In this paper, an ultrasonic positioning system is presented and characterized, based on the usage of a portable grid of beacons and of a few fixed anchors. Since the beacon grid can be moved to guarantee line of sight transmissions, the proposed strategy is potentially suitable for accurate positioning of a mobile object in an environment with a complex geometry. The system was tested experimentally, exhibiting a sub-centimeter positioning accuracy in a range up to 4 m.

Many algorithms and methods have been proposed for Computed Tomography (CT) image reconstruction, partic- ularly with the recent surge of interest in machine learning and deep learning methods. The majority of recently proposed methods are, however, limited to the image domain processing where deep learning is used to learn the mapping from a noisy image data set to a true image data set. While deep learning-based methods can produce higher quality images than conventional model-based post-processing algorithms, these methods have lim- itations. Deep learning-based methods used in the image domain are not sufficient for compensating for lost information during a forward and a backward projection in CT image reconstruction especially with a presence of high noise. In this paper, we propose a new Recurrent Neural Network (RNN) architecture for CT image reconstruction. We propose the Gated Momentum Unit (GMU) that has been extended from the Gated Recurrent Unit (GRU) but it is specifically designed for image processing inverse problems. This new RNN cell performs an iterative optimization with an accelerated convergence. The GMU has a few gates to regulate information flow where the gates decide to keep important long-term information and discard insignificant short- term detail. Besides, the GMU has a likelihood term and a prior term analogous to the Iterative Reconstruction (IR). This helps ensure estimated images are consistent with observation data while the prior term makes sure the likelihood term does not overfit each individual observation data. We conducted a synthetic image study along with a real CT image study to demonstrate this proposed method achieved the highest level of Peak Signal to Noise Ratio (PSNR) and Structure Similarity (SSIM). Also, we showed this algorithm converged faster than other well-known methods.

Nuclear power is advancing slowly because of public concerns about nuclear accidents, radioactive waste, fuel supply, cost, and nuclear proliferation. The development of molten salt reactors could alleviate most of these concerns and prevent water-cooled reactor accidents like those at Three Mile Island, Chernobyl, and Fukushima. The purpose of this article is to provide information about the potential advantages and problems of molten salt reactors.

The dataset used in this work is obtained from a semiconductor factory, SECOM (Semiconductor Manufacturing) dataset and is publicly available online.

In this paper, a new cascade one-dimensional convolution neural network (1DCNN) and bidirectional long short-term memory (BLSTM) model has been developed for binary and ternary classification of mental workload (MWL). MWL assessment is important to increase the safety and efficiency in Brain-Computer Interface (BCI) systems and professions where multi-tasking is required. Keeping in mind the necessity of MWL assessment, a two-fold study is presented, firstly binary classification is done to classify MWL into Low and High classes. Secondly, ternary classification is applied to classify MWL into Low, Moderate, and High classes. The cascaded 1DCNN-BLSTM deep learning architecture has been developed and tested over the Simultaneous task EEG workload (STEW) dataset. Unlike recent research in MWL, handcrafted feature extraction and engineering are not done, rather end-to-end deep learning is used over 14 channel EEG signals for classification. Accuracies exceeding the previous state-of-the-art studies have been obtained. In binary and ternary classification accuracies of 96.77% and 95.36% have been achieved with 7-fold cross validation, respectively.

In this paper, we present an integrated and fire neuron designed in a 22-nm FDSOI technology. In this novel design, we deploy the back-gate terminal of FDSOI technology for a tunable design. For the first time, we show analytically and with pre- and post-layout simulations a neuron with tunable spiking frequency using the back-gate voltage of FDSOI technology. The neuron circuit is designed in the sub- hreshold region and dissipates an ultra-low energy per spike of the order of Femto Joules per spike. With the layout area of only 30 um2, this is the smallest neuron circuit reported to date.

We present a cache attack monitoring methodology that leverages statistical machine learning models to detect n-day hardware attacks by analyzing the electromagnetic emanations of a device. Experimental results from a Raspberry Pi 4 hosting Linux and a Jetson TX2 development board running a Linux guest hosted by seL4 demonstrate that our approach can sense Spectre attacks with a concordance statistic of 97% and 95%.

In this manuscript, novel methods for the detection of multiple superimposed tones in noise and the estimation of their parameters are derived, and their application to colocated multiple-input multiple-output radar systems is investigated.

