Grid forming converters can offer fast voltage support during grid disturbances through a prompt injection of reactive current. Nevertheless, the large overcurrents experienced during grid faults can trigger the converter current limiter, which is often responsible of transient instability problems. While over-dimensioning the power rating of the converter enhances its grid supporting capabilities, it significantly increases the costs. Online monitoring of the semiconductor junction temperature, instead, enables the converter to perform a safe over-current operation for a limited time interval, potentially improving the transient stability without converter overdimensioning. Nevertheless, the transient stability analysis and current limitation design considering thermal dynamics is very challenging due to the high order of the model, which makes conventional methodologies based on second-order models not applicable. This article proposes a temperaturedependent current limitation in grid-forming converters, enabling overcurrent operation and significantly improving transient stability. A methodology for transient stability analysis is proposed based on numerical methods applied to the nonlinear state-space model of the converter under thermally stressed conditions during over-current scenarios caused by grid faults. The proposed method is implemented in a simulation of a grid-forming medium voltage Modular Multilevel Converter (MMC) considering the dynamics of junction temperatures.

Objective. Non-invasive solutions for providing artificial sensory feedback to lower-limb prosthesis users are compact and convenient for clinical translation because they do not require additional surgery. However, they are mostly simpler feedback schemes characterized by limited information bandwidth and low spatial resolution. Additionally, feedback is often assessed using specialized tasks and conditions, which sometimes promote the use of feedback, limiting comprehensive psychophysical and ecological insights. Approach. This study introduces OmniFeel, a novel feedback system composed of eight vibration motors and a sensorized insole, to intuitively convey omnidirectional foot pressure information. It was evaluated psychophysically to test pattern recognition (static and dynamic) and holistically by tracking biomechanical, task load, and user experience outcome measures during an ecological walking task that resembled real-life scenarios. The holistic assessment included walking in a building (overground walking, stairs) with and without a parallel cognitive task and with and without feedback. Ten able-bodied participants, two participants with transtibial amputations (TT1 and TT2), and one with transfemoral amputation (TF1) took part in both assessments. Main results. The feedback scheme was easy to interpret, with a high success rate in recognizing six static and four dynamic spatial patterns, even before systematic training (81.5 ± 7.87% and 95.75 ± 4.42%, respectively). Functional evaluation demonstrated that feedback decreased the task load (NASA-TLX) in most conditions and participants. In two participants with lower-limb amputation, the feedback also improved stance time symmetry (from 55.88% to 74.34% in TF1 and 65.41% to 74.71% in TT2) and substantially increased confidence in TF1, especially in stair ambulation. Significance. The proposed feedback conveys rich information about the foot sole pressure and yet it is easy to interpret. The holistic assessment demonstrates that the feedback can improve the use of a lower limb prosthesis in realistic conditions, but that this depends on the task and participant characteristics.

The fabrication and assembly of High Temperature Superconducting (HTS) magnets can be significantly streamlined by (1) direct deposition of HTS onto modular components and (2) laser-ablated grooves to bound and guide the electric currents over the superconducting surfaces. “Coils” bounded by consecutive grooves can be individually powered, or connected in series, or in parallel. Applications include plasma-confinement devices such as stellarators and magnets for particle accelerators. On the way to HTS magnets generating more complicated three-dimensional fields, this paper validates the technique for two cylindrically symmetric magnets made of standard conductors. The optimized grooving pattern is considered as an inverse problem that is resolved using a least squares approach with Tikhonov regularization. This approach was first applied to design a magnet that replicates the magnetic field configuration of a microwave source of the gyrotron type. Gyrotrons require a particular profile of magnetic field, which our aluminium prototype successfully reproduced with 10−1 precision. The second one, in copper, is a small-size, reduced-field magnet for Magnetic Resonance Imaging (MRI). This application requires highly uniform longitudinal fields. The achieved precision (about 50 ppm) is exceptional for an MRI magnet just getting out of the factory, and can easily meet hospital standards after minimal shimming or other error field correction. Future work and other potential applications of wide, patterned conductors or superconductors are also discussed.

