Endoscopic Lumbar Discectomy (ELD) is a minimally invasive surgical procedure to remove the herniated material of a disc in the lumbar spine [1]. In the traditional approach, the surgeon holds the endoscope and inserts the instruments into its lumen to reach the surgical window and perform the discectomy. Some steps, such as opening of the ligamentum flavum, require the help of an assistant to hold the endoscope while the surgeon is occupied handling other tools: this operation results in an inevitable waste of time. To overcome this and other drawbacks of the traditional procedure, a robotic approach has been investigated, which involves the use of a collaborative robot to hold the endoscope throughout the procedure. In this context, the gripper and its actuation system must be safe, reliable and lightweight. The following paper describes a novel mechanism for manual actuation of the gripper responsible for holding the endoscope, specifically designed to allow the surgeon to have full control over the gripping force applied.

In this study, we investigate the optical parameters of plasmonic Titanium Nitride (TiN) thin films using the RefFIT software, employing the Drude-Lorentz model for data fitting. TiN is an emerging material for plasmonic applications due to its high melting point, wear resistance, and stability at various temperatures, positioning it as a viable alternative to traditional noble metals. Our research focuses on fabricating, characterizing and extracting optical and electrical properties of TiN thin films deposited on silicon wafers using ellipsometry and Fourier Transform Infrared (FTIR) spectroscopy.We analyze six different samples, varying in thickness and composition, to determine key parameters such as plasma frequency and damping constants. The simultaneous fitting of reflectance and ellipsometry data in RefFIT provided results consistent with existing literature, establishing the software’s reliability for such analysis. Additionally, we validated the RefFIT results by comparing them with alternative models and experimental data, demonstrating the necessity of incorporating Lorentz terms alongside Drude components for accurate optical characterization.Our findings reveal that TiN films exhibit desirable plasmonic properties, with carrier concentration and relaxation times aligning with known trends in nitrides. The robustness of RefFIT in modeling these properties confirms its utility for future research in plasmonic materials. This study underscores the potential of TiN in advanced plasmonic devices and encourages further exploration using comprehensive modeling techniques.

In the era of digital healthcare, the exchange of medical data is crucial for enhancing patient care and advancing medical research. However, traditional data exchange systems often face challenges such as security breaches, data silos, and lack of interoperability. This paper proposes a blockchainbased solution for revolutionizing medical data exchange by creating a secure and decentralized marketplace. Leveraging blockchain technology ensures data integrity, enhances privacy, and facilitates seamless data sharing among stakeholders. Our approach includes smart contracts for automating transactions and ensuring compliance with regulatory standards. The proposed marketplace fosters trust, reduces costs, and promotes innovation in healthcare by enabling efficient and secure data exchange. This paper also discusses the implementation details, potential challenges, and future directions for blockchain in medical data management.

Researchers have sought solutions to interconnect different radio access technologies, enabling them to coexist in the current and future heterogeneous network environments to obtain additional benefits. This paper presents SVORA, an innovative approach that integrates the concept of Virtualised/Open (V/O)-Radio Access Network (RAN) with a Service-Based Architecture (SBA) in order to optimize the energy consumed by the network and minimize the communication delay. In the context of SVORA, the Delay-aware Energy Efficiencybased RAN selection algorithm (DEER) is proposed. DEER employs utility functions and considers Quality of Service (QoS) metrics such as throughput, resource block utilization, delay, and energy consumption measurements when finding the RAN to best support the requested service. DEER is highly flexible, being able to be tuned to support critical delay services, prioritizes energy saving, or observe a balanced approach. Testing shows that the SVORA-based DEER was able to increase the number of users and the aggregate data rate by about 10%, whereas the delay and energy consumption decreased by approximately 90% and 50%, respectively.

Automatic building segmentation from high resolution remote sensing images is currently a research hotspot. In recent years, deep learning, particularly CNNs (convolutional neural networks), has shown promising results. However, the state-of-the-art models currently available still struggle with boundaries, sharp edges, and complex geometry shapes. In order to address these challenges, we propose Boundary and Edge Aware U-Net (BEAU-Net), a boundary and edge attention network that is based on UNet. BAM focuses on the features with rich boundary information. It enhances the models' capability to differentiate between buildings and their surroundings, specially useful in urban regions where buildings are concentrated in proximity. EAM emphasizes refined features such as sharper edges and complex geometric shapes. It enables the model to capture complex geometrical building architectures and detect abrupt changes in edge details. We have conducted extensive experiments on the Massachusetts Building, WHU Building, and INRIA Aerial Image Labeling Dataset to validate the effectiveness of our framework. The results demonstrate significant improvement in the IoU and F1 scores, especially where narrow boundaries and complex shapes are present.

