Search

This contribution illustrates the preliminary results of the research activities carried out within the PhD in Information and Communication Technology for Health (ICTH) on the topic of Delivery Manager of Cognitive Computing for Neuroncology, at the Department of Electrical Engineering and Information Technologies (DIETI) of the University of Naples “Federico II”. Main goal of this research activity is the development of computational models based on Visual Recognition and Cognitive Computing algorithms for the analysis of Magnetic Resonance Imaging (MRI) derived data in patients with brain neoplasm of unknown origin, in order to obtain a semi-supervised classification model which allows the timely identification of various brain tumors types, requiring therapeutic strategies and clinical-radiological monitoring different from each other. The final purpose of this research activity is to propose a cognitive computing-based support able to supplement information that can help the neuroradiologist in decision-making for daily clinical practice.

This manuscript introduces a compact electrothermal model for SiC power MOSFETs that can be easily scaled to devices of different voltage ratings. The model is implemented as a subcircuit containing mainly SPICE native components. Both the static and dynamic performance can be tuned by adjusting a small set of parameters. The model is validated on 1.7 kV-60 A-rated devices.

Normally-off p-GaN HEMTs with Schottky-type gate are becoming of common adoption in both industrial and consumer electronic market. The presence of a gate current with a remarkable dependence on the temperature of the device allows embedding a temperature sensing feature with very fast response into the gate-driving circuitry. In this work, a prototype of a gate driver circuit with temperature monitoring capability is presented and its applicability in a broad range of operating conditions is validated by means of a thorough experimental campaign.

This paper presents the results achieved in four years of activity within the NERD? project, which was created by IBM Italy and La Sapienza University of Rome and is carried on in collaboration with several universities and schools, in Italy. The long-term objective of the project is bridging the gender gap in the so-called STEM disciplines (science, technology, engineering and mathematics) starting from the secondary education. Specifically, we refer to the experience conducted in the field of informatics and computer engineering by the Federico II University of Naples, which empowered the original educational model of the NERD? project, introducing collaboration issues within a smart learning community. This involved people from industry, university and secondary schools, including both teachers and students at all the levels.

In this paper, a coherent multiple-input multiple-output (MIMO) radar system with widely separated antennas is simulated in three-dimensional (3D) surveillance scenarios. This analysis proceeds in parallel with the design and development of a system demonstrator based on photonic technology. Microwave photonics permits to conceive an architecture where all radar signals are coherently generated and received in a photonics-based central unit and optically distributed/collected to/from the radar heads where antennas are located, through optical fiber links, maintaining the coherence. Albeit centralized coherent data fusion becomes possible, preliminary results show how secondary lobes represent the major limiting factor for the radar's performance. Nonetheless, diversity in the spatial and/or frequency domains become a crucial resource to circumvent this problem. A MIMO radar simulator is proposed to estimate the coherent ambiguity function in 3D multi-target multi-band scenarios. Preliminary simulation results, concerning 3D scenarios considered of high interest for automotive and other applications, show that coherence in MIMO systems allows to achieve an angular and height resolution very close to the theoretical limit given by the used RF wavelength and baseline length, thus confirming the huge potential of coherent MIMO radars in 3D applications.

In this study, an electrothermal compact model for SiC power MOSFETs is experimentally verified for the first time on a 3.3 kV device. Both the parameters determining the static behavior and those controlling the dynamic performances are calibrated over experimental data. The good agreement achieved proves the model scalability form 1.2 kV devices to 3.3 kV ones. Successively, to validate the good convergence properties of the model, an electrothermal simulation of a three-phase inverter consisting of 36 hard switched MOSFETs is performed.

In this paper, a temperature-dependent SPICE model of SiC power MOSFET is used for statistical electrothermal (ET) simulations of parallel devices for multichip power module applications. The proposed approach is applied to a power module in a half-bridge configuration, where parallel MOSFETs are subject to mismatched parameters. The impact of threshold voltage and current factor statistical fluctuations on energy dissipation during hard switching conditions is evaluated by means of Monte Carlo ET simulations. Thereafter, the mean time-to-failure (MTTF) has been estimated by using the lifetime models from literature for SiC chip. Finally, the effect of current imbalance on the acceleration factor due to changes in devices self-heating is analyzed and discussed.

In this paper the first coherent dual-band 2x4 MIMO radar experiment is presented. Range/cross-range maps demonstrate the higher cross-range resolution due to the coherence and the enhanced performance introduced by dual-band operation

Photonics-enabled coherent MIMO radars have been numerically investigated to evaluate the benefits of coherence and multi-band operation in 2/3D imaging.

In this paper we present an effective simulation strategy to analyze the behavior of power semiconductor devices during high electro-thermal stress events. This technique can enable device designers and circuit engineers to foresee the device behavior during stressful events in order to assess eventual reliability issues or lifespan reduction.

This paper introduces a statistical analysis of the impact of devices parameters dispersion on the performances of parallel connected SiC MOSFETs. To this purpose, the statistical fluctuations of threshold voltage and current factor are evaluated on a set of 20 MOSFETs. In order to assess the effects of parameters spread in a real operating condition, the electrothermal simulation of a 200kHz synchronous buck converter is performed. Subsequently, an investigation of the switching energy unbalance, as a function of parameters distribution tolerances, is achieved through several sets of Monte Carlo electrothermal simulations. The results aim at aiding both the design of multi-chip configurations and the selection of appropriate fabrication process rejection boundaries.

This work discusses about the short-circuit capability of new generation Enhancement-mode p-GaN power HEMTs. The electrothermal behavior of two commercially available devices is experimentally verified up to the failure events. Mission profile compact thermal simulations are used to estimate temperature increase during short-circuit tests. The assumption of a temperature-dependent gate current is then investigated by means of 2D electro-thermal TCAD simulations on a reference structure. Finally, a possible trade-off between gate driver resistance and short-circuit capability is discussed.

Spreadsheets are widely used in industries for tabular data analysis, visualization and storage. Users often exchange spreadsheets' semi-structured data to collaborative analyze them. Recently, office suites integrated a software module that enables collaborative authoring of office files, including spreadsheets, to facilitate the sharing process. Typically spreadsheets collaborative authoring applications, like Google Sheets or Excel online, need to delocalize the entire file in public cloud storage servers. This choice is not secure for enterprise use because it exposes shared content to the risk of third party access. Moreover, available platforms usually provide coarse grained spreadsheet file sharing, where collaborators have access to all data stored inside a workbook and to all the spreadsheets' formulas used to manipulate those data. This approach limits users' possibilities to disclose only a small portion of tabular data and integrate data coming from different sources (spreadsheets or software platforms). For these reasons enterprise users prefer to control fine grained confidential data exchange and their updates manually through copy, paste, attach-to-email, extract-from-email operations. However unsupervised data sharing and circulation often leads to errors or, at the very least, to inconsistencies, data losses, and proliferation of multiple copies. We propose a model that gives business users a different level of spreadsheet data sharing control, privacy and management. Our approach enables collaborative analytics of tabular data focusing on fine grained spreadsheet data sharing instead of coarse grained file sharing. This solution works with a platform that implements an end to end encrypted protocol for sensitive data sharing that prevents third party access to confidential content. Data are never shared into public clouds but they are transferred encrypted among the administrative domains of collaborators. In this paper we describe the model and the implemented system that enable our solution. We focus on two enterprise use cases we implemented describing how we deployed our platform to speed up and optimize industry processes that involve spreadsheet usage.