Artificial Intelligence has made significant strides, leading to the creation of advanced models like ChatGPT, developed by OpenAI. ChatGPT interacts with users in a conversational style, generating human-like responses across various domains. Despite its versatility, the accuracy of ChatGPT's responses can vary depending on the context, the complexity of the queries, and the repetition of prompts within a short time frame. To address these challenges and highlight potential issues, propose the Precision Answer Comparison and Evaluation Model (PACEM), which is designed to thoroughly assess and evaluate ChatGPT's performance. PACEM focuses on evaluating ChatGPT's accuracy and consistency across diverse domains, including literature, history, law or ethics, and sports or athletics. By systematically comparing responses and measuring performance, PACEM provides a comprehensive analysis of ChatGPT's capabilities and limitations in delivering reliable information. Additionally, this model is equipped to measure the processing time during evaluation, considering the time taken by ChatGPT to generate responses compared to the actual answers. The model reveals that ChatGPT generally delivers high-performance and accurate responses, often surpassing those generated by humans or composed from alternative ideas. However, the evaluation also indicates that the processing time may increase when the responses are lengthy, despite the accuracy of the answers provided by ChatGPT. Finally, this study concludes with key findings and suggests future work to address new challenges that may arise in similar contexts.

Detecting and characterizing oil spills in arctic environments is challenging due to the presence of sea ice. Methods based on Ground-Penetrating Radars and acoustic sensors were used in the literature to detect oil in ice-covered waters; however, their implementation is costly, and their capabilities are limited in terms of oil-thickness estimation. To address this problem, we propose a new low-cost sensing system based on a planar capacitive sensor that can detect the presence of oil under ice sheets and measure its thickness and depth. Our proposed sensor is based on movable dual electrodes mounted next to each other on the same sensing plane. The electric field created between the electrodes extends beneath the sensing plane and is affected by the electric properties of the sensed material (ice, oil, or water). Changes in the mutual capacitance of the electrodes are related to changes in the thickness and depth of the embedded oil phase. The capacitance of the sensor is measured at two different excitation frequencies while changing the horizontal separation distance between the electrodes. These measurements are collected to create a dataset for training machine-learning-based classification and regression models to detect the presence of oil and measure its thickness and depth. In comparison with the available sensing techniques, our proposed sensor has several advantages, such as being noninvasive or non-intrusive, simple to manufacture, safe to operate, and having low cost and low maintenance requirements. The experimental evaluation described in this paper demonstrates the effectiveness of our proposed system, which showed a very high detection accuracy of more than 90% and an accurate thickness and depth estimation capability with a Mean Absolute Error (MAE) of around 0.5 cm for thickness and depth estimations for oil thicknesses ranging between 0.5 cm and 8 cm and for oil depths ranging between 2 and 5 cm.

The deployment of fifth-generation (5G) networks across various industry verticals is poised to transform communication and data exchange, promising unparalleled speed and capacity. However, the security concerns related to the widespread adoption of 5G, particularly in mission-critical sectors, present significant challenges. This article investigates the potential of a Zero Trust (ZT) security philosophy as a viable countermeasure to these concerns. It delves into the practicalities of implementing ZT principles within 5G networks, with a specific focus on harnessing AI/ML technologies for proactive security measures, dynamic policy adaptations, and advanced risk assessments. Further, the article underscores the importance of developing a tailored ZT maturity model for 5G networks. Furthermore, the paper outlines key future research directions aimed at improving the ZT maturity of 5G deployments, contributing to the safe and secure integration of 5G technology in various sectors.

Unmanned aerial vehicles (UAVs) can enhance wireless access by dynamically positioning themselves closer to users while maintaining a backhaul connection to cellular base stations. In scenarios where users are geographically dispersed, multiple UAVs can be orchestrated to establish multi-hop integrated access and backhaul (IAB) connections. Traditionally, finding optimal UAV placement has required computationally demanding methods, such as combinatorial optimization or reinforcement learning, which are often impractical for real-time applications due to their complexity and training requirements. Additionally, UAV operators may lack the capability to solve complex optimization problems during live operations. This paper presents a novel framework that leverages large language models (LLMs) for optimizing the placement of multiple UAVs through iterative structured prompting. The proposed method achieves nearoptimal solutions in significantly fewer iterations compared to traditional methods, making it suitable for real-time deployment without extensive mathematical modeling. Through simulations, we demonstrate that this approach can efficiently manage UAV placements and ensure robust network connectivity and service quality. The LLM-driven framework simplifies problem-solving for UAV network operators, paving the way for its application in more complex real-world scenarios.