Owing to the shorter wavelength in the millimeter-wave (mmWave) spectrum, miniaturized antennas can receive power with a higher efficiency than UHF bands, promising sustainable mmWave-powered Internet of Things (IoT) devices. Nevertheless, the performance of a mmWave power receiver has not been compared, numerically or experimentally, to its sub-6 GHz counterpart. In this paper, the performance of mmWave-powered receivers is evaluated based on a novel wearable textile-based higher-order mode microstrip antenna, showing the benefits of wireless power transmission (WPT). Firstly, a broadband antenna is proposed maintaining a stable wearable measured bandwidth from 24.9 to 31.1 GHz, over three-fold improvement compared to a conventional patch. The proposed antenna has a measured 8.2 dBi co-polarized gain with the highest thickness-normalized efficiency of a wearable antenna. When evaluated for compact power receivers, the measured path gain shows that WPT at 26 GHz outperforms 2.4 GHz by 11 dB. A rectenna array based on the proposed antenna is then evaluated analytically showing the potential for up to 6.3x higher power reception compared to a UHF patch, based on the proposed antenna’s gain and an empirical path-loss model. Both use cases demonstrate that mmWave-powered rectennas are suitable for area-constrained and large-area wearable IoT applications.

In this paper, we propose a symmetric series convolutional neural network (SS-CNN), which is a novel deep convolutional neural network (DCNN)-based super-resolution (SR) technique for ultrasound medical imaging. The proposed model comprises two parts: a feature extraction network (FEN) and an up-sampling layer. In the FEN, the low-resolution (LR) counterpart of the ultrasound image passes through a symmetric series of two different DCNNs. The low-level feature maps obtained from the subsequent layers of both DCNNs are concatenated in a feed forward manner, aiding in robust feature extraction to ensure high reconstruction quality. Subsequently, the final concatenated features serve as an input map to the latter 2D convolutional layers, where the textural information of the input image is connected via skip connections. The second part of the proposed model is a sub-pixel convolutional (SPC) layer, which up-samples the output of the FEN by multiplying it with a multi-dimensional kernel followed by a periodic shuffling operation to reconstruct a high-quality SR ultrasound image. We validate the performance of the SS-CNN with publicly available ultrasound image datasets. Experimental results show that the proposed model achieves an exquisite reconstruction performance of ultrasound image over the conventional methods in terms of peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM), while providing compelling SR reconstruction time.

Real world signal acquisition through sensors, is at the heart of modern digital revolution. However, almost every signal acquisition systems are contaminated with noise and outliers. Precise detection, and curation of data is an essential step to reveal the true-nature of the uncorrupted observations. With the exploding volumes of digital data sources, there is a critical need for a robust but easy-to-operate, low-latency, generic yet highly customizable, outlier detection and curation tool, easily accessible, adaptable to diverse types of data sources. Existing methods often boil down to data smoothing that inherently cause valuable information loss. We have developed a C++ based, software tool to decontaminate time- series and matrix like data sources, with the goal of recovering the ground-truth. The SOCKS tool would be made available as an open-source software for broader adoption in the scientific community. Our work calls for a philosophical shift in the design pipelines of real- world data processing. We propose, raw data should be decontaminated first, through conditional flagging of outliers, curation of flagged points, followed by iterative, parametrically tuned, asymptotic converge to the ground-truth as accurately as possible, before performing traditional data processing tasks.

This paper introduces a novel transient simulation scheme and its extension for stability analysis of distribution systems with converter-based distributed generation. Efficiency and accuracy of the simulation scheme with extension are investigated. It is demonstrated that millisecond-level step sizes are applicable to efficient electromagnetic transient simulation for stability analysis of unbalanced power systems while maintaining satisfactory accuracy. The paper also supports the extension of the simulation scheme by showing that the extension has only an insignificant impact on the accuracy but significantly enhances the efficiency. This paper has been accepted by the 2021 IEEE IAS Annual Meeting.

In this paper, we aim to design a deep reinforcement learning (DRL) based control solution to navigating a swarm of unmanned aerial vehicles (UAVs) to fly around an unexplored target area under partial observation, which serves as Mobile Base Stations (MBSs) providing optimal communication coverage for the ground mobile users. To handle the information loss caused by the partial observability, we introduce a novel network architecture named Deep Recurrent Graph Network (DRGN), which could obtain extra spatial information through graph-convolution based inter-UAV communication, and utilize historical features with a recurrent unit. Based on DRGN and maximum-entropy learning, we propose a stochastic DRL policy named Soft Deep Recurrent Graph Network (SDRGN). In SDRGN, a heuristic reward function is elaborated, which is based on the local information of each UAV instead of the global information; thus, SDRGN reduces the training cost and enables distributed online learning. We conducted extensive experiments to design the structure of DRGN and examine the performance of SDRGN. The simulation results show that the proposed model outperforms four state-of-the-art DRL-based approaches and three heuristic baselines, and demonstrate the scalability, transferability, robustness, and interpretability of SDRGN. We release our code at https://github.com/yerfor/Soft-DRGN

