Search Results (168)

In this paper, a new approach to topology optimization using the parameterized level set function and genetic algorithm optimization methods is presented. The impact of a number of parameters describing the level set function in the representation of the model was examined. Using the B-spline interpolation function, the number of variables describing the level set function was decreased, enabling the application of evolutionary methods (genetic algorithms) in the topology optimization process. The traditional level set method is performed by using the Hamilton–Jacobi transport equation, which implies the use of gradient optimization methods that are prone to becoming stuck in local minima. Furthermore, the resulting optimal shapes are strongly dependent on the initial solution. The proposed topology optimization procedure, written in MATLAB R2013b, utilizes a genetic algorithm for global optimization, enabling it to locate the global optimum efficiently. To assess the acceleration and convergence capabilities of the proposed topology optimization method, a new genetic algorithm penalty operator was tested. This operator addresses the slow convergence issue typically encountered when the genetic algorithm optimization procedure nears a solution. By penalizing similar individuals within a population, the method aims to enhance convergence speed and overall performance. In complex examples (3D), the method can also function as a generator of good initial solutions for faster topology optimization methods (e.g., level set) that rely on such initial solutions. Both the proposed method and the traditional methods have their own advantages and limitations. The main advantage is that the proposed method is a global search method. This makes it robust against entrapment in local minima and independent of the initial solution. It is important to note that this evolutionary approach does not necessarily perform better in terms of convergence speed compared to gradient-based or other local optimization methods. However, once the global optimum has been found using the genetic algorithm, convergence can be accelerated using a faster local method such as gradient-based optimization. The application and usefulness of the method were tested on typical 2D cantilever beams and Michell beams. Full article

This research describes a multi-objective trajectory planning method for robotic arms based on time, energy, and impact. The quintic Non-Uniform Rational B-Spline (NURBS) curve was employed to interpolate the trajectory in joint space. The quintic NURBS interpolation curve can make the trajectory become constrained within the kinematic limits of velocity, acceleration, and jerk while also satisfying the continuity of jerk. Then, based on the Parrot Optimization (PO) algorithm, through improvements to reduce algorithmic randomness and the introduction of appropriate multi-objective strategies, the algorithm was extended to the Multi-Objective Parrot Optimization (MOPO) algorithm, which better balances global search and local convergence, thereby more effectively solving multi-objective optimization problems and reducing the impact on optimization results. Subsequently, by integrating interpolation curves, the multi-objective optimization of joint trajectories could be performed under robotic kinematic constraints based on time–energy-jerk criteria. The obtained Pareto optimal front can provide decision-makers in industrial robotic arm applications with flexible options among non-dominated solutions. Full article

This paper investigates the derivation of global shape functions in T-meshed quadrilateral patches through transfinite interpolation and local elimination. The same shape functions may be alternatively derived starting from a background tensor product of Lagrange polynomials and then imposing linear constraints. Based on the nodal points of the T-mesh, which are associated with the primary degrees of freedom (DOFs), all the other points of the background grid (i.e., the secondary DOFs) are interpolated along horizontal and vertical stations (isolines) of the tensor product, and thus, linear relationships are derived. By implementing these constraints into the original formula/expression, global shape functions, which are only associated with primary DOFs, are created. The quality of the elements is verified by the numerical solution of a typical potential problem of second order, with boundary conditions of Dirichlet and Neumann type. Full article

White blood cell (WBC) classification plays a crucial role in hematopathology and clinical diagnostics. However, traditional methods are constrained by limited receptive fields and insufficient utilization of contextual information, which hinders classification performance. To address these limitations, this paper proposes an enhanced WBC classification algorithm, CCE-YOLOv7, which is built upon the YOLOv7 framework. The proposed method introduces four key innovations to enhance detection accuracy and model efficiency: (1) A novel Conv2Former (Convolutional Transformer) backbone was designed to combine the local pattern extraction capability of convolutional neural networks (CNNs) with the global contextual reasoning of transformers, thereby improving the expressiveness of feature representation. (2) The CARAFE (Content-Aware ReAssembly of Features) upsampling operator was adopted to replace conventional interpolation methods, thereby enhancing the spatial resolution and semantic richness of feature maps. (3) An Efficient Multi-scale Attention (EMA) module was introduced to refine multi-scale feature fusion, enabling the model to better focus on spatially relevant features critical for WBC classification. (4) Soft-NMS (Soft Non-Maximum Suppression) was used instead of traditional NMS to better preserve true positives in densely packed or overlapping cell scenarios, thereby reducing false positives and false negatives. Experimental validation was conducted on a WBC image dataset acquired using the Fourier ptychographic microscopy (FPM) system. The proposed CCE-YOLOv7 achieved a detection accuracy of 89.3%, showing a 7.8% improvement over the baseline YOLOv7. Furthermore, CCE-YOLOv7 reduced the number of parameters by 2 million and lowered computational complexity by 5.7 GFLOPs, offering an efficient and lightweight model suitable for real-time clinical applications. To further evaluate model effectiveness, comparative experiments were conducted with YOLOv8 and YOLOv11. CCE-YOLOv7 achieved a 4.1% higher detection accuracy than YOLOv8 while reducing computational cost by 2.4 GFLOPs. Compared with the more advanced YOLOv11, CCE-YOLOv7 maintained competitive accuracy (only 0.6% lower) while using significantly fewer parameters and 4.3 GFLOPs less in computation, highlighting its superior trade-off between accuracy and efficiency. These results demonstrate that CCE-YOLOv7 provides a robust, accurate, and computationally efficient solution for automated WBC classification, with significant clinical applicability. Full article

