Geopotential and orthometric height differences between distant points can be measured via timescale comparisons between atomic clocks. Modern optical atomic clocks have residual instabilities on the order of 10^-18}, allowing height differences of around 1 cm to be measured. Frequency transfer via free-space optical links will be needed for measurements where linking the clocks via optical fiber is not possible, but requires line of sight between the clock locations, which is not always practical due to local terrain or over long distances. We present an active optical terminal, phase stabilization system, and phase compensation processing method robust enough to enable optical frequency transfer via a flying drone, greatly increasing the flexibility of free-space optical clock comparisons. We demonstrate a residual instability of 2.5x10^-18 after 3 s of integration, corresponding to a height difference of 2.3 cm, suitable for applications in geodesy, geology, and fundamental physics experiments.

When equipped with a reliable calibration model, Low-Cost Sensor (LCS) can be relied upon as an effective option for gas concentration estimation, providing robust and high spatio-temporal resolution data to replace large-scale analytical instruments. In this paper, we present ProxySense, a rapid and efficient approach for gas concentration estimation. The ProxySense pipeline consists of gas sensing unit made up of array of metal oxide LCS, data pre-processing including an effective approach based on Variatioanl Autoencoders (VAE) for handling missing sensor data and Long Short Term Memory Reccurrent Neural Network (LSTM-RNN) prediction model. We investigate the capability of ProxySense in exploiting the deep characteristics that exist in multi-sensors’ responses to predict the concentration of a gas for which no specific sensor is included in a multi-sensor device. We evaluated ProxySense for benzene (C6H6) and carbon monoxide (CO) concentration predictions and compared the performances to multiple baselines by means of prediction error characterization. We further studied the relationship between model performance and training length and showed ProxySense to be highly accurate for gas concentration prediction even for small number of training period.

Technology, like phones, computers, and game systems, is part of everyone's daily life. But storing all our stuff can be hard. One solution is using computer resources when we need them, called cloud storage. It lets us access our stuff from any device. Amazon Web Services (AWS) is one place that offers this. It's safe and lets businesses make cool apps easily. And it's cheaper than buying a fancy storage device. Other companies like Google and Oracle also offer cloud hosting. We're writing about comparing these three big cloud services—AWS, Google, and Azure. We want to figure out which is best for people and companies. We'll look at what they all offer and what makes each one special.

This paper proposes a novel fluxible spoked Mecanum wheel structure and a mobile robot based on this wheel, and introduces in detail the motion mode and corresponding control methods of the fluxible spoked Mecanum wheel in the omni-directional movement mode and crawl mode of the mobile robot.

We use the dynamic Green’s function to produce a frequency-dependent magnetic vector potential $\vec{A}(\omega)$ and derive expressions for the efficient (accurate and fast) computation of cylindrical components of the magnetic flux density vector $\vec{B}(\omega)$ as a function of the solenoid’s geometric and material parameters. $\vec{A}(\omega)$ may be used to efficiently compute the frequency-dependent flux linkage $\Phi(\omega)$, the complex inductance $L(\omega)$, and the radiation patterns of the solenoid anywhere in space including both near-field and far-field regions, excluding the (source) regions of conducting wire. Additionally, we propose the complex calibration coefficient $\chi(\omega)$ to account for the finite-radius conductor. Several numerical examples are provided to validate the proposed helical model against the superposition of circular loops. The proposed model is demonstrated for a wide range of applications across the spectrum from 60 Hz to 170 GHz, representing low-frequency power systems to high-frequency mm-wave communication systems. A plan is being developed for experimental validation of the model.

Change detection and diagnosis based on information-based features is considered. As benchmark case study, the faults of bearings are selected. Information from the working states, including those generated by faults in bearings, are transformed and carried by the vibration's signals, which are further processed by advanced techniques of information and signal processing, as e.g., statistical models, Renyi and Tsallis entropies, and other complexity measures.

