Abstract The planetary threaded roller bearings (PTRB) sustain higher axial loads than traditional thrust roller bearings but suffer from large contact stresses at the initial thread teeth. In this study, the PTRB is redesigned by incorporating combined tooth profiles and the convex-concave contact mode, effectively reducing contact stresses at the initial thread teeth. A load distribution model is developed to characterize contact stresses at the inner ring-roller and roller-outer ring interfaces. A multi-objective optimization model using the feasibility identification-based NSGA-II algorithm is established to improve static axial load rating and fatigue life. Since some Pareto-optimal solutions perform worse than the initial design, a high-quality solution region is defined, whose solutions consistently enhance the transmission performance and durability of the PTRB.

Abstract Machine Learning (ML) and Deep Learning (DL) have been used for anomaly detection in Industrial Internet of Things (IIoT) environments. The presence of imbalanced data, high noise levels, missing values, and high dimensionality poses an enormous challenge for existing methods, leading to inconsistent reliability in detecting anomalies in real-world industrial environments. Current anomaly detection solutions suffer from high false-negative rates due to class imbalance and noisy sensor data, limiting their practical applicability. This paper proposes the Ensemble Wasserstein Generative Adversarial Network for IIoT (EWAD-IIoT) framework, which is uniquely designed to address these challenges. The aim is to build a robust anomaly detection model with high recall (94.7%) and precision (93.6%) while minimizing miss rates in complex IIoT settings. Evaluations on two benchmark datasets, SECOM (industrial sensor data) and MNIST (image data), demonstrate EWAD-IIoT’s superiority over traditional methods like standalone WGAN and WGAN-GP. To highlight its efficacy, we compare results against these benchmarks, showcasing improvements in F1-score (95.8%) and noise robustness. The framework leverages advanced preprocessing (Z-score filtering, Min-Max scaling), SMOTE-based balancing, and WGAN-generated synthetic samples to handle data imbalance and dimensionality. The results validate EWAD-IIoT’s capability to detect rare anomalies in IIoT environments, with a balanced trade-off between recall and precision, making it a scalable solution for predictive maintenance and fault diagnosis.

Abstract As the core component responsible for high-frequency power switching in photovoltaic inverters, accurately predicting the remaining useful life (RUL) of insulated gate bipolar transistors (IGBTs) has become a key factor in ensuring the stable operation of photovoltaic systems. However, existing methods struggle to precisely characterize the degradation characteristics and processes of IGBTs at different time points. To address these issues, this paper proposes a MIG-PI-Pathformer RUL prediction method that integrates physical degradation models with deep learning. This method establishes a multi-stage Inverse Gaussian degradation model based on the physical failure mechanisms of IGBTs and couples it with the dual attention mechanism of the Pathformer model to capture complex degradation features, adaptively divide time scales, and thereby correct prediction errors in the physical model; Additionally, physical rule constraints are incorporated into the Pathformer loss function to ensure that RUL predictions align with degradation mechanisms. Simulation results show that, on NASA’s IGBT aging dataset, compared to the single Pathformer, the proposed method reduces MSE and MAE by 70.21% and 17.84%, respectively, and improves R2 by 7.66%. This method provides more accurate and physically interpretable technical support for fault warning and optimized maintenance of photovoltaic inverters.

Abstract With the widespread application of rotating machinery in intelligent manufacturing, aerospace, and other industrial fields, accurate and reliable fault diagnosis and maintenance have become increasingly critical for ensuring system safety and operational efficiency. However, existing domain-adaptation-based cross-domain intelligent fault diagnosis methods primarily focus on achieving feature transfer at the global domain level, often overlooking the complexity, imbalance, and significant class-level variability arising from the simultaneous distribution of samples across the source and target domains. This oversight can lead to inaccurate recognition of fine-grained class-level features, thereby limiting diagnostic accuracy. To address these challenges, this paper presents a class-level domain alignment method (CDD_DANN) that combines Classifier Deterministic Difference (CDD) loss with a dual-classifier structured Domain-Adversarial Neural Network (DANN), effectively improving class-level feature alignment and transfer in cross-domain fault diagnosis. Additionally, to effectively address the challenge of sparse marginal samples at deeper levels, we propose the PMCDAN method, which replaces CDD with a proxy-based metric learning approach, Proxy Neighborhood Component Analysis (ProxyNCA), to capture deeply shared features between the source and target domains more robustly. This enables global domain alignment and class alignment under challenging conditions. Furthermore, to tackle the data imbalance, this paper incorporates a Diffusion-GAN-based fault sample augmentation method, which facilitates both domain and class-level alignment when data is scarce, thus enabling more accurate fault diagnosis. The effectiveness and superiority of the proposed approach are validated through experimental evaluations against existing methods using the Paderborn University bearing dataset and a self-collected gear fault dataset. The proposed method provides valuable insights and practical guidance for fault diagnosis in complex real-world industrial scenarios.

Abstract Vision-language models with large-scale image-text pairs have shown significant potential on representation learning. Human pose estimation task, which is highly sensitive to pixel-wise transformation, requires effective methods for mining pose-specific knowledge. In this paper, we investigate the homologous human pose retrieval task relying on large-scale annotated datasets to enhance pose knowledge extraction. We propose Pose Prompt (PosePro), which leverages vision-language models to categorize global pose configuration of an image, build compatible design, generate pose embedding as proposals. We then aim to integrate the learned knowledge as visual and textual prompt to facilitate the learning processing of newly unseen tasks. We demonstrate the effectiveness of fundamental PosePro model through extensive experiments on both pose retrieval and human pose estimation, showing significant improvements in accuracy and generalization ability, especially in scenarios with limited samples.

