In this study, a robust method for 3D pose estimation of immature green apples (fruitlets) in commercial orchards was developed, utilizing the YOLO11 object pose detection model alongside Vision Transformers (ViT) for depth estimation. For object detection and pose estimation, performance comparisons of YOLO11 (YOLO11n, YOLO11s, YOLO11m, YOLO11l and YOLO11x) and YOLOv8 (YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l and YOLOv8x) were made under identical hyperparameter settings among the all configurations. Likewise, for RGB to RGB-D mapping, Dense Prediction Transformer (DPT) and Depth Anything V2 were investigated. It was observed that YOLO11n surpassed all configurations of YOLO11 and YOLOv8 in terms of box precision and pose precision, achieving scores of 0.91 and 0.915, respectively. Conversely, YOLOv8n exhibited the highest box and pose recall scores of 0.905 and 0.925, respectively. Regarding the mean average precision at 50% intersection over union (mAP@50), YOLO11s led all configurations with a box mAP@50 score of 0.94, while YOLOv8n achieved the highest pose mAP@50 score of 0.96. In terms of image processing speed, YOLO11n outperformed all configurations with an impressive inference speed of 2.7 ms, significantly faster than the quickest YOLOv8 configuration, YOLOv8n, which processed images in 7.8 ms. This demonstrates a substantial improvement in inference speed over previous iterations, particularly evident when comparing YOLO11n and YOLOv8n. Subsequent integration of ViTs for the green fruit's pose depth estimation revealed that Depth Anything V2 outperformed Dense Prediction Transformer in 3D pose length validation, achieving the lowest Root Mean Square Error (RMSE) of 1.52 and Mean Absolute Error (MAE) of 1.28, demonstrating exceptional precision in estimating immature green fruit lengths. Following this, the DPT showed notable accuracy improvements with a RMSE of 3.29 and an MAE of 2.62. In contrast, measurements derived from Intel RealSense point clouds exhibited the highest discrepancies from the ground truth, with a RMSE of 9.98 and an MAE of 7.74. These findings emphasize the effectiveness of YOLO11 in detecting and estimating the pose of immature green fruits, illustrating how Vision Transformers like Depth Anything V2 adeptly convert RGB images into RGB-D data, thus enhancing the precision and computational requirement of 3D pose estimations for future robotic thinning applications in commercial orchards.

The simultaneous progress of integrated optical frequency comb (OFC) and radio frequency (RF) photonic signal processing technique have promoted the rapid development of real-time signal processing. Integrated optical frequency comb offer multiple wavelengths as a powerful source for RF photonic signal transversal filter. Here, we review development of real-time signal processing system consisting of integrated OFC and RF photonic signal transversal filter in chronological order, and focus on the applications of this system such as differentiator, integrator, Hilbert transformer, and image processor. We also discuss and present our outlook on more parallel functions and further integration of real-time signal processing system.

How to build the arithmetic finding 3-uncolorable vertices in polynomial time in 3-coloring is studied. For planar graph G[V,E] the diagrammatic arithmetic computing jammed configurationwith the independent clip (DJC) is built in the color space according to the Grötzseh theorem. Each DJC corresponds a triplet in G[V,E] , which is the invariant set of three sets of monochromatic vertices with different colors under coloring transformation. The algorithmcancomplete the computation to found all DJC in 8 O(V ) . The vertex adjoins vertices of three monochromatic sets is 3-uncolorable. They are only the part of 3-uncolorable vertices. So, 3-coloring has 3-uncolorable vertices can be recognized in polynomial time.

This paper probes the validity up to 15 kHz of the analytical formulae published in the CIGRE Technical Brochure 908 related to the calculation of the armour losses in singlearmoured AC cables, which have been tested in the Brochure at 50 Hz only. After briefly reviewing the theoretical background, and carefully explaining the approximations performed by the Brochure, it is shown that for a typical cable the analytical formulae are reasonably accurate in the estimation of losses up to 15 kHz, but the absolute error increases linearly with frequency.

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.

In this study, the method used involves segmenting images into fixed-sized square regions called superpixels and then analyzing features of each superpixels to determine whether it belongs to the foreground or background. The study also discusses the process of selecting relevant features and testing different combinations. These features include characteristics of the superpixel itself—its local relationships with neighboring superpixels, and global features of the entire image. Importantly, it emphasizes making the decision-making process understandable—achieved by using a decision tree model for analyzing the collected features. This study focuses on teaching a self-driving system how to accurately separate foreground (movable areas like floors) from background (non-movable areas) in images captured inside buildings. This approach aims to strike a balance between entirely automated segmentation models like Convolutional Neural Networks (CNNs)—which lack explainability, and purely knowledge-based segmentation models, where all information is predefined. This information is stored in a database for analysis. Specifically, it addresses the scenario of an autonomous vehicle navigating public spaces without endangering people.