Sound speed profiles (SSPs) must be detected simultaneously to perform a multibeam depth survey. Accurate real-time sound speed profile (SSP) acquisition remains a critical challenge in deep-sea multibeam bathymetry due to the limitations regarding direct measurements under harsh operational conditions. To address the issue, we propose a joint inversion framework integrating World Ocean Atlas 2023 (WOA23) temperature–salinity model data, historical in situ SSPs, and surface sound speed measurements. By constructing a high-resolution regional sound speed field through WOA23 and historical SSP fusion, this method effectively mitigates spatiotemporal heterogeneity and seasonal variability. The artificial lemming algorithm (ALA) is introduced to optimize the inversion of empirical orthogonal function (EOF) coefficients, enhancing global search efficiency while avoiding local optimization. An experimental validation in the northwest Pacific Ocean demonstrated that the proposed method has a better performance than that of conventional substitution, interpolation, and WOA23-only approaches. The results indicate that the mean absolute error (MAE), root mean square error (RMSE), and maximum error (ME) of SSP reconstruction are reduced by 41.5%, 46.0%, and 49.4%, respectively. When the reconstructed SSPs are applied to multibeam bathymetric correction, depth errors are further reduced to 0.193 m (MAE), 0.213 m (RMSE), and 0.394 m (ME), effectively suppressing the “smiley face” distortion caused by sound speed gradient anomalies. The dynamic selection of the first six EOF modes balances computational efficiency and reconstruction fidelity. This study provides a robust solution for real-time SSP estimation in data-scarce deep-sea environments, particularly for underwater autonomous vehicles. This method effectively mitigates the seabed distortion caused by missing real-time SSPs, significantly enhancing the accuracy and efficiency of deep-sea multibeam surveys. Full article

Human motion recognition is crucial for applications like navigation, health monitoring, and smart healthcare, especially in weak GNSS scenarios. Current methods face challenges such as limited sensor diversity and inadequate feature extraction. This study proposes a CNN–Transformer–Attention framework with multimodal enhancement to address these challenges. We first designed a lightweight wearable system integrating synchronized accelerometer, gyroscope, and magnetometer modules at wrist, chest, and foot positions, enabling multi-dimensional biomechanical data acquisition. A hybrid preprocessing pipeline combining cubic spline interpolation, adaptive Kalman filtering, and spectral analysis was developed to extract discriminative spatiotemporal-frequency features. The core architecture employs parallel CNN pathways for local sensor feature extraction and Transformer-based attention layers to model global temporal dependencies across body positions. Experimental validation on 12 motion patterns demonstrated 98.21% classification accuracy, outperforming single-sensor configurations by 0.43–7.98% and surpassing conventional models (BP-Network, CNN, LSTM, Transformer, KNN) through effective cross-modal fusion. The framework also exhibits improved generalization with 3.2–8.7% better accuracy in cross-subject scenarios, providing a robust solution for human activity recognition and accurate positioning in challenging environments such as autonomous navigation and smart cities. Full article

As an important branch of machine learning, federated learning still suffers from statistical heterogeneity. Therefore, personalized federated learning (PFL) is proposed to deal with this obstacle. However, the privacy of local and global gradients is still under threat in the scope of PFL. Additionally, the correctness of the aggregated result is unable to be identified. Therefore, we propose a secure and verifiable personalized federated learning protocol that could protect privacy using homomorphic encryption and verify the aggregated result using Lagrange interpolation and commitment. Furthermore, it could resist the collusion attacks performed by servers and clients who try to pass verification. Comprehensive theoretical analysis is provided to verify our protocol’s security. Extensive experiments on MNIST, Fashion-MNIST and CIFAR-10 are carried out to demonstrate the effectiveness of our protocol. Our model achieved accuracies of 88.25% in CIFAR-10, 99.01% in MNIST and 96.29% in Fashion-MNIST. The results show that our protocol could improve security while maintaining the classification accuracy of the training model. Full article