Probabilistic graphical models are powerful frameworks that represent and reason through uncertainty in complex domains. The research addresses the challenges of representation, inference, and learning of PGMs within the settings of high-dimensional and incomplete data. We will be using both directed and undirected models in capturing these complex relationships between variables and make use of sophisticated inference techniques-for instance, belief propagation and Markov Chain Monte Carlo methods-to improve computational efficiency. Results of this paper prove that the integration of prior knowledge and the use of structured conditional probability distributions significantly improve model accuracy and robustness. We further apply PGMs to problems of medical diagnosis and image segmentation to portray their prowess in supporting informed decisions under uncertainty. The results indicate that, apart from helping in a better comprehension of complex systems by providing a versatile tool for knowledge discovery and predictive modelling, PGMs also support making informed decisions under uncertainty. Ultimately, the research offers continuous developments in PGMs, particularly insight into their practical applications and potentials for improvement in artificial intelligence and machine learning.

The work describes the creation of the original model of the RPD Module, a component of the Remote PHY Node. This model incorporates GCP/R-DEPI/R-UEPI integration, which has been implemented using cutting-edge packet processing technologies for the enhanced data transmission between CCAP Core and the cable modem via DOCSIS 4.0. The model is open-source and, therefore, does not impose limitations on various applications and improvements for hybrid fiber-coaxial networks. The developed model enhances the performance of processing data packets and improves the reliability of next-generation broadband cable telecommunication networks.

This paper presents a cost-effective and scalable hybrid methodology for evaluating retrieval-augmented generation (RAG) systems using specialized pretrained models and advanced metrics, designed for critical domains like healthcare and finance. LLM judge-based evaluation approaches are hindered by significant score inconsistencies across identical input runs (up to 35% variations) and high computational costs while traditional NLP approaches, overly reliant on entity or phrase matching, lack a multi-faceted perspective and fail to capture deeper semantic understanding. Our methodology addresses these challenges with a novel approach that combines an ensemble of fine-tuned pretrained models and advanced NLP techniques, ensuring consistent, reproducible evaluations across multiple dimensions, while being tailored for offline scenarios to eliminate reliance on internet-connected systems or proprietary LLMs, offering a cost-effective solution with high accuracy in assessing semantic relevance, factual correctness, and context adherence. To enhance reliability, the approach employs a robust weighted scoring system using harmonic means combined with PCA, adaptive, and entropy-weighting techniques for trustworthy and consistent evaluation enabling seamless integration with existing systems, continuous metric adaptation and domain-specific customization, ideal for high-stakes applications that demand rigorous quality assessments without relying on external APIs.

Stress tracing is an important research domain that supports many applications, such as health care and stress management; and its closest related works are derived from stress detection. However, these existing works cannot well address two important challenges facing stress detection. First, most of these studies involve asking users to wear physiological sensors to detect their stress states, which has a negative impact on the user experience. Second, these studies have failed to effectively utilize multimodal physiological signals, which results in less satisfactory detection results. This paper formally defines the stress tracing problem, which emphasizes the continuous detection of human stress states. A novel deep stress tracing method, named DST, is presented. Note that DST proposes tracing human stress based on physiological signals collected by a noncontact ultrawideband radar, which is more friendly to users when collecting their physiological signals. In DST, a signal extraction module is carefully designed at first to robustly extract multimodal physiological signals from the raw RF data of the radar, even in the presence of body movement. Afterward, a multimodal fusion module is proposed in DST to ensure that the extracted multimodal physiological signals can be effectively fused and utilized. Extensive experiments are conducted on three real-world datasets, including one self-collected dataset and two publicity datasets. Experimental results show that the proposed DST method significantly outperforms all the baselines in terms of tracing human stress states. On average, DST averagely provides a 6.31% increase in detection accuracy on all datasets, compared with the best baselines.