In this work, a high-bandwidth very-low input capacitance original topology buffer was designed. The proposed buffer circuitry is conceived as part of a novel non-invasive non-immersed real-time dielectric spectroscopy biomass sensor. Both a circuital-based simulation model and an experimental PCB prototype were implemented and successfully tested. The buffer effective parasitic input capacitance was lowered to a 2% of its original value by means of power supply bootstrapping. This improvement will allow performing differential voltage measurements with the required accuracy and without compromising the necessary bandwidth.

In this letter, to handle the over-generation issue of photovoltaic (PV) systems based on maximal power point tracking (MPPT) algorithms, and to provide the power reservation for power grids, an optimal point shifting design based on virtual impedance is proposed for PV systems with flexible power point tracking. With the introduction of the virtual impedance, the virtual PV power is provided to auto-shift optimal power points on PV power-voltage curves. With a MPPT algorithm applied to the virtual PV power, the actual PV power is auto-turned without the estimation of maximum available PV power. The proposed design is validated through simulation studies of a PV-battery charging system and experimental studies of a grid-tied PV system, where both the overcharging of the battery and the over PV power generation are effectively avoided by reducing actual PV power according to the increasing of the battery SOC or the grid RMS voltage, respectively.

The rapid integration of the Internet of Things (IoT) into high-end equipment has dramatically transformed fault diagnosis processes, enabling real-time monitoring and predictive maintenance. This technological advancement, while enhancing operational efficiency, has also introduced critical security vulnerabilities, particularly within supply chains susceptible to cyberattacks. This paper investigates the intersection of IoT security and fault diagnosis in high-end equipment, proposing a systematic secure software engineering framework. Through an extensive literature review and a detailed analysis of existing datasets on supply chain attacks, an innovative secure software engineering approach is developed. The methodology is validated through a case study on IoT-Enabled Fault Diagnosis in a High-End Car Manufacturing Supply Chain, showcasing the practical application of the proposed framework. Business Process Modelling and Notation (BPMN) simulations of the car manufacturing process reveal that 1,000 cars can be assembled and repaired in 467.33 minutes (approximately 7.79 hours or onethird of a day) using this approach. Additionally, the study introduces key metrics for process quality, data quality, and predictive model accuracy, offering a quantitative assessment of the framework's effectiveness. The findings emphasize the critical need for robust security measures in IoT-enabled fault diagnosis systems to mitigate potential risks and ensure operational integrity.

Digitalization is an important part of Industry 4.0 in which data sharing is essential to exploit the full potential of data and enable the use of modern technologies like machine learning or predictive maintenance. In addition to trust, other boundary conditions, e.g., incentives, must be taken into account. This work investigates, compares and groups different approaches that can be used to increase data sharing. Designing a data market, which depends on the respective use case, appears to be a suitable high level solution for this. The paper points out that technological approaches to realize a trustful data market can be grouped in data spaces and data intermediaries as superordinated categories. Afterwards, different combinations in order to build a centralized, decentralized or hybrid trustful data market are shown. Finally, initial abstract design decision rules are derived.

Ransomware continues to pose a significant threat to the security of Windows systems, as its rapidly evolving nature enables attackers to bypass traditional signature-based detection methods. The use of opcode-based features, combined with K-Nearest Neighbors (KNN) classification, offers a novel and effective approach to identifying ransomware by analyzing the low-level behavior of executables rather than relying solely on external characteristics. Through extracting opcode sequences from both ransomware and benign Windows executables, and converting them into high-dimensional feature vectors, the method captures behavioral distinctions that facilitate accurate classification. The study demonstrates that feature selection techniques, such as Chi-Square and Information Gain, not only improve computational efficiency but also maintain the model's detection accuracy, significantly enhancing the practicality of deploying machine learning models in real-world cybersecurity environments. Comparative evaluations show that KNN performs competitively with other machine learning approaches, making it a suitable choice for detecting ransomware while balancing efficiency and accuracy. The results emphasize the importance of opcode-based analysis in providing deeper insights into executable behavior, offering robust defense mechanisms against ransomware attacks.

Stochastic Computing has turned out to be a very suitable hardware design approach for Morphological Neural Networks (MNNs), given its easiness in implementing maxima, minima and products through simple logic gates that perform bitwise operations. This work enhances a previous MNN design by using approximate binary adders, reducing the required hardware resources of adders which are the most area-consuming part of the MNN. The experimental results show a minimal decrease in accuracy, a reduction in circuit area and increases in speed and energy efficiency. The proposed approach has been validated on a 180nm UMC CMOS chip for the MNIST experiment, showing highly competitive performance and energy efficiency values with respect other works present in the literature.

This study introduces a novel transformer architecture that integrates Gaussian Attention, adapter modules, and Rotary Position Embeddings (RoPE) to enhance efficiency in natural language processing tasks. Preliminary experiments with a low-quality dataset demonstrated the model's ability to generate responses, though the output lacked coherence and relevance. These findings highlight the necessity for further training on high-quality datasets, which is currently hindered by limited computational resources. Consequently, this research seeks collaborative opportunities and funding to facilitate comprehensive pretraining and fine-tuning of the model, aiming to develop efficient and effective language models accessible to researchers with constrained capabilities.

