The release of the Citizens Broadband Radio Service (CBRS) spectrum has enabled private network deployments, but signal strength in this band is highly sensitive to weather conditions. While traditional signal propagation models are often constrained by their assumptions and lack adaptability to varying environments, Machine Learning (ML) models offer a flexible alternative for predicting signal strength under complex conditions. In this work, we develop a Neural Network (NN) based predictive model for CBRS signal strength, leveraging a comprehensive set of weather parameters, collected over a 24month period from a live CBRS network in Buffalo, NY. Our model demonstrates a Mean Absolute Error (MAE) of 1.83 dB in predicting hourly signal strength, and we further evaluate its generalization capability by applying a transfer learning approach to another CBRS network in Elmira, NY, achieving a MAE of 2.13 dB. This study highlights the potential of ML methods in modeling weather-induced signal variability and their applicability across different network environments.

This paper introduces a new theoretical framework for edge detection and pioneers a novel technical route grounded in this framework. In this paper, texture is defined as the disparity in optical properties between two distinct regions. Based on this definition, an XOR operator is devised to precisely extract gradient and orientation information of textures from binary images. By exploiting the bit independence property and the bit difference weighted sum property of binary numbers, the application scope of the XOR operator is expanded, resulting in a new algorithm applicable to all types of image —the Bitwise Texture Extraction Algorithm (BTE). Edge is defined as the texture that delineates the contours of visible objects. Based on this definition, an algorithm named the Bitwise Edge Extraction Algorithm (BEE) is proposed, which is for extracting edge information from the output of BTE. Subsequently, the application of the output of the BEE algorithm, the BEE Map, in instance segmentation tasks is discussed, and the BEE algorithm is optimized by leveraging the linear continuity of edges, enhancing its accuracy in assigning edge affiliations. Finally, this paper tackles the challenge of Edge-to-Noise Separation by defining image noise as all elements in an image that do not constitute target elements. Based on this novel definition, a new approach to noise filtering is introduced: by applying operations that preserve only valid edge clusters for each object within the BEE Map individually, and then combining these outcomes, a noise-free, pure edge map is achieved.

A physics-based compact model for the metal-ferroelectric-insulator-metal (MFIM) tunneling junctions is developed, covering their essential dynamic features. The dual-insulator introduces a complex shape of tunneling barriers, which also shows a historical dependence on ferroelectric polarization. Voltage control of the ferroelectric tunnel junction (FTJ) is formulated using the dynamic nucleation-limited switching theory. Key operational voltages are analyzed, especially for those with zero internal fields. The tunneling transport is analytically formulated, covering the dual insulator barriers and the above special bias points based on the WKB approximations. Key features of the developed model include quasi-static and dynamic state transitions, historical dependence, and time-modulated multiple states. The model is verified through calibrated TCAD simulations and facilitates possible applications of MFIM FTJ in memory and computing.

The integration of vision and radar sensors is crucial for enhancing the perception and decision-making capabilities of autonomous cars. This survey paper reviews state-ofthe-art techniques in vision-radar fusion, categorizing and analyzing early, late, and hybrid fusion strategies. We discuss the strengths and limitations of individual sensors and how their complementary characteristics can improve overall system performance. Key topics include data preprocessing, feature extraction, and sensor alignment, as well as machine learning and deep learning approaches for effective fusion. Practical challenges such as sensor calibration, synchronization, and heterogeneous data handling are addressed. We evaluate fusion techniques in real-world scenarios, assessing their accuracy, robustness, and computational efficiency. This survey aims to guide researchers and practitioners in developing more reliable and efficient autonomous driving systems through advanced vision-radar fusion techniques, offering insights into future research directions and potential improvements.

This paper investigates an average secrecy rate (ASR) maximization problem for an unmanned aerial vehicle (UAV) enabled wireless communication system, wherein a UAV is employed to deliver confidential information to a ground destination in the presence of a terrestrial passive eavesdropper. By employing an artificial noise (AN) injection based secure two-phase transmission protocol, we aim at jointly optimizing the UAV’s trajectory, network transmission power, and AN power allocation over a given time horizon to enhance the ASR performance. Specifically, we divide the original non-convex problem into four subproblems, and propose a successive convex approximation based efficient iterative algorithm to solve it suboptimally with guaranteed convergence. Simulation results demonstrate significant security advantages of our designed scheme over other known benchmarks, particularly for stringent flight durations.