An accelerated Integral Equations (IE) field solver for determining scattered fields from electrically large electromagnetic metasurfaces utilizing Fast Multipole Method (FMM) is proposed and demonstrated in 2D. In the proposed method, practical general metasurfaces are expressed using an equivalent zero thickness sheet model described using surface susceptibilities, and where the total fields around it satisfy the Generalized Sheet Transition Conditions (GSTCs). While the standard IE-GSTC offers fast field computation compared to other numerical methods, it is still computationally demanding when solving electrically large problems, with a large number of unknowns. Here we accelerate the IE-GSTC method using the FMM technique by dividing the current elements on the metasurface into near- and far-groups, where either the rigorous or approximated Green’s function is used, respectively, to reduce the computation time without losing solution accuracy. Using numerical examples, the speed improvement of the FMM IE-GSTC method O(N1.5) over the standard IE-GSTC O(N3) method is confirmed. Finally, the usefulness of the FMM IE-GSTC is demonstrated by applying it to solve electromagnetic propagation inside an electrically large radio environment with strategically placed metasurfaces to improve signal coverage in blind areas, where a standard IE-GSTC solver would require prohibitively large computational resources and long simulation times.

This paper investigates the problem of localization of co-channel transmitters or primary users (PUs) using an array mounted on a moving aerial platform. As a practical alternative for a sensor network to pursue the localization task, the proposed Phase Interferometric Source Localization (PISL) technique utilizes a moving sensor that measures phase difference between two antennas mounted on the platform. Due to the sparse nature of PUs’ distribution in the region, we model the localization task as a simple basis-pursuit denoising framework and introduce a reconstruction method using a sparse recovery algorithm to discover locations of unknown PUs based on the phase difference measurements. We show that the ratio of distance between two antennas to the carrier-frequency wavelength is the critical parameter making the localization feasible. We also propose a scheme for sensor motion design in order to maximize the number of detectable PUs based on mutual coherence property. Since the motion optimization problem is very hard to address we develop a simple geometric relaxation to address the problem. The simulation results show that PISL can precisely recover the map of PUs with only few measurements and also reveal that sensor motion path can have determinate effect on the localization accuracy. PISL performance is compared with an state-of-the-art technique that utilizes adaptive beamforming and results show the superiority of PISL results in localization accuracy.

This work proposes a methodology for the energy-and cost-efficient 3-D deployment of an unmanned aerial vehicle (UAV)-based aerial access point (AAP), that exchanges a given amount of independent data with a set of ground user equipment (UE). Considering a fly-hover-communicate transmission scheme, the most energy-efficient 3-D hovering points (HPs) of the AAP are determined by decoupling the problem in the horizontal and vertical dimensions. First, we derive analytically the optimal hovering altitude that jointly maximizes the downlink and uplink global energy efficiency (GEE) of the system. Next, we propose the multilevel circle packing (MCP) algorithm to determine the minimal number of HPs and their associated horizontal coordinates, such that the AAP covers all the UEs in the given geographical area. A cost analysis is carried out to observe the variation of both fixed and variable costs; these are then minimized by suitably selecting the AAP’s battery parameters, like the depth of discharge (DOD), defined as the portion of battery capacity that is consumed during a discharge cycle, and the velocity of the UAV. Simulation results show that: the UAV energy consumption has a significant impact on the 3-D HPs of the AAP; the time spent during the substitution swap of an out of power AAP has a major influence on the operational cost; the cost of the system can be optimized by suitably selecting the onboard battery and the UAV flight parameters.

In this paper, we propose a novel reinforcement-learning-based framework for left ventricle contouring, which mimics how a cardiologist outlines the left ventricle in a cardiac image. Since such a contour drawing process is simply moving a paintbrush along a specific trajectory, it is thus analogized to a path finding problem.

In this paper, we demonstrate the significance of restoring harmonics of the fundamental frequency (pitch) in deep neural network (DNN) based speech enhancement. We propose a sliding-window attention network to regress the spectral magnitude mask (SMM) from the noisy speech signal. Even though the network parameters can be estimated by minimizing the mask loss, it does not restore the pitch harmonics, especially at higher frequencies. In this paper, we propose to restore the pitch harmonics in the spectral domain by minimizing cepstral loss around the pitch peak. The network parameters are estimated using a combination of the mask loss and cepstral loss. The proposed network architecture functions like an adaptive comb filter on voiced segments, and emphasizes the pitch harmonics in the speech spectrum. The proposed approach achieves comparable performance with the state-of-the-art methods with much lesser computational complexity.