This study presents an enhanced deep learning model, Multi-Scale Fusion MaxViT (MSF-MaxViT), designed for medical image classification. The aim is to improve both the accuracy and robustness of the image classification task. MSF-MaxViT incorporates a Parallel Attention mechanism for fusing local and global features, inspired by the MaxViT Block and Multihead Dynamic Attention, to improve feature representation. It also combines lightweight components such as the novel Multi-Scale Fusion Attention (MSFA) block, Feature Boosting (FB) Block, Coord Attention, and Edge Attention to enhance spatial and channel feature learning. To optimize the hyperparameters in the network model, the Super Beluga Whale Optimization (SBWO) algorithm is used, which combines bi-interpolation and adaptive parameter tuning, and experiments have shown that it has a relatively excellent convergence performance. The network model, combined with the improved SBWO algorithm, has an image classification accuracy of 92.87% on the HAM10000 dataset, which is 1.85% higher than that of MaxViT, proving the practicality and effectiveness of the model. Full article

The application of robotic arms in the industrial field is continuously becoming greater and greater. The impact force generated by a robotic arm in a gripping operation leads to vibration and wear. To address this problem, this paper proposes a trajectory planning method based on the improved Slime Mould Algorithm. An interpolation curve under the joint coordinate system is constructed by using seven non-uniform B-spline functions, with time and impact force as the optimization objectives and angular velocity, angular acceleration, and angular acceleration as the constraints. The original algorithm introduces Bernoulli chaotic mapping to increase the diversity of the population, adaptively adjusts the feedback factor, improves the crossover operator to accelerate the global convergence, and combines the original algorithm with an improved artificial bee colony search strategy guided by the global optimal solution, adding a quadratic interpolation method to increase the diversity of the population and to accelerate the global convergence speed. Combined with the improved artificial swarm search strategy guided by the global optimal solution, the quadratic interpolation method is added to enhance the local utilization ability. The simulation and real-machine experimental results show that the improved algorithm shortens the movement time of the robotic arm, reduces the joint impacts, minimizes the vibration and wear, and prolongs the service life of the robotic arm. Full article

Federated learning, with its unique privacy protection mechanisms and distributed model training capabilities, provides an effective solution for data security by addressing the challenges associated with the inability to directly share private data due to privacy concerns. It exhibits broad application potential across various fields, particularly in scenarios such as autonomous vehicular networks, where collaborative learning is required from data sources distributed across different clients, thus optimizing and enhancing model performance. Nevertheless, in complex real-world environments, challenges such as data poisoning and labeling errors may cause some clients to introduce label noise that significantly exceeds ordinary levels, severely impacting model performance. The following conclusions are drawn from research on extreme label noise: highly polluted data severely affect the generalization capability of the global model and the stability of the training process, while the reweighting strategy can improve model performance. Based on these research conclusions, we propose a method named Enhanced Knowledge Distillation and Particle Swarm Optimization for Federated Learning (FedDPSO) to deal with extreme label noise. In FedDPSO, the server dynamically identifies extremely noisy clients based on uncertainty. It then uses the particle swarm optimization algorithm to determine client model weights for global model aggregation. In subsequent rounds, the identified extremely noisy clients construct an interpolation loss combining pseudo-label loss and knowledge distillation loss, effectively mitigating the negative impact of label noise overfitting on the local model. We carried out experiments on the CIFAR10/100 datasets to validate the effectiveness of FedDPSO. At the highest noise ratio under Beta = (0.1, 0.1), experiments show that FedDPSO improves the average accuracy on CIFAR10 by 15% compared to FedAvg and by 11% compared to the more powerful FOCUS. On CIFAR100, it outperforms FedAvg by 8% and FOCUS by 5%. Full article

This paper accompanies the initial public release of the EDI multi-modal SLAM dataset, a collection of long tracks recorded with a portable sensor package. These include two global shutter RGB camera feeds, LiDAR scans, as well as inertial and GNSS data from an RTK-enabled IMU-GNSS positioning module—both as satellite fixes and internally fused interpolated pose estimates. The tracks are formatted as ROS1 and ROS2 bags, with separately available calibration and ground truth data. In addition to the filtered positioning module outputs, a second form of sparse ground truth pose annotation is provided using independently surveyed visual fiducial markers as a reference. This enables the meaningful evaluation of systems that directly utilize data from the positioning module into their localization estimates, and serves as an alternative when the GNSS reference is disrupted by intermittent signals or multipath scattering. In this paper, we describe the methods used to collect the dataset, its contents, and its intended use. Full article

