, H. (2018). 3-D Placement of an Unmanned Aerial Vehicle Base Station for Maximum Coverage of Users with Different QoS Requirements. IEEE Wireless Communications Letters, 7(1), 38–41. https://doi.org/10.1109/LWC.2017.2752161 Hayajneh, K. F., Bani-Hani, K., Shakhatreh, H., Anan, M., & Sawalmeh, A. (2021). 3D deployment of Unmanned aerial vehicle-base station assisting ground-base station. Wireless Communications and Mobile Computing, 2021. https://doi.org/10.1155/2021/2937224 Contact Prospective students are to send their CV, master coursework grades, and code portfolio (if present) to: prof. Fredy Ruiz, Ph.D. Associate Professor fredy.ruiz@polimi.it prof. Lorenzo Fagiano, Ph.D. Associate Professor lorenzo.fagiano@polimi.it

Hide Description

Description:
The thesis is In collaboration with ST-Microelectronics, and provides for reimbursement of expenses. The thesis will start practicing on some specific technologies (MEMS mirrors, lasers, SPADs and Silicon-photonics), but also optical and embedded systems, then he/she will join the MEMS LiDAR or Silicon-Photonics program Role may vary upon specific project needs. Task examples that can be assigned to the intern are: Design and develop a control system for a LIDAR projection system realized in Silicon Photonics; build a test environment; characterize the device in real conditions. Design and develop a control system for a gyroscope realized in Silicon Photonics; build a test environment; characterize the device in real conditions. Design and develop a new CFD (Constant Fraction Discriminator); build a test environment; characterize the device in real conditions. Technical skills: Willing to learn and autonomy HW-SW skills (mandatory): Digital and analog electronics, schematic diagrams, C, C++, Matlab Other skills (appreciated): VHDL, 3D modeling (Solidworks), optics ST contact: daniele.caltabiano@st.com

Hide Description

Description:
A huge amount of information is generated every day by a countless number of sensors. These sensor data are used to take important decisions and improve the welfare of users and citizens. However, the most of these data is never used and its storage and maintenance have a relevant cost in terms of finance and energy. This thesis aims at explore a different approach for sensor data management and will investigate methods and tools to adapt the data collection from sensor in terms of data generation frequency and level of detail according to several factor: actual usage of the data, data location, and energy efficiency of data storage and movement. This thesis will provide a methodology to solve issues related to where, when, and how to store sensor data in the cloud continuum. KEYWORDS: Green IT, IoT, adaptive monitoring system, cloud continuum, fog computing

Hide Description

Description:
Modern applications are composed of several components (microservices, functions, monoliths) that cooperate together for reaching the overall application goal. Most of the existing infrastructures don’t provide information about the energy consumption at the application component level. Thus, some models to derive this information are needed if we want to envision a real sustainable application future. The aim of this thesis is to explore the current state of the art and the existing tools for deriving energy consumption of single-component applications and their emissions and to propose a method to extend these tools for measuring the impact of multi-component applications. KEYWORDS: Green IT, CO2 estimation, energy consumption, cloud native applications

Hide Description

Description:
The adoption of AI is sensitively increasing the number of computing instances in the cloud. The training phase uses a huge amount of data coming from different data sources. The training data volume is one of the most relevant contributors to the power consumption of model training in ML, while having a limited impact on the model accuracy. At the same time, distributed ML approaches are becoming more and more popular (e.g., federated learning). In this context, the thesis should provide an approach to decrease the energy consumption of AI without affecting the accuracy of the model by taking decisions on where and how to store the training data and to execute the model training. KEYWORDS: data preparation, Green AI, Data quality, distributed learning

Hide Description

Description:
Sustainability and Energy Efficiency are a driver for the improvement of our society and a necessity for a better future. While Data Centers are getting greener, this is not the case for other computing locations like fog and edge computing, composed of small and inefficient data centers. Considering applications as composed of several microservices interconnected in a workflow and exchanging data, the research question is how it is possible to adaptively allocate each component by considering the trade-off between QoS and Energy Efficiency. KEYWORDS: micro services, fog computing, cloud continuum, green IT, scheduler

Hide Description

Description:
Images, audio tracks and videos can be manipulated in many different ways (e.g., object insertion and removal, local retouching, laundering operations, copy-move, deepfakes, etc.). We are interested in developing methods to detect and localize possible editing operations on multimedia objects exploiting model-based and data-driven solutions.

