Search Results (20)

This paper presents a novel framework for optimizing multimodal large language model (MLLM) inference through task offloading and resource allocation in UAV-assisted satellite edge computing (SEC) networks. MLLMs leverage transformer architectures to integrate heterogeneous data modalities for IoT applications, particularly real-time monitoring in remote areas. However, cloud computing dependency introduces latency, bandwidth, and privacy challenges, while IoT device limitations require efficient distributed computing solutions. SEC, utilizing low-earth orbit (LEO) satellites and unmanned aerial vehicles (UAVs), extends mobile edge computing to provide ubiquitous computational resources for remote IoTDs. We formulate the joint optimization of MLLM task offloading and resource allocation as a mixed-integer nonlinear programming (MINLP) problem, minimizing latency and energy consumption while optimizing offloading decisions, power allocation, and UAV trajectories. To address the dynamic SEC environment characterized by satellite mobility, we propose an action-decoupled soft actor–critic (AD-SAC) algorithm with discrete–continuous hybrid action spaces. The simulation results demonstrate that our approach significantly outperforms conventional deep reinforcement learning methods in convergence and system cost reduction compared to baseline algorithms. Full article

Per- and polyfluoroalkyl substances (PFASs) are ubiquitous environmental contaminants with potential endocrine-disrupting properties. This study examines the association between exposure to multiple PFASs and pan-cancers associated with sex hormones (PCSH) while accounting for potential non-linear relationships and interactions. We analyzed data from the National Health and Nutrition Examination Survey (NHANES), spanning two-year cycles from 1999 to 2012 and including 14,373 participants. Serum concentrations of six PFAS—perfluorooctanoic acid (PFOA), perfluorooctanesulfonic acid (PFOS), perfluorohexanesulfonic acid (PFHxS), perfluorodecanoic acid (PFDE), perfluorononanoic acid (PFNA), and perfluoroundecanoic acid (PFUA)—were assessed for their relationship with PCSH. The statistical analyses included descriptive statistics, Spearman and Pearson correlation analyses, and both linear and logistic regression models. Additionally, Bayesian kernel machine regression (BKMR) was applied to capture potential nonlinear relationships and interactions. The initial t-tests showed a statistically significant difference in PFOS levels between individuals with and without PCSH (p = 0.0022), with higher mean PFOS levels in the PCSH group. Chi-square tests revealed a significant association between ethnicity and PCSH (p < 0.001). Linear and logistic regression analyses revealed significant associations for PFOS. BKMR analysis identified PFOA as having the highest posterior inclusion probability, indicating its importance in explaining PCSH risk. Univariate exposure-response analysis revealed limited individual PFAS effects. However, bivariate analysis indicated a complex U-shaped interaction pattern among many joint PFAS assessments. The overall exposure effect analysis suggested that the combined impact of all PFASs was more strongly associated with PCSH at exposure levels below the 0.5 quantile compared to higher levels. Single-variable interaction analyses highlighted PFOA and PFOS as the most interactive PFASs when evaluating their interaction with combined exposure to all other PFASs. In summary, while the initial findings suggested a positive association between PFOS and PCSH, the BKMR analysis revealed complex non-linear relationships and interactions among PFAS. These findings highlight the importance of evaluating PFASs as a mixture rather than as individual chemicals and using techniques that can capture non-linear relationships and interactions. Full article