Eight-node quadrilateral isoparametric elements of the serendipity type have frequently been used in finite-element analyses of two-dimensional seepage problems. The shape functions for these elements are quadratic. Hence, nonlinear variation in the potential and stream function values across each element could be approximated to a high degree of accuracy. This also necessitates a commensurate high-order interpolation function to locate, in a straightforward way, equipotential lines and streamlines. In this paper, a quadratic interpolation algorithm for locating deformation contours is modified to suit flow net generation. The modification lies in the procedure for identifying the pairs of the points of intersection to be joined when there are four, six, or eight points of intersection of the contour segments of the same level and the edges of an element. The original algorithm finds the pairs of intersection points in a local coordinate system by testing all possible cases that may be encountered. The modified algorithm considers that in most, if not all, scenarios, equipotential lines and streamlines extend monotonically from one impervious boundary of the flow domain to another and from an inflow boundary to an outflow boundary, respectively. The intersection points are rapidly paired by converting their local coordinates to global coordinates and sorting the order of the intersection points according to their global coordinates. The modified algorithm eliminates the need for an exhaustive search and complex matching process, enhancing computational efficiency. The modified algorithm is verified against an exact analytical solution to the flow net for a levee under-seepage flow. Excellent agreement is obtained. Two additional illustrative examples are analyzed. One is unconfined seepage through a rectangular dam, and the other is confined seepage beneath unsymmetrical cofferdams. The equipotential lines and streamlines obtained from the modified algorithm are shown to be smoother and more accurate than those obtained using popular commercial software (GeoStudio 24.2.0), especially when a coarse finite-element mesh is adopted. Full article

Based on the idea of a surface moving frame in differential geometry, a surface envelopment approximation method is proposed for the integrated design of point-contact tooth surfaces. This method utilizes the envelopment characteristic curve of the first tooth surface as the spline curve and adopts the local structure of the second tooth surface along a predesigned contact path as the surface interpolation condition. Through motion transformation described by the motion invariants of the first tooth surface, a conjugate motion space for the second tooth surface is fully defined by the motion invariants of the first tooth surface. This constitutes the basis of the integrated optimization design space and ensures the global optimization and machinability of the tooth surface design method. Using the experimental data of the point meshing tooth surface loading contact, the gap between the two tooth surfaces during no-load meshing is used as the design target parameter to predict and control the shape and size of the contact area under heavy load and further the symmetry requirements of the tooth surface design. Consequently, a variational inequality model for the global optimal design of the point meshing tooth surface is established. Full article

To improve the intelligence level of particleboard inspection lines, machine vision and artificial intelligence technologies are combined to replace manual inspection with automatic detection. Aiming at the problem of missed detection and false detection on small defects due to the large surface width, complex texture and different surface defect shapes of particleboard, this paper introduces image super-resolution technology and proposes a super-resolution reconstruction model for particleboard images. Based on the Transformer network, this model incorporates an improved SRResNet (Super-Resolution Residual Network) backbone network in the deep feature extraction module to extract deep texture information. The shallow features extracted by conv 3 × 3 are then fused with features extracted by the Transformer, considering both local texture features and global feature information. This enhances image quality and makes defect details clearer. Through comparison with the traditional bicubic B-spline interpolation method, ESRGAN (Enhanced Super-Resolution Generative Adversarial Network), and SwinIR (Image Restoration Using Swin Transformer), the effectiveness of the particleboard super-resolution reconstruction model is verified using objective evaluation metrics including PSNR, SSIM, and LPIPS, demonstrating its ability to produce higher-quality images with more details and better visual characteristics. Finally, using the YOLOv8 model to compare defect detection rates between super-resolution images and low-resolution images, the mAP can reach 96.5%, which is 25.6% higher than the low-resolution image recognition rate. Full article

Currently, flexible robots, exemplified by parallel robots, play a crucial role in the automated packaging of agricultural products due to their rapid, accurate, and stable characteristics. This research systematically explores trajectory planning strategies for parallel robots in the high-speed tomato-grabbing process. Kinematic analysis of the parallel robot was conducted using geometric methods, deriving the coordinates of each joint at various postures, resulting in a kinematic forward solution model and corresponding equations, which were verified with data. To address the drawbacks of the point-to-point “portal” trajectory in tomato grabbing, a 3-5-5-3 polynomial interpolation method in joint space was proposed to optimize the path, enhancing trajectory smoothness. To improve the efficiency of the tomato packaging process, a hybrid algorithm combining particle swarm optimization (PSO) and genetic algorithms (GA) was developed to optimize the operation time of the parallel robot. Compared to traditional PSO, the proposed algorithm exhibits better global convergence and is less likely to fall into local optima, thereby ensuring a smoother and more efficient path in the robot-grabbing tomato process and providing technical support for automated tomato packaging. Full article