Complex Query Answering (CQA) on knowledge graphs is a fundamental yet challenging task, which can be formalized as answering a subset of first-order logic queries containing logical conjunction, disjunction, negation, and existential quantifiers. Recent research reveals that Link Predictors (LPs) trained on 1-hop queries can generalize to various types of complex queries. However, existing methods neglect crucial characteristics of LPs' outputs, including the effects of highly relevant entities and uncertainty. What's worse, as they model logical operations by fuzzy set operations, these methods suffer from problems like inflexibility, sensitivity to noise, and inconsistency with priority in human cognition, which limits their performance, especially on queries with negation. To address these challenges, we propose an efficient fuzzy system for CQA that requires no extra training overheads and is plug-and-play with existing LP-based methods. Firstly, we expand the output of LPs by two complementary membership functions of weak and strong relevance, which help to distinguish the target entities from highly relevant and irrelevant entities. Subsequently, we model logical operations through fuzzy rule bases and infer the final predictions via defuzzification, providing a flexible and tractable scheme for modeling logical operations. Finally, the effectiveness of the proposed fuzzy system is validated by its outstanding performance on benchmark datasets when compared to state-of-theart methods. The source code of our proposed method is available at https://anonymous.4open.science/r/FuzzSys-CQA-E285.

In this study, seven types of Chameleon Cloud compute nodes (ARM64, Haswell, Skylake, Cascade Lake, Ice Lake R650, Ice Lake R750, and Zen 3) are benchmarked using an HPC software suite that is meant to study polymorphic computing architectures. This benchmark is sensitive to the compute nodes' integer, memory, and threading performance. Each of the compute nodes is evaluated for execution time, speedup, and efficiency across a sweep of worker thread counts. The ARM64 compute node is comparatively the least performing compute node by a wide margin. The best performing compute nodes were the Ice Lake and Zen 3 systems. These nodes performed roughly equally.

The GEDI spaceborne lidar system was specifically designed to study forest ecosystems. Inference on forest attributes using GEDI data was mostly addressed through model-assisted, model-based and hybrid approaches. In this study, we applied a double sampling for post-stratification (DSPS) design-based approach to combine GEDI and national forest inventory data. Although widely used in the field of forest inventories, the use of such a design-based approach relying on GEDI data has not yet been investigated. This method is advantageous because it requires neither precise geolocation nor co-location between GEDI footprints and inventory plots. We evaluated the impact of the bridge variable and the impact of GEDI’s spatial sampling pattern on the results of the DSPS approach by comparing our GEDI-based results to reference airborne-laser-based results. We employed maximum tree height as the bridge variable and chose a complex study area in northeastern France with relief and highly diverse forest stands. We used 202,808 GEDI footprints as the first-phase sample and 476 National Forest Inventory (NFI) plots as the second-phase sample to estimate the growing stock volume (GSV). Compared with estimates based solely on NFI field plots, the DSPS approach reduced the GSV variance by up to 54% without any additional cost, aside from the negligible additional time required to download and process the GEDI data. GEDI can thus be considered as an effective data source to post-stratify NFI data and provide forest attribute estimates with a greater precision or at a finer spatial scale.

A document by Hamid Davoudi. Click on the document to view its contents.

This paper addresses the growing interest in approximate computing circuits, recognized as promising alternatives for accelerating hardware for applications such as machine learning and image processing that demonstrate an ability to tolerate a certain level of noise and variations within the system. While numerous papers delve into diverse techniques for designing approximate arithmetic circuits, a notable gap exists in unified hardware libraries and design automation tools essential for enabling comprehensive design space exploration and comparative analysis across various applications. Our work herein involves the development of highly parameterized register transfer level (RTL) and high-level synthesis (HLS) libraries, encompassing a diverse array of approximate adders and multipliers. Additionally, we introduce a Python-based design automation tool designed to streamline the hyperparameter tuning, synthesis, and implementation processes for these circuits. The frameworks presented in this paper facilitate an unprecedented scale of comprehensive, comparative and quantitative analysis for approximate circuits, which to the best of our knowledge, have not been previously achieved in the literature.

A document by Shuhan Li. Click on the document to view its contents.

