Search

A nonlinear embedding model for a body-grounded version of the industry standard ASM-HEMT model is presented in this paper for accelerated power amplifier (PA) design. This embedding model is able to simulate any PA mode of operation at its intrinsic reference planes and predict the impedances and harmonic terminations required at the extrinsic reference planes to sustain that targeted mode for the modeled transistor. The ASM-HEMT embedding model is further extended to account for the transistor's manifold using six-port ABCD parameters, enabling more accurate design of MMIC PAs.

In this work we consider robust stabilization of uncertain dynamical systems and show that this can be achieved by solving a non-classically constrained analytic interpolation problem. In particular, this non-classical constraint confines the range of the interpolant, when evaluated on the imaginary axis, to a frequency-dependent set. By considering a sufficient condition for when this interpolation problem has a solution, we derive an approximate solution algorithm that can also be used for controller synthesis. Finally, the theory is illustrated on a numerical example with a plant with uncertain gain, phase, and output delay., 8 pages, 2 figures

This paper presents for the first time an investigation of the ASM-HEMT model large-signal accuracy across a wide range of operating frequencies. Comparisons between measured and simulated power sweeps are presented at 10, 20 and 30 GHz. Load-pull measurements and simulations are compared spanning the same frequency range. The large-signal simulations of a single ASM-HEMT model exhibit excellent agreement with all power measurements.

Building on a recent paper by Georgiou and Lindquist [1] on the problem of rank deficiency of spectral densities and hidden dynamical relations after sampling of continuous-time stochastic processes, this paper is devoted to understanding related questions of feedback and Granger causality that affect stability properties. This then naturally connects to CARMA identification, where we remark on certain oversights in the literature., Comment: 7 pages, 2 figures

Gaussian elimination is a key technique for solving dense, non-symmetric systems of linear equations. Pivoting is used to ensure numerical stability but can introduce significant overheads. We propose replacing pivoting with recursive butterfly transforms (RBTs) and iterative refinement. RBTs use an FFT-like structure and randomized elements to provide an efficient, two-sided preconditioner for factoring. This approach was implemented and tested using Software for Linear Algebra Targeting Exascale (SLATE). In numerical experiments, our implementation was more robust than Gaussian elimination with no pivoting (GENP) but failed to solve all the problems solvable with Gaussian elimination with partial pivoting (GEPP). Furthermore, the proposed solver was able to outperform GEPP when distributed on GPU-accelerated nodes.

In this work, a multi-sensor approach to separate oil palm plantations from forest cover using NDFI and a modified Pauli Decomposition technique is presented. The main contribution of this research is the potential to reduce misclassification of both classes, in the context of automated-base supervised classification algorithms, to decrease uncertainties derived through the detection and mapping process of forest cover. The hereby proposed method includes the generation of a primary forest map cover defining thresholds from a high resolution multi -spectral satellite image, and then the palm oil plantation will be filtered out from this classification using scattering mechanisms by a Pauli Decomposition approach. Preliminary results shown the capabilities of this approach in order to generate complementary information to separate the oil palm plantations from the forest cover classification.

The variety and scope of available Internet of Things (IoT) devices is constantly increasing, along with the number of manufacturers that offer these things. As these numbers increase, so does the fragmentation of these ecosystems; either through incompatible means of communication and operation, or through the innate dependencies of a device’s ecosystem (things with similar programming interfaces may be tied to different "clouds", rendering them static and unable to form external interactions). We argue that the future of IoT depends on the ability of things to work together, and to be dynamic and reprogrammable to create exciting and opportune applications; but that to achieve this, things must share some level of common ground. To this end, we introduce the Internet of Things Device Description Language (IoT-DDL) to harness these commonalities while minimizing friction and new requirements on existing thing functionality. IoT-DDL, a human- and machine-readable schema, is designed to reflect these shareable capabilities while providing a common structure that things can use to self-discover and share their entities, services, and cloud-based extensions within a smart space. In this paper, we present an overview of the IoT-DDL, along with rationale and examples for its parts. We also introduce a tool to ease the creation of IoT-DDL files for thing vendors, developers, and end users.