The cyclic prefix (CP) is a fundamental feature that enables orthogonal frequency division multiplexing (OFDM) to achieve efficient frequency-domain equalization and resilience against delay spread (DS). However, the trade-off is a reduction in spectral efficiency (SE), prompting the need for methods to minimize CP length without compromising performance. This paper introduces an innovative CP reduction approach leveraging array non-stationarity in extra-large MIMO (XL-MIMO) systems, particularly through the concept of visibility regions (VRs). Unlike conventional methods that adopt a worstcase scenario for CP length, our approach customizes the CP length based on the maximum excess delay of each VRs, thereby reducing the CP overhead. To enhance diversity gain, maximal ratio combining (MRC) is used to optimize the contributions from antenna elements. Simulation results validate our method's ability to improve SE while reducing latency and maintaining low computational complexity. The proposed approach demonstrates practical gains with standard CP configurations, marking a significant step toward more efficient and robust XL-MIMO OFDM systems.

Society’s dependence on telecommunications networks has increased significantly over the past 30 years. The expectations of always-available access, coupled with the demands for real-time video communications has driven a significant increase in network complexity. At the same time, there has been a shift to move industrial and machine communications from dedicated networks to shared networks, and an explosion of devices accessing these networks for a multitude of consumer and enterprise applications. The increased reliance on very complex networks requires a holistic approach to understanding the potential critical failure points in order to build the required network resilience. Driven by, and driving, this complex network transformation has been the the adoption of Internet Protocol (IP) not only for data services, but also mobile voice, Voice over Internet Protocol (VoIP) and interactive video services. In addition, IP networks are being increasingly used for critical industrial and commercial applications. This transition, however, exposes the fragility of centralised IP routing protocols, common distributed IP functions with common configuration options and the respective management systems. This paper examines the shift from traditional telecommunication redundancy models, which emphasised physical redundancy and the geographic isolation of failures, to the current IP-centric networks where even minor configuration errors can precipitate national-scale outages. Through a detailed analysis of the critical functions employed by mobile and fixed-line operators, we identify specific IP and mobile functions that demand heightened scrutiny to prevent cascading failures or failures that are difficult to troubleshoot. We present case studies illustrating the severe consequences of such failures on network services and their broader economic and corporate ramifications. Furthermore, we explore potential mitigation strategies and highlight the necessity for ongoing research to develop more robust network architectures that can withstand the inherent vulnerabilities of IP networks. This study aims to contribute to the enhancement of service resiliency in the face of evolving mobile and fixed network demands and complexities.

Team composition is a critical factor influencing collaboration within agile software development. This study investigates the impact of gender distribution on teamwork quality in software engineering capstone courses. We examined the experiences of 240 students organized into 40 teams during an agile project course. We analyzed two surveys, one conducted before team composition and one at the end of the project work. As much as 91% of the students chose to use the practice of conducting stand-up meetings in their project work and the majority were satisfied with the practice. Further, our analysis reveals that while women tend to engage more in design and men in programming, an increase in the proportion of women within a team correlates with a higher involvement of women in programming tasks. Our findings highlight gender differences in perceptions and experiences related to project involvement in agile software engineering courses.

The recently proposed short-time impedance spectroscopy (short-time IS) measures impedance spectra with high temporal resolution using nonsinusoidal oscillator signals. This study applied the short-time IS system to biological tissue impedance measurement during muscle activity. The system measured the temporal impedance spectrum of the human forearm and analyzed impedance changes during wrist motion. The IS system captured highly reproducible, frequency-dependent impedance changes when the wrist moved up and down. The forearm's impedance response during wrist motion was visualized with a high temporal resolution of about 0.5 ms, displaying the impedance magnitude and phase angle change as a color map. The short-time IS system's mode-switching mechanism enables rapid impedance measurement at 40 frequencies, forming smooth Cole-Cole plots with high temporal resolution. While muscle tissue impedance measurement is typically affected by skin tissue's high impedance, we found that difference and differential Cole-Cole plots reduced skin impedance effects and produced a characteristic curve corresponding to wrist motion, even with two-terminal measurement. We theoretically confirmed that the difference and differential Cole-Cole plots represent a reasonable muscle tissue impedance variation. This characteristic curve is compatible with machine learning techniques like pattern recognition. We anticipate that the IS system and visualization method will be beneficial for myoelectric control in prosthetic hands and noninvasive muscle disease diagnosis.

Deep neural networks (DNNs) have emerged as powerful predictors, demonstrating impressive performance across a spectrum of applications. They efficiently map high-dimensional input data to an output. However, a notable drawback is the lack of transparency and interpretability in DNN outputs. In critical domains such as healthcare, finance, public safety, and autonomous systems, transparency in DNN output is paramount for informed decision-making and trustworthiness. Recognizing and quantifying uncertainty in outputs can significantly improve decision-making capabilities. In crowd counting, where deep learning models are prevalent, uncertainty estimation is vital for robust decision-making in resource allocation, event planning, and public safety. However, existing work often overlooks the associated uncertainty in crowd count predictions, focusing solely on estimation. In this work, we propose a method to not only estimate the count but also the uncertainty in these estimated counts, considering both data uncertainty and model uncertainty. Furthermore, to enhance the calibration of model uncertainty, we utilize adaptive dropout rates within a Monte Carlo dropoutbased Bayesian framework. By utilizing isotonic regression, we calibrate overall predictive uncertainty, covering both data and model uncertainties. This leads to well-calibrated confidence intervals for the model predictions. Through extensive simulations on widely used datasets, our method sets a new benchmark.