Generative Adversarial Networks (GANs) are increasingly becoming essential tools for improving the perception of autonomous vehicles in adverse environmental conditions, like low light, fog, rain, and snow. This survey discusses recent trends in applying GANs to tackle these problems and classifies the available work based on environmental conditions and GAN architectures. Studies showcase the impact of different GAN models such as CycleGAN, Super-Resolution GANs, and DeblurGAN in enhancing image quality, visibility, and detection accuracy of objects under challenging conditions. For instance, CycleGAN has been remarkably effective in nighttime image enhancement; DeblurGANs significantly reduce motion blur at capture. Furthermore, multi-modal fusion approaches have been applied very successfully with RGB and infrared (IR) data creating strong perception systems that work well in diverse operational conditions. However, practically deployed driving solutions must process information quickly while remaining computationally efficient. This survey emphasizes how GAN technology has made autonomous vehicle research more adaptable and safer; it also points out potential graduate study directions involving lighter-weight models of GANs or domain-adaptive architectures to make them more applicable. We hope our comprehensive review will help researchers utilize GAN technology to enhance perception for autonomous vehicles.

An adaptive, a variable code rate,can respond to channel changes promptly channel coding method is proposed,a discrete memoryless binary source sequence 𝑿 is transformed losslessly to make it capable of error checking. After the transformation, 𝑿 has variable and fixed code rates through weighted probability arithmetic coding. Compared with a fixed code rate, a variable code rate can respond to channel changes promptly and effectively improve the transmission performance, thereby being suitable for more complex communication environments. We prove that when the code length is long enough, the decoding error probability of the weighted probability arithmetic coding is 0 and the encoding rate can reach 1. When the symbols in the sequence 𝑿 are uniformly distributed, the transmission rate can reach the channel capacity. In an additive-white-Gaussian-noise channel binary-phase-shiftkeying signal simulation experiment, the error correction performances under various fixed code rates were compared. When the fixed code rate was 0.5 and the block error rate (BLER) was 10 −3 the weighted probability arithmetic coding improved by 0.37 dB compared to that obtained using Polar CA-SCL, and when BLER was 10 −5 , it improved by 0.48 dB. The algorithm can be widely used in communication systems, data storage, and other fields.

The proliferation of small cells (SCs) is ubiquitous in boosting system capacity, coverage, and quality of services (QoS) for smart applications. However, due to the dynamic users and constrained resources, SCs cannot afford to have a large quantization codebook, which is mostly suitable for the macro cells (MCs) with finite rate feedback (FRF)-based multiple input single output (MISO) systems. This paper proposes a per-receiver antenna-based adaptive quantization approach for SCs where the codebook size changes w.r.t. the number of receiver antennas. We further address the code quantization error (CQE), which could occur due to the quantization of two unique channels with the same code and average system error (AvgSysErr), which could increase due to the rise in the CQE. We show that the convergence of AvgSysErr with FRF-based MISO systems requires SCs to have the non-unique code probability of the quantization codebook less than 1 N , where N is the number of antennas at the transmitter. Similarly, the lower-bound for the non-unique code probability needs to be less than or equal to ε, where ε is the difference between the non-unique code probability of 1 N 1 and 1 N 2 , given 1 N 1 > 1 N 2 (where N1 and N2 are the number of antennas at the transmitter). Additionally, interference from the adjacent SCs is inevitable due to their dense deployments satisfying the low signal-to-interference-plus-noise ratio (SINR) regime. We quantify the minimum interference constraint (IC) for SCs, which is critical for the multiplexing gain in the low SINR regime. We use the iterative steepest ascent (SA) method and quantify ICs for SCs. With the proof of convergence and the comparative analysis of the ideal and real cases, supported with our simulated results, we show that the expected value of IC is the best IC to be considered for SCs.

The integration of increasing amounts of volatile renewable energy sources leads to a growing amount of uncertainty. In addition, the number of small-scale renewable units grows rapidly, and the resulting decentralization of supply and demand patterns affects especially the underlying electricity networks. Consequently, reliable and performative forecasts are required, especially for network operators, to cope with the stochastic nature of highly decentralized electricity systems. Therefore, this paper addresses the challenge of selecting among a set of competing forecasting models to predict congestions in power systems with high shares of distributed and uncertain generation. Here, an event-based evaluation framework is applied to different, inter alia, copula-based forecasting models, and compared to other probabilistic metrics, with regard to the resulting congestions. To demonstrate the applicability of this event-based approach, empirical data for a high-voltage distribution network in Germany is used together with intra-day weather forecasts. Furthermore, the different forecasting models are benchmarked to each other with regard to their computational performance.