We present techniques of generating data for mixed precision solvers that allows to test those solvers in a scalable manner. Our techniques focus on mixed precision hardware and software where both the solver and the hardware can take advantage of mixing multiple floating precision formats. This allows taking advantage of recently released generation of hardware platforms that focus on ML and DNN workloads but can also be utilized for HPC applications if a new breed of algorithms is combined with the custom floating-point formats to deliver performance levels beyond the standard IEEE data types while delivering a comparable accuracy of the results.

Twin study is one of the major parts of human brain research that reveals the importance of environmental and genetic influences on different aspects of brain behavior and disorders. Accurate characterization of identical and fraternal twins allows us to infer on the genetic influence in a population. In this paper, we propose a novel pair-wise classification pipeline to identify the zygosity of twin pairs using the resting state functional magnetic resonance images (rs-fMRI). The new feature representation is utilized to efficiently construct brain network for each subject. Specifically, we project the fMRI signal to a set of cosine series basis and use the projection coefficients as the compact and discriminative feature representation of noisy fMRI. The pair-wise relation is encoded by a set of twin-wise correlations between functional brain networks across brain regions. We further employ hill climbing variable selection to identify the most genetically affected brain regions. The proposed framework has been applied to 208 twin pairs in Human Connectome Project (HCP) and we achieved 92.23(±4.43)% classification accuracy.

Nowadays, the challenges that teachers face on different levels of education are varied and some are more complex than others, such is the case of the inclusion of students who present some disability in high level education with the rest of the class. The role of teacher is of the utmost importance, as this is the first contact with the student, it should be the support and bridge which will best manage the relationships between students, as well as finding the best path that allows all students to advance at the same pace and above all to make possible a better development in the classroom by the students. The Information and Communication Technologies (ICT's) play an important role in education, since they're one of the primary technological tools which the students count with to comprehend, communicate and exchange information between them, socialize and develop better and better as much as outside and inside the classroom.

In this article, we call for a "Walk Before You Run" adjustment in the Internet-of-Things (IoT) research and development exercise. Without first settling the quest for what thing is or could be or do, we run the risk of presumptuous visions, or hypes, that can only fail the realities and limits of what is actually possible, leading to customers and consumers confusion as well as market hesitations. Specifically, without a carefully-designed Thing architecture in place, it will be very difficult to find the "magic" we are so addicted and accustomed to - programming! Programming the IoT, as we once programmed the mainframe, the workstation, the PC and the mobile devices, is the natural way to realize a fancy IoT scenario or an application. Without Thing architectures and their enablement of new programming models for IoT - we will continue to only envision fancy scenarios but unable to unleash the IoT full potential. This article raises these concerns and provides a view into the future by first looking back into our short history of pervasive computing. The article focuses on the domain of "Personal" IoT and will address key new requirements for such Thing architecture. Also, practicing what we preach, we present our ongoing efforts on the Atlas Thing Architecture showing how it supports a variety of thing notions, and how it enables novel models for programmability.