To investigate the dynamic wave propagation characteristics and dynamic response of heterogeneous layered slopes under a blasting vibration, a modeling method considering the slope’s layered dip angle and heterogeneity was proposed. Different dip jointed slope models were established using the Weibull random distribution function introduced to realize the stochastic distribution of rock mechanics parameters, representing heterogeneity. Taking the background project of the Sijiaying Yanshan Open-Pit Iron Mine as an example, through numerical simulation, the effects of different joint dip angles and rock hardness on the slope’s dynamic response were analyzed in detail. The sensitivity of the elastic modulus, cohesion, and friction angle to the slope dynamic response was also investigated. A comparative analysis of the amplification effects between a jointed slope and heterogeneous slope was conducted. Finally, the dynamic stability of the jointed slope and heterogeneous slope under a blasting load was analyzed. The results indicate that the Peak Ground Acceleration (PGA) of jointed slopes with dip angles of 45° and 60° is generally higher than that of slopes with a 0° dip angle and without joints. The smaller the rock mass heterogeneity, the smaller the PGA at the measuring points, and the less sensitive the PGA is to variations in the three quantities. Under the same physical and mechanical parameters of the rock, the amplification factor of jointed slopes is generally greater than that of heterogeneous slopes. Under the blasting load, the overall dynamic time-series safety factors of both slopes decrease first and then increase, with the safety factor reaching its lowest value at the location of the strongest blasting vibration wave. This study can provide guidance for the blasting design and safety protection of layered dip slopes and serve as a reference for the analysis of blasting impact laws in similar mines. Full article

Resistance spot-welded joints are crucial parts in contemporary manufacturing technology due to their ubiquitous use in the automobile industry. The necessity of improving manufacturing efficiency and quality at an affordable cost requires deep knowledge of the resistance spot welding (RSW) process and the development of artificial neural network (ANN)- and machine learning (ML)-based modelling techniques, apt for providing essential tools for design, planning, and incorporation in the welding process. Tensile shear force and nugget diameter are the most crucial outputs for evaluating the quality of a resistance spot-welded specimen. This study uses ML and ANN models to predict shear force and nugget diameter responses to RSW parameters. The RSW analysis was executed on similar and dissimilar AISI 304 and grade 2 titanium alloy joints with equal and unequal thicknesses. The input parameters included welding current, pressure, welding duration, squeezing time, holding time, pulse welding, and sheet thickness. Linear regression, Decision tree, Support vector machine (SVM), Random forest (RF), Gradient-boosting, CatBoost, K-Nearest Neighbour (KNN), Ridge, Lasso, and ElasticNet machine learning algorithms, along with two different structures of Multilayer Perceptron, were utilized for studying the impact of the RSW parameters on the shear force and nugget diameter. Different validation metrics were applied to assess each model’s quality. Two equations were developed to determine the shear force and nugget diameter based on the investigation parameters. The current research also presents a prediction of the Relative Importance (RI) of RSW factors. Shear force and nugget diameter predictions were examined using SHapley (SHAP) Additive Explanations for the first time in the RSW field. Trainbr as the training function and Logsig as the transfer function delivered the best ANN model for predicting shear force in a one-output structure. Trainrp with Tansig made the most accurate predictions for nugget diameter in a one-output structure and for shear force and diameter in a two-output structure. Depending on validation metrics, the Random forest model outperformed the other ML algorithms in predicting shear force or nugget diameter in a one-output model, while the Decision tree model gave the best prediction using a two-output structure. Linear regression made the worst ML predictions for shear force, while ElasticNet made the worst nugget diameter forecasts in a one-output model. However, in two-output models, Lasso made the worst predictions. Full article

Accurate peach detection is essential for automated agronomic management, such as mechanical peach harvesting. However, ubiquitous occlusion makes identifying peaches from complex backgrounds extremely challenging. In addition, it is difficult to capture fine-grained peach features from a single RGB image, which can suffer from light and noise in scenarios with dense small target clusters and extreme light. To solve these problems, this study proposes a multimodal detector, called CRLNet, based on RGB and depth images. First, YOLOv9 was extended to design a backbone network that can extract RGB and depth features in parallel from an image. Second, to address the problem of information fusion bias, the Rough–Fine Hybrid Attention Fusion Module (RFAM) was designed to combine the advantageous information of different modes while suppressing the hollow noise at the edge of the peach. Finally, a Transformer-based Local–Global Joint Enhancement Module (LGEM) was developed to jointly enhance the local and global features of peaches using information from different modalities in order to enhance the percentage of information about the target peaches and remove the interference of redundant background information. CRLNet was trained on the Peach dataset and evaluated against other state-of-the-art methods; the model achieved an $mA{P}_{50}$ of 97.1%. In addition, CRLNet also achieved an $mA{P}_{50}$ of 92.4% in generalized experiments, validating its strong generalization capability. These results provide valuable insights for peach and other outdoor fruit multimodal detection. Full article

