Minglun Ren

Rotating machinery prognostics is essential for ensuring the reliability and operational safety of industrial systems. Although gated recurrent units (GRUs) have achieved competitive performance in remaining useful life (RUL) prediction, whether their internal dynamics are consistent with irreversible degradation mechanisms remains largely unexplored. This study revisits GRU-based prognostics from a physics-consistency perspective and analyzes the potential mismatch between gating mechanisms and degradation evolution. A full-life benchmarking framework is developed based on the XJTU-SY bearing run-to-failure dataset. A training-based health indicator (HI) is constructed through multi-domain vibration feature extraction and principal component analysis, where the degradation-state representation and RUL prediction objective are explicitly distinguished to avoid physically inconsistent supervision. Several representative approaches, including statistical models and deep learning architectures (LSTM, GRU, TCN, and Transformer), are evaluated using both prediction accuracy metrics (RMSE, MAE, and R²) and physical consistency criteria (monotonicity index, monotonicity violation index, and degradation trend consistency). Experimental results demonstrate that superior prediction accuracy does not necessarily guarantee physically consistent degradation modeling. Although GRU provides competitive RUL prediction performance, its hidden-state evolution and gating responses exhibit noticeable non-monotonic behaviors during degradation progression. These findings reveal a potential discrepancy between prediction-oriented recurrent learning mechanisms and irreversible degradation dynamics, highlighting the importance of incorporating physics-consistency evaluation into reliable data-driven prognostic models.

The trajectory prediction of the subject vehicle contributes to the autonomous vehicle’s recognition and evaluation of potentially risky situations in advance, serving as an essential basis for subsequent decision-making and planning. Accurately predicting future trajectories of traffic participants has become a significant issue in enhancing autonomous vehicles’ efficiency and safety. Purely physics-driven or data-driven methods are the main research direction for vehicle lane-changing trajectory prediction on highways at present. However, these methods are either not applicable to complex and highly interactive driving environments or raise concerns about model interpretability and physical implications. To fully utilize the strengths and mitigate the disadvantages of both data-driven and physics-driven models, a physics-informed deep learning (PIDL) model that integrates physical information into a deep learning model is proposed to accurately predict vehicle lane-changing trajectories in this paper. An attention-temporal convolutional network model and a kinematic model are utilized as the data-driven and physics-driven components of the PIDL model, respectively. The HighD dataset is utilized to validate the outstanding performance of the model. The comparison results with baseline models indicate that the model can achieve more precise and interpretable predictions of vehicle lane-changing trajectories. The results of the multi-feature input experiments demonstrate that inputting more feature information on vehicles into the model can improve its prediction performance. The excellent generalization of the model in both temporal and spatial dimensions is verified by the transfer experiments. This PIDL model offers a promising avenue for enhancing the accuracy, interpretability, and generalization of trajectory prediction for autonomous vehicles.

The application of blockchain technology in e‐platforms makes products more reliable and traceable, especially products from third‐party retailers without brand authorization. However, this technology will also intensify the competition between authorized retailers and third‐party retailers. On the other hand, as intermediaries, e‐platforms' profits have also become elusive. We aim to investigate how blockchain technology impacts optimal pricing strategies and profit distribution in a dual‐channel supply chain composed of one manufacturer, one e‐platform, and two e‐tailers. By adopting a Stackelberg game framework, the results show that (1) the e‐platforms may not necessarily benefit from blockchain; only when the commission rate and unit usage cost are both small (or large) enough can the platform achieve higher profit. (2) The application of blockchain can yield extra reputation benefits for third‐party retailer, while the retail prices of both channels may also be higher. (3) The Pareto improvements show that when manufacturer sells products to official retailer and third‐party retailer at different prices, the possibility of successful blockchain implementation increases. (4) When the third‐party retailer shares blockchain usage cost, its optimal channel price rises and demand decreases. For the official retailer, changes in pricing depend on competitive intensity. The manufacturer tends to reduce wholesale price under the strategy, benefiting both e‐tailers and enhancing product competitiveness and supply chain efficiency.

Due to the high cost of obtaining 3D annotations and the accumulation of many 2D datasets with 2D semantic labels, deploying multi-view 2D images for 3D semantic segmentation has attracted widespread attention. Fusion of multi-view information requires establishing local-to-local as well as local-to-global dependencies among multiple views. However, previous methods that are based on 2D annotations supervision cannot model local-to-local and local-to-global dependencies simultaneously. In this letter, we propose a novel multi-view fusion framework with graph neural networks (MVF-GNN) for multi-view interaction and integration. First, a multi-view graph based on the associated pixels in multiple views is constructed. Then, a multi-scale multi-view graph attention network (MSMVGAT) module is introduced to perform graph reasoning on multi-view graphs at different scales. Finally, an attention multi-view graph aggregation (AMVGA) module is introduced to learn the importance of different views and integrate multi-view features. Experiments on the ScanNetv2 benchmark dataset show that our method outperforms state-of-the-art 2D/3D semantic segmentation methods based on 2D annotations supervision.