Beta-phase gallium oxide $(\beta-\mathrm{Ga}_{2}\mathrm{O}_{3})$ has shown promise as a next-generation wide-bandgap semiconductor for use in power electronics. It possesses a bandgap and expected critical field strength of ~4.8 eV and ~8 MV/cm, respectively, surpassing the same measured characteristics of GaN and SiC [1]. Early work has been successful in demonstrating lateral metal-oxide-semiconductor field-effect transistors (MOSFETs), predominately for depletion mode operation, with high critical field strength [2], high current density [3], and high breakdown voltage [4], [5]. One limitation of $\beta-\mathrm{Ga}_{2}\mathrm{O}_{3}$ MOSFETs is source access resistance $(\mathrm{R}_{\mathrm{S}})$ , the ungated region between source and gate, with sheet resistance $(\mathrm{R}_{\mathrm{SH}})$ typically in the $\mathrm{k}\Omega/\mathrm{sq}$ range. The $\mathrm{R}_{\mathrm{S}}$ affects key device performance parameters such as transconductance $(\mathrm{G}_{\mathrm{M}})$ and drain-current density $(\mathrm{I}_{\mathrm{DS}})$ . Higher DC and RF performance can be expected from eliminating $\mathrm{R}_{\mathrm{S}}$ by self-aligning the gate and source contacts. We present, for the first time, a self-aligned gate (SAG) $\beta-\mathrm{Ga}_{2}\mathrm{O}_{3}$ MOSFET using a refractory metal gate-first design. MOSFET devices with $2\mathrm{x}50\mu \mathrm{m}$ gate periphery, $7.5\ \mu \mathrm{m}$ source-drain distance $(\mathrm{L}_{\mathrm{SD}})$ and $2\ \mu \mathrm{m}$ gate length $(\mathrm{L}_{\mathrm{G}})$ were directly compared with and without the SAG features. MOSFETs with SAG show a substantial increase in $\mathrm{G}_{\mathrm{M}}$ and $\mathrm{I}_{\mathrm{DS}}$ , from ~2 mS/mm to ~14 mS/mm and ~10 mA/mm to ~45 mA/mm, respectively, up to $\mathrm{V}_{\mathrm{GS}}=4$ V. Lastly, we report a laterally scaled device $(2\mathrm{x}50\mu \mathrm{m})$ with $\mathrm{L}_{\mathrm{SD}}=2.5\mu \mathrm{m}$ and $\mathrm{L}_{\mathrm{G}}=2\mu \mathrm{m}$ , achieving high current density $(\sim 140\mathrm{mA}/\mathrm{mm})$ , and high G M $(\sim 35\mathrm{mS}/\mathrm{mm})$ .

The power switching potential of Ga 2 O 3 has been well established in the last several years due to the material's high critical electric field $(\mathrm{E}_{\mathrm{C}})$ strength. This allows for dramatic device scaling and operation at high voltage with reduced series resistance in power switching applications. The same argument applies for RF power amplification, where the high $\mathrm{E}_{\mathrm{C}}$ of Ga 2 O 3 allows FETs to be scaled to smaller dimensions while operating at higher operating voltages than existing RF FETs. At the same time, ab initio calculations published in 2017 for velocity-field characteristics of Ga 2 O 3 indicate a peak electron velocity of $2\mathrm{x}10^{7}$ cm/s which suggests the power-frequency product for Ga 2 O 3 is significantly greater than that of GaN. In 2017, the first demonstrated CW RF operation for Ga 2 O 3 FETs was in part limited by the thermal resistance and immaturity of heat extraction techniques to ~300 mW/mm. Since then, pulsed RF power measurement techniques have been implemented to accelerate electrical characterization of RF Ga 2 O 3 FETs. Pulsed RF output power $> 500$ mW/mm has been measured at 1 GHz. Advances in laterally scaled devices and topology design to ensure electrons achieve saturated velocity in the channel will be discussed.

In this paper, a reconfigurable microstrip patch antenna (MPA) in the S-band is presented. Three different nematic liquid crystals (NLCs) with birefringence Δn/Δe in the range of (0.05-0.25)/(0.18-0.80) were used as a grounded substrate below the microstrip patch. The full-wave numerical results of the MPA showed a tuning range from 1.8% to 6.8%.

In magnetomotive ultrasound (MMUS), solenoids have been extensively used to generate varying magnetic fields, that act to move nanoparticles and their surrounding tissue. Here we show, compare and contrast two magnetic devices, one based on a solenoid, another containing a rotating permanent magnet. Suitable field characteristics for this application include high flux density and gradient, and uniform field laterally in the image plane, and the corresponding design features are highlighted. Field measurements show that both magnets generate suitable fields. The magnetic force potentially exerted by the permanent magnet even exceeds that of the solenoid up to a depth of 4 mm. Considering the substantial heat generation in high inductance electromagnets, a permanent magnet would be preferable as it eliminates the need for cooling. Thus, we suggest the use of permanent magnets in MMUS devices to reduce the size, while maintaining a high field and gradient required to move nanoparticles.