Hide Description

Description:
Cooperation with Prysmian S.p.A.

Hide Description

Description:
Copperation with Terna S.p.A.

Hide Description

Description:
Load forecasting plays a crucial role in the day-to-day operations of electric utilities, especially in modern power systems, where a significant share of power generation is attributable to renewable sources. Over the years, several algorithms have been developed to tackle this problem, on time scales ranging from a few hours to several months. Most recent solutions have employed machine learning techniques such as deep learning to increase the granularity of the prediction, down to the single-building level. The main goal of this thesis is the implementation of several machine learning algorithms for the prediction of load consumption of a single building over a 24-hour time horizon. The data set that will be used for training is very rich and consists of over a year of recordings covering load consumption, load generation (due to photovoltaic panels installed on the building), and weather data. The applicant will implement and test the performance of both conventional machine learning algorithms such as Autoregressive Integrated Moving Average (ARIMA) and deep-learning-based approaches, for instance using Long Short-Term Memory (LSTM) networks.

Hide Description

Description:
The displacement of conventional generators by renewable generation technologies implies a reduction in rotating inertia, which may jeopardize system stability. Monitoring inertia in real-time is thus becoming a major concern for Transmission System Operators (TSOs). The main goal of this thesis is the development of a framework for the online estimation of inertia in an electric power system by using state-of-the-art artificial intelligence (AI) techniques. Unlike existing algorithms for inertia estimation, which rely on disturbances (such as load or generator trips or line faults), the aim of this thesis is to provide a continuous estimation of inertia during the normal, unperturbed operation of a power system. To achieve this goal, the applicants will develop an approach based on deep convolutional neural networks for the continuous estimation of system inertia and apply it to synthetic data obtained by simulating models of several power networks. The obtained results will be compared with existing methods of inertia estimation and with other conventional (i.e., not based on deep-learning) machine learning algorithms.

Hide Description

Description:
Several black-box models of circuit breakers have been proposed since the pioneering work by Cassie in 1943. Nevertheless, the nowadays need for a ``further'' black-box model is due to two main reasons: (i) accurate tools to simulate the complex physics that describes the interruption phenomena in circuit breakers are very CPU intensive; (ii) available black-box models condense in a few simple equations the complex physics of circuit breakers, but they are inaccurate, and do not cope with the full set of working conditions. The focus of this research thesis is on medium voltage vacuum breakers. The first goal is to formulate a proper electric arc model grounding on a modeling approach based on the prey-predator paradigm originally used to describe dust-forming plasmas that are found in flames, the interstellar medium, and comet tails, and then adopted to model SF6 circuit breakers and air miniature circuit breakers. Then, the arc model will be linked to the geometric, kinematic, and magnetic field properties of the vacuum breaker structure to obtain an aggregate model at the two electrical terminals.

Hide Description

Description:
Constant ON-Time (COT) buck converters are popular in wide input-voltage range applications for many reasons, e.g., the ease of control design over varying input voltage conditions, the absence of jittery ON-time behavior, which affects constant-frequency control methods when incurring in minimum controllable ON-time limitations, and, provided that the ON-time is adapted to the input voltage, the ability to maintain a reasonably constant switching frequency over the input voltage range. Many wide-input-range adaptive COT regulators, targeted to automotive and other wide-input-range applications, are available from various IC vendors. COT buck converters are switching circuits and may exhibit complex dynamics induced by several types of bifurcations: period-doubling, grazing, contact, and border-collision. This thesis aims to study the stability of this circuit by resorting to proper mathematical tools, such as one-dimensional discrete-time maps, to identify suitable regions in the circuit parameter space in which it could be profitably used.

Hide Description