Based on the characteristic of ubiquitous radar emitting low-gain wide beam, a method of long-time coherent integration (LTCI) is required to enhance target detection capability. However, high-speed and highly maneuverable targets can cause Doppler frequency migration (DFM), range migration (RM), and velocity ambiguity (VA), severely degrading the performance of LTCI. Additionally, the number of targets is unknown and variable, and the presence of clutter further complicates the target tracking problem. To address these challenges, we propose a hybrid integration method to achieve joint detection and estimation of multiple high-speed, and highly maneuverable targets. Firstly, we compensate for first-order RM using the keystone transform (KT) and generate corresponding sub-range-Doppler (SRD) planes with different folding factors to achieve VA compensation. These SRD planes are then stitched together to form an extended range-Doppler (ERD) plane, which covers a broader velocity range. Secondly, during the track-before-detect (TBD) process, tracking is performed directly on the ERD plane. We use the sequential Monte Carlo (SMC) approximation of the probability hypothesis density (PHD) to propagate multi-target states. Additionally, we propose an amplitude-based adaptive prior distribution method and a line spread model (LSM) observation model to compensate for DFM. Since the acceleration of the target is included in the particle state, using particles to search for DFM does not increase the computational load. To address the issue of misclassifying mirror targets as real targets in the SRD plane, we propose a particle space projection method. By stacking the SRD planes to create a folding range-Doppler (FRD) space, particles are projected along the folding factor dimension, and then, the particles are clustered to eliminate the influence of the mirror targets. Finally, through simulation experiments, the superiority of the LSM for targets with acceleration was demonstrated. In comparative experiments, the proposed method showed superior performance and robustness compared to traditional methods, achieving a balance between performance and computational efficiency. Furthermore, the proposed method’s capability to detect and track multiple high-speed and highly maneuverable targets was validated using actual data from a ubiquitous radar system. Full article

Software-Defined Wide Area Networks (SD-WANs) are an emerging Software-Defined Network (SDN) technology to reinvent Wide Area Networks (WANs) for ubiquitous network interconnections in cloud computing, edge computing, and the Internet of Everything. The state-of-the-art overlay-based SD-WANs are simply conjunctions of Virtual Private Network (VPN) and SDN architecture to leverage the controllability and programmability of SDN, which are only applicable for specific platforms and do not comply with the extensibility of SDN. This paper motivates us to refactor traditional VPNs with SDN architecture by proposing an overlay-based SD-WAN solution named Software-Defined Virtual Private Network (SD-VPN). An SDN-based auto-constructed VPN model and its evaluating metrics are put forward to automatically construct overlay WANs by node placement and service orchestration of SD-VPN. Therefore, a joint placement algorithm of VPN nodes and algorithms for overlay WAN service loading and offloading are proposed for SD-VPN controllers. Finally, a three-layer SD-VPN system is implemented and deployed in actual network environments. Simulation experiments and system tests are conducted to prove the high-efficiency controllability, real-time programmability, and auto-constructed deployability of the proposed SD-VPN. Performance trade-off between SD-VPN control channels and data channels is evaluated, and SD-VPN controllers are proven to be extensible for other VPN protocols and advanced services. Full article

Multimodal datasets are ubiquitous in modern applications, and multimodal Variational Autoencoders are a popular family of models that aim to learn a joint representation of different modalities. However, existing approaches suffer from a coherence–quality tradeoff in which models with good generation quality lack generative coherence across modalities and vice versa. In this paper, we discuss the limitations underlying the unsatisfactory performance of existing methods in order to motivate the need for a different approach. We propose a novel method that uses a set of independently trained and unimodal deterministic autoencoders. Individual latent variables are concatenated into a common latent space, which is then fed to a masked diffusion model to enable generative modeling. We introduce a new multi-time training method to learn the conditional score network for multimodal diffusion. Our methodology substantially outperforms competitors in both generation quality and coherence, as shown through an extensive experimental campaign. Full article