Flying networks, composed of Unmanned Aerial Vehicles (UAVs) acting as mobile Base Stations and Access Points, have emerged to provide on-demand wireless connectivity, especially due to their positioning capability. Still, existing solutions are focused on improving aggregate network performance using a best-effort approach. This may compromise the use of multiple services with different performance requirements. Network slicing has emerged in 5G networks to address the problem, allowing to meet different Quality of Service (QoS) levels on top of a shared physical network infrastructure. However, Mobile Network Operators typically use fixed Base Stations to satisfy the requirements of different network slices, which may not be feasible due to limited resources and the dynamism of some scenarios. We propose an algorithm for enabling the joint placement and allocation of communications resources in Slicing-aware Flying Access and Backhaul networks – SurFABle. SurFABle allows the computation of the amount of communications resources needed, namely the number of UAVs acting as Flying Access Points and Flying Gateways, and their placement. The performance evaluation carried out by means of ns-3 simulations and an experimental testbed shows that SurFABle makes it possible to meet heterogeneous QoS levels of multiple network slices using the minimum number of UAVs.

The exponential growth of textual data across various domains necessitates the development of efficient and accurate summarization techniques to facilitate quick comprehension and information retrieval. The novel automated system for summarizing multiple document abstracts and titles using advanced neural architectures addresses this need by leveraging large language models to generate concise and coherent summaries. The methodology involved comprehensive data collection, preprocessing, model selection, and summarization processes, evaluated through a combination of quantitative and qualitative metrics. Results demonstrated high efficacy in handling shorter documents and strong performance in technical domains such as healthcare and science, although challenges in coherence and readability were noted. Domain-specific performance highlighted the necessity for tailored adaptations, and the study contributed valuable insights into hybrid summarization techniques combining extractive and abstractive methods. Future research directions include the development of advanced attention mechanisms, domain-specific fine-tuning, and reinforcement learning techniques to optimize summarization quality, alongside addressing ethical considerations to ensure responsible deployment.

Cellular Vehicle-to-Everything (C-V2X) has become one of the highest anticipated technologies for vehicular safety network. In LTE C-V2X Basic Safety Messages (BSMs) are transmitted on radio resources that are allocated with a periodic resource reusability. This allocation is based on a semi persistent sensing-based scheduling scheme (SPS) algorithm. But, due to this reuse of periodic resources, the possibility of loss of consecutive packets between the same vehicle pair is significant.  This study dis-cusses different approaches proposed to solve this consecutive loss problem. Based on this investigation, this article suggests an efficient One-Shot based solution with a new control parameter, that performs superior to the state-of-the-art solution that is standardized in SAE J3161/1 which this article analyzes and shows to have limitation in case of high-density scenario.

Improving Cataract Surgery Procedure using Machine Learning and Thick Data Analysis

Federal Energy Regulatory Commission (FERC) Orders 841 and 2222 have recommended that distributed energy resources (DERs) should participate in energy and reserve markets; therefore, a mechanism needs to be developed to facilitate DERs’ participation at the distribution level. Although the available reserve from a single distribution system may not be sufficient for tertiary frequency regulation, stacked and coordinated contributions from several distribution systems can enable them participate in tertiary frequency regulation at scale. This paper proposes a deep reinforcement learning (DRL)- based approach for optimization of requested aggregated reserves by system operators among the clusters of DERs. The cooptimization of cost of reserve, distribution network loss, and voltage regulation of the feeders are considered while optimizing the reserves among participating DERs. The proposed framework adopts deep deterministic policy gradient (DDPG), which is an algorithm based on an actor-critic method. The effectiveness of the proposed method for allocating reserves among DERs is demonstrated through case studies on a modified IEEE 34-node distribution system.

The penetration of weather dependent renewable energy sources which are highly stochastic in nature create new challenges with system security, reliability, flexibility, and sustainability. This research focuses on the development of an artificial intelligence-based control method for a laboratory-scale hardware-in-the-loop microgrid setup. This setup features a variety of loads, battery banks, protection relays, renewable and nonrenewable sources of energy, and advanced metering devices interfaced through standard communication protocols. A resilient, cost-effective distributed control strategy allows flexible and reliable autonomous operation and control of the microgrid. Thus, a deep reinforcement learning based multi-agent system is the primary focus for grid control in this research. These intelligent agents learn from the microgrid environment and take actions to maximize their cumulative rewards based on prior experiences. The agents receive rewards for taking good actions and are penalized for bad actions. The results obtained in this research prove the feasibility of this approach.