Advances in deep learning (DL)-based models for brain image analysis have significantly enhanced the accuracy of Alzheimer's disease (AD) diagnosis, allowing for more timely interventions. Despite these advancements, most current DL models suffer from performance degradation when inferring on unseen domain data owing to the variations in data distributions, a phenomenon known as domain shift. To address this challenge, we propose a novel approach called the dynamic frequency mixup scheduler (DyMix) for unsupervised domain adaptation. Contrary to the conventional mixup technique, which involves simple linear interpolations between predefined data points from the frequency space, our proposed DyMix dynamically adjusts the magnitude of the frequency regions being mixed from the source and target domains. Such an adaptive strategy optimizes the model's capacity to deal with domain variability, thereby enhancing its generalizability across the target domain. In addition, we incorporate additional strategies to further enforce the model's robustness against domain shifts, including leveraging amplitude-phase recombination to ensure resilience to intensity variations and applying self-adversarial learning to derive domaininvariant feature representations. Experimental results on two benchmark datasets quantitatively and qualitatively validated the effectiveness of our DyMix in that we demonstrated its outstanding performance in AD diagnosis compared to state-of-the-art methods. The code is available at: https://github.com/ku-milab/DyMix.

In addition to the growing need for wireless connectivity, the advancement of 6G technology is also driven by the dynamic and varied requirements of different vertical sectors. The vision of 6G extends beyond enabling pervasive and ubiquitous wireless intelligence and addresses seamless communication between humans, machines, and everything in between. This paper introduces key performance indicators (KPIs) as crucial metrics for identifying vertical demand and conducting digitalization requirements analysis within the energy sector. Moreover, it highlights the smart grid use cases with their specified communication KPI values. Finally, the digitization of energy systems will be discussed, along with its potential challenges, opportunities, and solutions.

In many contexts, particularly in Electromagnetic Compatibility (EMC), a common-mode Current Probe (CP) is intended to be a non-invasive transducer. We introduce its applications generally, but focus on transient waveform measurements. The complex form of our 10 Hz to 1 GHz CP Transfer Impedance (Zt) is measured using Vector Network Analysers. Measured data are compared to manufacturer calibration data and a choice made on the best form of the complex Zt for transient analysis. We then inject a time waveform into a test system and use the Zt to reconstruct the signal, using the inverse Fast Fourier Transform algorithm. The reconstructed current waveform is compared to a reference observation for validation. For frequencies between DC and 240 MHz, a good correlation can be achieved. We raise relevant discussions on filtering, calibration system resonance and de-embedding.  

Voice preference, unlike other areas of human voice research, has been relatively underexplored, despite being a major category in voice evaluation. Current studies on voice preference have not yet leveraged music harmony theory, which provides a clear standard for sounds that are generally considered pleasant. In this study, I investigated the relationship between the harmonic structure of voices and their perceived preferences. Ten voice samples in Korean were evaluated for preference using the Mean Opinion Score (MOS) method, and the results were compared to voice harmony analyses through two approaches: the cent method and the weight method. The results demonstrated that voices with higher preference ratings tended to align more closely with musical harmony, offering a novel approach to understanding voice likability. These findings could have broad implications for fields such as voice synthesis and human-computer interaction.

An effective and important technique, Deep Learning has been found to be recurrently used in different fields, including computer vision, machine vision, and natural language processing. As a result of this technology, deepfakes generate visual media of a particular person that are nearly similar to the original material. Over the past few years, a variety of studies have examined how deepfakes operate, and numerous methods utilizing deep learning have come forward to detect both deepfakes videos and images. In this work, the dangers deepfakes present to corporate cybersecurity are investigated, underscoring the need to understand the technology related to deepfakes and the importance of constructing effective detection and mitigation methods. The article includes coverage of the merits and shortcomings of several methods for locating deepfakes, including deepfake detection algorithms along with human review and contextual analysis. A technique that unites automated detection with human judgment is critical for discovering deepfakes. The article brings to attention the necessity for businesses and individuals to strengthen their awareness and evasion capabilities against deepfake deception, decrease the risk that deepfakes can get, and take advantage of the potential opportunities deepfakes might create. Key takeaways reveal that facial landmark analysis, lip synchronization detection, and micro-expression analysis play key roles in deepfake detection and that a zero-trust and multi-factor authentication protocol is necessary to address the risks associated with deepfakes.