A document by Xiucheng Wang. Click on the document to view its contents.

This review explores the potential of telepsychiatry (TP) and artificial intelligence (AI) in addressing mental health issues in India, showing the opportunities and challenges of implementing digital health solutions to address the indirect effects of the COVID-19 pandemic on mental health. The paper analyzes the system of VOS viewer, co-citation maps, and keyword hotspots. 2019 led to the widespread introduction of lockdowns and social separation as public health measures. These actions, however, also carry hazards to mental health, such as a rise in loneliness and a decrease in access to psychiatric care. Methods to close this gap include telepsychiatry (TP) and other advanced technologies that enable access to care and remote mental health support. The combination of technology such as virtual reality, artificial intelligence, and mobile apps with TP has great promise for improving mental health treatments. Using these digital technologies could transform the way mental healthcare is delivered and provide improved support for patients, families, and healthcare providers as the world works through the COVID-19 pandemic’s issues.

This study presents an approach to developing a model for classifying abnormalities in video capsule endoscopy (VCE) frames. Given the challenges of data imbalance, we implemented a tiered augmentation strategy using the albumentations library to enhance minority class representation. Additionally, we addressed learning complexities by progressively structuring training tasks, allowing the model to differentiate between normal and abnormal cases and then gradually adding more specific classes based on data availability. Our pipeline, developed in PyTorch, employs a flexible architecture enabling seamless adjustments to classification complexity. We tested our approach using ResNet50 and a custom ViT-CNN hybrid model, with training conducted on the Kaggle platform. This work demonstrates a scalable approach to abnormality classification in VCE.

In this paper, we propose a thorough analysis of an advanced binary tree algorithm with multipacket reception (MPR) capabilities, along with timeliness metrics such as age of information (AoI) and packet loss probability (PLP) in a gated access setup scenario. We consider a finite number of users with Bernoulli arrivals and leverage packet preemption capabilities in their buffers, highlighting a novel and significant aspect. Initially, we explore the expected length of the collision resolution interval, conditioned on the anticipated number of users who have arrived beforehand. Notably, the preemption of user packets occurs both inside and outside of the corresponding collision resolution intervals. Subsequently, we illustrate the state transitions of contending users, conditioned on the MPR scenarios. Due to the capability of employing MPR, the receiver can successfully decode all the packets in an error-free channel. Furthermore, we demonstrate that, in gated access, the normalized throughput increases with an increase in the upper bound of MPR capabilities. Nonetheless, delay, PLP, and AoI exhibit superior performance in the unnormalized case. This demonstrates that enhancing physical resource capabilities in the unnormalized case leads to improved performance in a random-access channel when using binary tree algorithms.

The IETF Manufacturer Usage Description (MUD) standard seeks to enhance IoT security by restricting device communication to specified network flows. However, while MUD profiles define permitted flow destinations, they lack visibility into the content exchanged, leaving devices vulnerable to unintended payloads. This paper introduces a systematic approach to categorize MUD flows by their payload types-transparent, encoded, or encrypted-providing insights crucial for cyber-risk assessment, anomaly detection, and selective packet inspection. Our contributions are threefold: (1) We collect a dataset consisting of over 123000 network flows involving 24 different protocols, label each record with a ground-truth label of payload type, and analyze them to identify 17 attributes predictive of three classes-we make our data publicly available; (2) We train and evaluate a baseline classifier with over 90% accuracy for detecting flow payload types, and show encoded and encrypted flows demand higher confidence scores to pass a practical threshold 0.8; (3) We demonstrate how detailed payload characterization in representative MUD flows significantly improves security profiling for IoT devices at computing costs higher than our baseline. We develop an inference pipeline with three specialized classifiers that are dynamically and progressively applied to the payload of input MUD flows. Our evaluation results show that highly (more than 99%) accurate and confident inferences can be made across three payload classes by spending 44% more computing resources.