This study presents a novel application of a Probabilistic Bayesian Neural Network (PBNN) for estimating vocal function variables and enhancing non-invasive ambulatory voice monitoring by addressing aleatoric and epistemic uncertainties in regression tasks. The proposed PBNN allows for estimating key physiological parameters including subglottal pressure, vocal fold contact pressure, thyroaritenoid and cricothyroid muscle activations, from seven aerodynamic and acoustic features. The PBNN is trained on the Triangular Body-Cover Model (TBCM) of the vocal folds to produce a non-linear inverse mapping between its inputs and outputs. Furthermore, the selected aerodynamic and acoustic features can be obtained in ambulatory settings, thus enhancing the practical applicability of the proposed method. Transfer Learning is then applied to integrate real voice data into the initially synthetic-trained network to refine subglottal pressure estimations. The confidence intervals generated by the PBNN illustrate its ability to identify uncertain estimations, as the results show correlations between prediction errors and the estimated aleatoric and epistemic uncertainties. This correlation is advantageous because it shows that the network can effectively predict potential inaccuracies in its estimations. Increased uncertainty is mainly observed at operating points where the TBCM is likely to exhibit non-linear behaviors, at higher subglottal pressures. This suggests that the selected input features may not be robust enough for capturing the nonlinear effects in the TBCM. These results highlight the potential for future research to assess the viability of incorporating new features and additional measurements that could better capture non-linear responses.

As Cyber Physical Systems (CPSs), notably autonomous vehicles, generate increasing volumes of image-based data, efficient storage solutions become paramount. Leveraging high-density white-box SSDs and tailored data deduplication techniques presents a promising approach. However, conventional deduplication methods are ill-equipped to handle image file characteristics effectively, while existing image-specific solutions like imgDedup have limitations in format support and SSD integration. To address these challenges, we propose an SSDbased Approximate Image Deduplication (AID) system. SSD-AID comprises an image-customized deduplication, called AID, and an SSD-aware Image Consolidation. AID combines exact and approximate matching methods, along with optimized content compression, to enhance deduplication efficacy while integrating seamlessly with white-box SSDs. Furthermore, our SSD-aware Image Consolidation optimizes flash page utilization by grouping images based on access frequencies. Experimental results across multiple datasets demonstrate significant improvements in deduplication ratio and SSD page utilization. SSD-AID offers a comprehensive solution for efficient image data management in CPSs, addressing challenges in large-scale storage utilization and minimizing unused space.

Low-power and cost-effective IoT sensing nodes enable scalable monitoring of extensive, inaccessible, remote, and dangerous environments. However, some of these environments impose rough and extreme operating conditions on the nodes and require continuous adaptation and reconfiguration of physical and link layer parameters. In this paper, we closely investigate the stability of wireless links established between low-power IoT nodes deployed on the surface of different water bodies and propose dynamic transmission power adaptation to deal with link quality fluctuation and frequent disconnections. Our model is based on Minimum Mean Square Estimation (MMSE) and relies on statistics pertaining to received power and the acceleration the movement of water induces on communicating nodes. The model achieves a prediction accuracy exceeding 97%. The cost of processing and communicating the 3D acceleration statistics can be forgone with a reduced accuracy of 90% with a Kalman Filter.

In recent years, the rapid expansion of transformer-based architectures has significantly advanced the capabilities of text generation, translation, and other language-centric tasks. However, scaling such models while maintaining task-specific adaptability has become increasingly challenging due to the growing complexity of architectures and computational demands. Token adapters, a novel mechanism explored in this research, present a modular solution to extending model functionality without requiring extensive retraining of the core architecture. By introducing task-specific token manipulation layers into the Mistral model, substantial improvements were observed in perplexity, BLEU scores, and overall accuracy across a variety of tasks. The experimental results demonstrated the effectiveness of token adapters in enhancing linguistic coherence and context sensitivity, particularly for machine translation and question answering, while also highlighting the trade-offs in terms of memory consumption and training time. This research contributes to the broader exploration of model extensibility and proposes token adapters as a promising avenue for task-specific optimization within scalable architectures. The findings demonstrate the potential for modular approaches to increase both flexibility and specialization in language models, positioning token adapters as a valuable tool for future innovations in the field.

In this work, an electronically reconfigurable onedimensional guided wave fed slot array antenna is proposed. The individual antenna element making up the array is designed at X-band and is capable of switching between radiating and non-radiating states by use of a PIN diode used to control the operating frequency of the element. In addition to this discrete control of the phase constant (β), the average ON and OFF time of the individual elements is also used to taper the attenuation constant (α), and therefore the amplitude distribution, of the array which allows the reduction of the side lobe level (SLL) by virtue of time modulation. A 21 element linear array is fabricated and measured to validate the proposed concept in terms of beam scanning in the H-plane and SLL reduction. The limitations of the structure are also evaluated.