A new synthesized magnetic focusing (SMF) technology for multiple coaxial circular coils is proposed in this paper. The fundamental principle of the proposed coaxial plane SMF is provided comparing with previous SMF as well as transcranial magnetic stimulation (TMS) in medical applications. The proposed SMF theory determines the appropriate currents in multiple coaxial circular coils, resulting in a focused magnetic field distribution across a three-dimensional space. Simulation results show that 4.2 times sharper magnetic field with a cost of reduced peak magnetic field. Experiments for five coils with a 50 cm diameter at a focal distance of 10 cm show very good agreement with simulation. A single-source-driver is firstly used replacing with multiple drivers of conventional SMF.

This article addresses whether the Lorentz force formula, as used in electrostatics and magnetostatics, is also applicable to non-relativistic electrodynamics. To answer this question, the article uses the Liénard-Wiechert potentials to calculate the general analytical solution of Maxwell's equations for arbitrarily moving, non-relativistic point charges. It is then shown that the obtained solution only produces the correct force of a direct current on a moving test charge under illogical assumptions. This problem does not arise from the Maxwell equations but is probably related to the fact that the Lorentz force is already contained in Maxwell's equations due to Maxwell's addition. The Lorentz force should therefore not be added a second time in the form of an explicit supplementary formula. Instead, it is reasonable to integrate an old hypothesis by Carl Friedrich Gauss from 1835 into Maxwell's electrodynamics. This integration turns Maxwell's electrodynamics into Weber-Maxwell electrodynamics and resolves the contradictions.

Electron cyclotron resonance (ECR) produced toroidal magnetized plasmas is investigated in a device named 'STARMA', an acronym for Simple Tight Aspect Ratio Machine Assembly, having major and minor radius of 10.4 cm and 6.8 cm respectively with aspect ratio (A) of ~ 1.5. The plasma is formed using a 6 kW magnetron based microwave source operating at 2.45 GHz, employing ECR techniques at a toroidal magnetic field of ~875G. The unique property of this device ensures a large variation of toroidal magnetic field (from 2 kG to 0.4 kG) across the poloidal cross section of the device and supports favorable magnetic fields of 875 G (for fundamental mode) and 437 G (for 2nd harmonic) for concurrent accommodation of fundamental and 2 nd harmonic resonance ECR breakdown, when microwave power at 2.45 GHz is used to form the plasma. In this scenario, the ECR wave-plasma interaction becomes more complex when the plasma production scheme includes simultaneously both, the fundamental and second harmonics of the ECR wave within the plasma volume. This novel technique of plasma production, involving both the fundamental and second harmonic of ECR waves, becomes relevant to small aspect ratio machines and therefore is studied in details. The contours generated from the measured plasma density suggests formation of slab geometry plasma in poloidal cross section of the machine and confirms elongated plasma in vertical direction. The radial profile at the equatorial plane suggests that plasma is formed with density (𝑛 𝑒)~ (1 − 7) × 10 16 𝑚 −3 and electron temperature (𝑇 𝑒) ~(3 − 30) 𝑒𝑉. The plasma is expected to exhibit complex behavior of mode conversion (O-X-B mode conversion), particularly when O and X-modes are launched from the low field side (LFS) at varied angles (𝜃) with respect to toroidal magnetic field (𝐵 𝜙). The subsequent conversion of X-mode into the Electron Bernstein Wave (EBW) at the upper hybrid resonance (UHR) emphasizes on mode conversion affecting parameters like density scale length (𝐿 𝑛 𝑒) and magnetic field scale length (𝐿 𝐵 𝜙) near the UHR. Therefore experiments, relevant to the mode conversion is also carried out by varying the toroidal launch angle up to 10 o. The Budden parameter (𝜂) for experimental conditions are estimated which serves as a milestone in the investigation to understand the OX -B mode conversion process. The verification of the OX mode conversion has been also done using the full wave simulation as WKB approximation is not applicable here. Experiments further suggest that the launching of the Xmode from the outboard side introduces large fluctuations and may be responsible for destabilizing the entire plasma production process.