Abstract Medical image registration is a critical problem in medical image analysis, enabling the spatial alignment of anatomical structures across different imaging modalities. However, existing algorithms often struggle with local registration of large deformations and exhibit limited feature extraction capabilities. To address these challenges, we propose a Multi-level Transformation Progressive Registration Algorithm (MTPR). This method incorporates the concept of multi-level transformations, the model performs four progressive registration steps, predicting the deformation field from coarse to fine. Initially, the model applies an enhancement process, introducing a hybrid filtering enhancement module based on wavelet transform and improved guided filtering to enhance image edges. During the registration phase, we propose a pyramid shared weight enhancement network (PWE-Net), which precisely extracts multimodal image features and implements a progressive deformation field prediction strategy. In order to increase the feature extraction capability of the model, we propose spatial feature fusion module in skip connection of encoder and decoder, which combines multi-scale information into a spatial feature representation with rich context information. Additionally, we introduce a dual-similarity metric to enhance model capability for local organ registration by incorporating structural similarity, increasing the model's attention to local organs. Experiments conducted on publicly available datasets (OASIS, LPBA40) and clinical CT/MR data, achieved an average dice similarity coefficient (DSC) of 0.822, average average symmetric surface distance (ASSD) of 0.741 mm, average standard deviation of jacobian determinant (Std. Jacobian) of 0.247 in the clinical CT/MR data. The Wilcoxon signed-rank test statistical analysis shows that the evaluation indicators of the MTPR algorithm are significantly better than other baseline methods (P < 0.05). The MTPR model's multi-scale information aggregation capability effectively handles large deformations, demonstrating excellent registration accuracy and generalization performance.

Abstract The Aggregate Nesting (AN) methodology is proposed to address the complexities of multi-material nesting optimization in industries such as shipbuilding, where both cost-efficiency and adaptability are critical. Unlike traditional single-material nesting approaches, AN dynamically allocates items across multiple plates, optimizing material yield and reducing overall production costs. This methodology utilizes a graph-based model, integrating heuristic techniques to solve complex nesting problems with greater scalability and efficiency. Real-world experiments conducted with data from a South Korean shipyard demonstrate that AN significantly reduces plate count, material waste, and overall production costs compared to traditional expert-based nesting methods. Although computational efficiency remains an area for further improvement, particularly in the aggregated allocation (P2QR) step, AN notably reduces manual workloads by automating the nesting process. Additionally, AN shows promise in improving decision-making through dynamic plate size adjustments based on material attributes. Future work aims to integrate reinforcement learning techniques to further enhance the adaptability and scalability of the methodology, enabling its application to larger-scale and more complex industrial scenarios.

Abstract Human Activity Recognition (HAR) plays a crucial role in identifying and digitizing human behaviors. Among various approaches, sound-based HAR offers distinct advantages, such as overcoming visual limitations and enabling recognition in diverse environments. This study introduces an innovative application of sound segmentation with SegNet, originally designed for image segmentation, to sound-based HAR. Traditionally, labeling sound data has been challenging due to its limited scope, often restricted to specific events or time frames. To address this issue, a novel labeling approach was developed, allowing detailed annotations across the entire temporal and frequency domains. This method facilitates the use of SegNet, which requires pixel-level labeling for accurate segmentation, leading to more granular and explainable activity recognition. A dataset comprising six distinct human activities—speech, groaning, screaming, coughing, toilet flushing, and snoring—was constructed to enable comprehensive evaluation. The trained neural network, utilizing this annotated dataset, achieved F1 scores ranging from 0.68 to 0.95. The model's practical applicability was further validated through recognition tests conducted in a professional office environment. This study presents a novel framework for quantifying daily human activities through sound segmentation, contributing to advancements in intelligent system technology.

Abstract This study presents a ​novel optimization framework leveraging the ​Adaptive Genetic Algorithm (AGA) to achieve ​global collaborative optimization of ​multi-parameter lubrication film texture morphology. By constructing a ​multi-degree-of-freedom parametric model, the framework enables ​independent control of texture depth and spatial orientation angles for each texture unit. A ​chromosome mapping mechanism is developed to encode ​54 geometric parameters across ​18 mirrored texture units, facilitating efficient optimization of complex surface textures. The ​Reynolds equation is numerically solved to evaluate the ​lubrication film bearing capacity, which serves as the ​fitness function for optimization. The results demonstrate a ​37.3% improvement in bearing capacity compared to uniform protrusion textures, achieving ​11.4 N with a ​7.1% reduction in resistance. The optimized texture configuration generates ​synergistic pressure superposition through ​strategically distributed inlet pits and outlet protrusions, forming ​continuous high-pressure zones that maximize hydrodynamic effects. This methodology not only addresses the limitations of traditional uniform texture designs but also provides a ​systematic and efficient approach for ​complex surface texture optimization in tribological applications. The proposed framework leverages the ​adaptive genetic algorithm's global search capabilities to achieve ​multi-parameter collaborative optimization, offering significant advancements in the design of ​high-performance lubricating surfaces for demanding industrial applications.