Osteoarthritis (OA) is the most common joint disease that causes local inflammation and pain, significantly reducing the quality of life and normal social activities of patients. Currently, there are no disease-modifying OA drugs (DMOADs) available, and treatment relies on pain relief agents or arthroplasty. To address this significant unmet medical need, we aimed to develop monoclonal antibodies that can block the osteoclast-associated receptor (OSCAR). Our recent study has revealed the importance of OSCAR in OA pathogenesis as a novel catabolic regulator that induces chondrocyte apoptosis and accelerates articular cartilage destruction. It was also shown that blocking OSCAR with a soluble OSCAR decoy receptor ameliorated OA in animal models. In this study, OSCAR-neutralizing monoclonal antibodies were isolated and optimized by phage display. These antibodies bind to and directly neutralize OSCAR, unlike the decoy receptor, which binds to the ubiquitously expressed collagen and may result in reduced efficacy or deleterious off-target effects. The DMOAD potential of the anti-OSCAR antibodies was assessed with in vitro cell-based assays and an in vivo OA model. The results demonstrated that the anti-OSCAR antibodies significantly reduced cartilage destruction and other OA signs, such as subchondral bone plate sclerosis and loss of hyaline cartilage. Hence, blocking OSCAR with a monoclonal antibody could be a promising treatment strategy for OA. Full article

In recent years, wave energy has gained attention for its sustainability and cleanliness. As one of the most important parameters of wave energy, significant wave height (SWH) is difficult to accurately predict due to complex ocean conditions and the ubiquitous chaotic phenomena in nature. Therefore, this paper proposes an integrated CEEMDAN-LSTM joint model. Traditional computational fluid dynamics (CFD) has a long calculation period and high capital consumption, but artificial intelligence methods have the advantage of high accuracy and fast convergence. CEEMDAN is a commonly used method for digital signal processing in mechanical engineering, but has not yet been used for SWH prediction. It has better performance than the EMD and EEMD and is more suitable for LSTM prediction. In addition, this paper also proposes a novel filter formulation for SWH outliers based on the improved violin-box plot. The final empirical results show that CEEMDAN-LSTM significantly outperforms LSTM for each forecast duration, significantly improving the prediction accuracy. In particular, for a forecast duration of 1 h, CEEMDAN-LSTM has the most significant improvement over LSTM, with 71.91% of RMSE, 68.46% of MAE and 6.80% of NSE, respectively. In summary, our model can improve the real-time scheduling capability for marine engineering maintenance and operations. Full article

Inverse kinematics problems (IKP) are ubiquitous in robotics for improved robot control in widespread applications. However, the high non-linearity, complexity, and equation coupling of a general six-axis robotic manipulator pose substantial challenges in solving the IKP precisely and efficiently. To address this issue, we propose a novel approach based on neural network (NN) with numerical error minimization in this paper. Within our framework, the complexity of IKP is first simplified by a strategy called joint space segmentation, with respective training data generated by forward kinematics. Afterwards, a set of multilayer perception networks (MLP) are established to learn from the foregoing data in order to fit the goal function piecewise. To reduce the computational cost of the inference process, a set of classification models is trained to determine the appropriate forgoing MLPs for predictions given a specific input. After the initial solution is sought, being improved with a prediction error minimized, the refined solution is finally achieved. The proposed methodology is validated via simulations on Xarm6—a general 6 degrees of freedom manipulator. Results further verify the feasibility of NN for IKP in general cases, even with a high-precision requirement. The proposed algorithm has showcased enhanced efficiency and accuracy compared to NN-based approaches reported in the literature. Full article