Ferroelectric Superconducting Quantum Interference Device (Fe-SQUID) has recently emerged as a viable option to realize superconducting computing due to its voltage-controlled switching, which is essential to build large-scale digital circuits. This is the first work to model Fe-SQUID-based logic circuits and develop standard cell libraries compatible with existing EDA tool flows. We provide a comprehensive evaluation of the power consumption and performance of a wide range of Fe-SQUIDbased circuits, benchmarking them against the state-of-the-art 5 nm FinFET-based circuits. Our 5 nm FinFET transistor model is validated against industrial measurements. The validation is conducted not only at room temperature but also at extremely low temperatures, down to 10K, for fair comparisons against Fe-SQUID superconducting circuits. Our findings revealed that contrary to CMOS-based circuits, circuits realized using Fe-SQUID dissipate significantly more power. This presents a substantial challenge within the constraints of limited cooling power budgets in state-of-the-art cryostats. Open source: we will release our developed Fe-SQUID compact model, Fe-SQUID standard cell libraries, and Fe-SQUID-based circuits and make them publicly available in a github repository.

The pansharpening task has always been a focal point for researchers. However, existing methods have yet to fully exploit the modality correlation between Panchromatic (𝑃 𝐴𝑁) and Multispectral (𝑀𝑆) images, leading to the loss of spatial high-frequency information and spectral distortion during fusion. Additionally, they apply the same processing standards to information at different scales, causing these pieces of information to mix and interfere. To address these issues, we propose a multi-scale fusion network based on a continuous information guidance mechanism named CIGformer. We focus on improving the interaction mechanism between 𝑃 𝐴𝑁 and 𝑀𝑆. First, we introduce an Intensity Substitute Block (ISB) to decompose the shared and unique information of 𝑃 𝐴𝑁 and 𝑀𝑆, initializing the subsequent information guidance. Then, we construct an Information Guidance Block (IGB) based on the Transformer architecture, which is the core of our information guidance. It enables adaptive balanced retention of the unique information of 𝑃 𝐴𝑁 and 𝑀𝑆 without affecting the utilization of their shared information. Furthermore, we build a multi-level Encoder-Decoder bidirectional pyramid structure to mitigate the impact of multi-scale information intermixing. At each encoder level, we continuously use IGB to achieve optimal information utilization. Finally, we introduce a consistency loss function to evaluate the retention of unique information, assisting in the training process. Experiments on the GaoFen-2 and WorldView-3 datasets show that based on unique information guidance, our proposed interaction method significantly enhances the utilization efficiency of 𝑃𝐴𝑁 and 𝑀𝑆. Our code is available at https://github.com/Xidian-AIGroup190726/CIGformer.

The focus of this paper is to determine if a common operating system can satisfy all the basic requirements of IoT devices, taking into consideration the limited capability of IoT devices, high customization needs, and security requirements, and if a single operating system can achieve this or not.

We propose and experimentally demonstrate the design of a compact source for DAS systems using a mini-EYDFA commonly used in CATV networks together with an integrated, low-phase-noise Direct Digital Synthesis (DDS) device that can generate readily programmable probe waveforms with a bandwidth of up to 1.4 GHz. The DDS module is synchronized with a NI-PXIe system for real-time acquisition of traces at a rate of up to 200MS/s and, thanks to the low phase noise DDS characteristics as well as high gain and stability of the EYDFA and jitter-suppressed acquisition of traces, the scheme enables measurements of representative vibration signatures with a bandwidth of up to 4 kHz at a distance of 9.71 km, with an SNR of ~24 dB without trace averaging, offering a performance near the Nyquist limit set by round-trip-time of trace acquisition. Analyzes of the spatial, temporal, and spectral responses of extracted vibrations confirm the distributed dynamic sensing capability of the technique. The proposed configuration enables the simplification of sources used in DAS systems, paving the way toward further miniaturization of the interrogation units and their scalable commercialization for wider use in several safety and integrity monitoring applications.

Rank metrics quantify machine learning model prediction errors in the setting of binary classification with a given dataset, but imbalances in the class distribution or model predictions can significantly influence the interpretation. We analyze and interpret the area under the receiver operating characteristic curve (auROC) and an area under the precision recall curve, or average precision score (AP), in the presence of two kinds of imbalance: dataset class imbalance and model prediction imbalance. We show that the auROC can have a symmetric but not necessarily uni-modal response to dataset imbalance, but it is insensitive to model prediction imbalance. The AP has an asymmetric response that we show is also not uni-modal. The AP and auROC scores can be challenging to interpret due to these properties. We propose a novel precision recall space, the PR*, and associated average precision score, called AP*, to address the identified shortcomings of the AP and auROC. The proposed AP* metric has a symmetric response to class imbalance, like the auROC, a sensitivity to model prediction imbalance, like the AP, and it adds both a symmetry to model prediction imbalance and unimodal response to class imbalance that makes it more reliable for model selection than the AP or auROC. For model selection, the AP* gives a higher score to the model that tends to make most of its errors in the majority class.