This paper presents a design methodology and prototyping of an angular speed sensor based on microwave polarimetry. The detection mechanism relies on the transmission response of a dual-polarized circular aperture irradiating an array of regularly spaced copper traces. The generated response is a regular pulse train defined by the location of a copper trace with reference to the center of the circular aperture. For speed measurement, the response was acquired within a 10s duration with the circular aperture response effectively sampled at 100 Hz. The acquired response was analyzed in the frequency domain using Fast Fourier Transform (FFT) and the rotor speed was obtained from the frequency spectrum of the acquired pulse train. The detection system can detect speeds as low as 0.4 RPM with an upper limit of 5000 RPM limited by the maximum sampling rate of the acquisition system. With the adoption of the FFT technique for speed estimation, the proposed angular speed measurement system can instantaneously detect angular speed with a maximum absolute estimation error of 0.0625 RPM within the 10 s acquisition time. The rotor occupies an area of 84.64 cm2.

This paper describes the historical trajectory of video conferencing systems, spanning from their earlier mechanical and analog origins in the 1920s to the sophisticated IP-based services delivered from the cloud in the 2020s. Each technological age is examined, highlighting the technical and functional aspects that characterized its evolution. Commercial landmarks of each age are presented providing a comprehensive overview of the most prominent offerings at pivotal moments in the timeline. By examining the past and speculating on the future, this paper aims to provide a holistic understanding of the development, current state, and forthcoming trends in video conferencing technology.

This paper investigates the occurrence of the double descent phenomenon within the domain of federated learning. It derives a closed-form solution to calculate the mean excess risk in the coefficients of a linear regression model, with the theoretical findings verified through simulations. The results confirm the presence of the double descent effect in the case of federated learning. This effect is especially pronounced in a heterogeneous scenario where local datasets of collaborating users have different sizes and properties. Furthermore, the results unveil unexpected and non-trivial performance behaviors. Interestingly, federated learning even outperforms centralized learning in particular scenarios when the expected excess risk of the former is lower than that of the latter. Finally, our analysis reveals that the typically used random user selection in federated learning exhibits asymptotically lower performance in contrast to the case of excluding only a single carefully chosen user from the learning process.

We apply machine learning (ML) techniques to identify the modes in rectangular waveguides from images of 2D modal field patterns injected with uniform, exponential, correlated exponential, and Gaussian noise distributions. A binary classifier is used to identify either transverse electric (TE) or transverse magnetic (TM) modes, and a Multi-class classifier is used to identify the mode numbers. Signal to noise ratios of 1, 0.1, and 0.01 are used to show the effectiveness of each model. Results show accuracy scores up to 99.95%. Several examples demonstrate that noisy modal patterns (unidentifiable to human eyes) may be successfully classified by the ML model.

The integration of renewable generation in the energy mix has led to a massive penetration of converter-interfaced generators (CIG) in the electrical network. However, CIG’s are reported to be susceptible to interactions with electrical resonances and  other equipment. It became common practice to conduct electromagnetic transient (EMT) studies to assess these risks, define the mitigation measures and design the control systems. To perform the controllers design efficiently, the EMT analysis can be complemented with an analysis of linearized systems, for which numerous efficient methods have been developed. The paper presents a method to construct the linearized system which is believed to be sufficiently accurate for being used for controllers tuning, and sufficiently simple to be efficiently used on real-life cases. It takes into account the frequency dependent electrical network, and is based on the approach previously used for torsional analyses, but applied to a wider spectrum. It is illustrated on a simple numerical example, demonstrating the challenges that the electrical resonances might pose on the stability margins. A comparison of the results obtained from linearized model and EMT model is presented, aiming to validate the findings from the linearized model and the proposed approach.

This research paper aims to address the development of a cost-effective Data Platform tailored for analytics, data science, and AI applications. While the focus of this paper lies predominantly on the retail industry, the approach presented can be applicable across various domains. The advent of cloud technology has significantly benefited small and mid-sized organizations, leading to discussions surrounding IT modernization in recent years. Within the context of this study, we concentrate specifically on small and mid-sized retail organizations that utilize ERP, allocation, Warehouse Management, and Sell applications, both in-store and online. The Information Technology (IT) sector strives to establish a scalable data platform that caters to reporting, data science, and AI initiatives. This research paper places special emphasis on mid-sized organizations, which often face budget constraints and stringent project timelines. Establishing an IT organization with limited resources and budget necessitates careful consideration of the available skill set. By addressing the challenges and opportunities associated with building a cost-effective Data Platform, this research paper aims to provide valuable insights and practical guidance for organizations operating within similar contexts. The subsequent sections will delve into the methodologies, technologies, and strategies employed to achieve a scalable and efficient data platform that aligns with the specific requirements of small and mid-sized retail organizations. More detail will be found at attached PDF.