To investigate the deformation and stability of the rock surrounding a tunnel excavated through layered rock, a three-dimensional numerical simulation was conducted to compare and analyze the right line of the YD2K100+560–630 two-stage docking cross-section of the Yuelongmen Tunnel using a ubiquitous joint model. Based on 3-D numerical simulations, the following conclusions were drawn: (1) The excavation on the construction entrance side significantly disturbed the construction exit side, and the exit side was more disturbed in the steeply dipping layer than in the gently dipping layer. (2) The influence of the dip angle of the laminae on the uplift of the tunnel invert was greater than the influence on the tunnel vault settlement when the laminae were oriented perpendicular to the trend of the tunnel and two-stage docking excavation was used. (3) When two-stage excavation was applied to gently inclined laminated rock, there was an obvious disturbance to both the top of the tunnel (less than one time the diameter of the through face; and to the bottom of the tunnel within 0.5 time the diameter of the through face on the side with the stopping construction exit side during the tunnelling process. When tunnel excavation was conducted against the trend of the layer, there was obvious disturbance to the top of the tunnel within 0.5 time the diameter of the through face and to the bottom of the tunnel within one time the diameter of the entrance face) on the construction exit side during the tunnelling process due to the presence of laminae. The excavation on the construction entrance side significantly disturbed the construction exit side. When docking and penetrating layered rock tunnels, priority should be given to docking and penetrating in the direction corresponding to the layers, and temporary support on the exit side should be strengthened to ensure the deformation stability of the surrounding rock mass. Full article

Multi-modal sensor fusion has become ubiquitous in the field of vehicle motion estimation. Achieving a consistent sensor fusion in such a set-up demands the precise knowledge of the misalignments between the coordinate systems in which the different information sources are expressed. In ego-motion estimation, even sub-degree misalignment errors lead to serious performance degradation. The present work addresses the extrinsic calibration of a land vehicle equipped with standard production car sensors and an automotive-grade inertial measurement unit (IMU). Specifically, the article presents a method for the estimation of the misalignment between the IMU and vehicle coordinate systems, while considering the IMU biases. The estimation problem is treated as a joint state and parameter estimation problem, and solved using an adaptive estimator that relies on the IMU measurements, a dynamic single-track model as well as the suspension and odometry systems. Additionally, we show that the validity of the misalignment estimates can be assessed by identifying the misalignment between a high-precision INS/GNSS and the IMU and vehicle coordinate systems. The effectiveness of the proposed calibration procedure is demonstrated using real sensor data. The results show that estimation accuracies below 0.1 degrees can be achieved in spite of moderate variations in the manoeuvre execution. Full article

Manipulation of deformable objects is a desired skill in making robots ubiquitous in manufacturing, service, healthcare, and security. Common deformable objects (e.g., wires, clothes, bed sheets, etc.) are significantly more difficult to model than rigid objects. In this research, we contribute to the model-based manipulation of linear flexible objects such as cables. We propose a 3D geometric model of the linear flexible object that is subject to gravity and a physical model with multiple links connected by revolute joints and identified model parameters. These models enable task automation in manipulating linear flexible objects both in simulation and real world. To bridge the gap between simulation and real world and build a close-to-reality simulation of flexible objects, we propose a new strategy called Simulation-to-Real-to-Simulation (Sim2Real2Sim). We demonstrate the feasibility of our approach by completing the Plug Task used in the 2015 DARPA Robotics Challenge Finals both in simulation and real world, which involves unplugging a power cable from one socket and plugging it into another. Numerical experiments are implemented to validate our approach. Full article

Depth imaging has, through recent technological advances, become ubiquitous as products become smaller, more affordable, and more precise. Depth cameras have also emerged as a promising modality for activity recognition as they allow detection of users’ body joints and postures. Increased resolutions have now enabled a novel use of depth cameras that facilitate more fine-grained activity descriptors: The remote detection of a person’s breathing by picking up the small distance changes from the user’s chest over time. We propose in this work a novel method to model chest elevation to robustly monitor a user’s respiration, whenever users are sitting or standing, and facing the camera. The method is robust to users occasionally blocking their torso region and is able to provide meaningful breathing features to allow classification in activity recognition tasks. We illustrate that with this method, with specific activities such as paced-breathing meditating, performing breathing exercises, or post-exercise recovery, our model delivers a breathing accuracy that matches that of a commercial respiration chest monitor belt. Results show that the breathing rate can be detected with our method at an accuracy of 92 to 97% from a distance of two metres, outperforming state-of-the-art depth imagining methods especially for non-sedentary persons, and allowing separation of activities in respiration-derived features space. Full article