摘要 : Traditional Simultaneous Localization and Mapping (SLAM) systems assume static environments, and their localization accuracy degrades significantly when encountering dynamic objects. Although dynamic SLAM based on semantic information can handle simple dynamic objects, it fails to address the impact of pseudo-dynamic objects and unrecognized dynamic objects on pose estimation. To address these issues, we propose a DC-SLAM system, comprehensively eliminating two types of dynamic features while preserving valid static features. The system implements a dual-category mechanism: for active dynamic features, it combines a dynamic object detection network with dense optical flow for detection and removal, preventing excessive elimination of static features' for passive dynamic features, the system first determines their attributes using multi-view geometry, then clusters all feature points based on depth information and object detection categories, and finally optimizes dynamic properties through Mahalanobis distance analysis of outlier similarity. This approach compensates for the limitations of multi-view geometry, enabling more effective suppression of passively dynamic objects. Additionally, an octree mapping module is developed to assist mobile robots in scene understanding for practical applications. Extensive experiments on the TUM dataset, Bonn dataset, and real-world dynamic scenes verify DC-SLAM's effectiveness. The results demonstrate significant improvements, compared with ORB-SLAM2, DC-SLAM reducing Absolute Trajectory Error (ATE) by 98.86 % and Relative Pose Error (RPE) by 97.28 %, while enabling the reliable construction of octree maps to enhance spatial understanding.

摘要 : While scholars and managers have long queried how a review message can be useful, the e-word-of-mouth (eWOM) literature yields inconclusive findings. Our study approaches this classic research question from the novel angle of diversity and investigate how topic diversity of online reviews affect review usefulness. Specifically, topic diversity manifests in both topic variety, the extent to which different topics are discussed in a review message, and content distribution across topics, the extent to which content is distributed across different topics. We leverage the aspect-based topic modeling technique to uncover the extent of different topics discussed in individual review messages as the nested subunits and operationalize topic variety and content distribution as two main unit-level predictors of review usefulness. This text-based approach helps analyze about 200k restaurant reviews from Yelp.com and generates six different topics relating to restaurant reviews. Results reveal that (1) on average, topic variety positively affected review usefulness, while content distribution across topics did not; (2) however, when topic variety is high (low), the impact of content distribution on review usefulness is significant and positive (significant and negative). Our study contributes to the e-WOM literature by offering new knowledge regarding topic diversity and review usefulness.

摘要 : This study employs a low-temperature solvothermal strategy to fabricate carbon nitride nanorod (LCM t ) photocatalysts, aiming to overcome the challenges of high energy consumption and morphology control in photocatalyst design. By precisely regulating the solvothermal reaction time, a synergistic optimization of morphology, elemental composition, and optical absorption properties is achieved. Density functional theory (DFT) calculations reveal that C O and Cl− groups significantly contribute to the conduction band energy level, whereas the effects of −OH and C−Cl groups can be neglected. Bader charge and charge difference density distribution indicate that the introduction of C O and Cl− groups remarkably change the electronic structure of carbon nitride, triggering intramolecular charge redistribution and strong electron transfer effects. Furthermore, DFT combined with in-situ XPS analysis results confirm that C O and Cl− groups act as dominant active sites, promoting O 2 adsorption and markedly lowering the energy barrier for the formation of the key intermediate OOH during photocatalytic H 2 O 2 production. Finally, the optimal LCM t achieves H 2 O 2 yield of 203.5 μM h −1 under visible light. This work presents an efficient, metal-free photocatalyst for sustainable H 2 O 2 generation and offers valuable insights into the rational design of advanced photoreactive nanomaterials.

摘要 : In this paper, we present an innovative method for the fabrication of multilayer microfluidic analytical devices (μPADs) in a single piece of paper. It involves hand-tearing a single piece of industrial oil filter paper (Sc-y102) into multiple layers. Subsequently, we employ a customized rubber stamp to print PDMS hydrophobic barriers on each layer, finally creating a functional and multilayered 3D microfluidic paper-based analytical device (3D μPAD). Parameter investigations showed an optimal resolution of 1200 μm for the hydrophilic channel and 840 μm for the hydrophobic barrier. Through using hand-tearing method, we can separate a single sheet of paper into up to six layers, each with a uniform thickness of ∼100 μm. Based on this method, we design and create a variety of single-layer 2D μPADs and multilayer 3D μPADs, and all derived from a single sheet of paper. Notably, the hydrophobic PDMS barriers, outperformed the conventional wax barriers in terms of chemical resistance, are able to effectively block the penetration of solvents and surfactants. Then, a 3D μPAD aimed for the timed detection of glucose is developed, which features an innovative multilayered design that precisely controls the timing of fluid flow to different detection zones. Finally, we designed a multilayer 3D μPAD that can achieve accurate and reliable blood typing by utilizing red blood cell agglutination and vertical flow filtration through single-sheet multiple microchannels. This hand-tearing approach expands the fabrication process for monolithic μPADs and provides a new way of realizing more complex and multifunctional device designs, may triggering more in-depth investigations into the monolithic devices and expanding research in the field.

摘要 : Clustering techniques face persistent challenges in balancing automation with human interpretability. Traditional methods require laborious parameter tuning and domain expertise to define similarity measures and validate results, while deep learning approaches trade transparency for performance. To bridge this gap, we propose a human-in-the-loop framework that synergizes domain knowledge with graph-based semi-supervised learning. Our system enables users to iteratively refine clusters through intuitive visual adjustments on a subset of data, guided by real-time quality metrics to reduce errors and decision fatigue. These sparse annotations propagate to unlabeled instances via a graph neural network (GNN) that models latent relationships through modularity-driven structural learning. By translating cluster adjustments into semi-supervised classification tasks, our method eliminates manual feature engineering and scales to large datasets without retraining. Evaluations on two subsets of the MNIST dataset demonstrated that the NMI (Normalized Mutual Information) of our method improved by 50.44% and 64.77% relative to baseline clustering method, respectively.

摘要 : The objective of this study is to analyze energy conversion in two configurations of piezoelectric material placement in acoustic black holes. These structures concentrate vibrational energy due to the gradual reduction in thickness, making them ideal for energy harvesting. In the first configuration, piezoelectric materials are placed at the outer edges of the hole; in the second, at the inner edges. The material is applied only to specific regions, rather than covering the entire inner or outer edge. The same amount of piezoelectric material is used in both cases, being able to act as both a vibration damper and an energy harvester. This study investigates the optimal position for piezoelectric material placement, comparing energy conversion at the outer vs. inner edges of a central elliptical hole. The finite element method was used to discretize the structural domain, considering elliptical hole geometries. Dynamic structural analysis was applied to compute energy distributions and conversions. The results showed that the placement of the piezoelectric material influences energy conversion, with the most suitable position being along the outer edge of the hole. These findings reinforce the importance of optimal piezoelectric placement for maximizing energy harvesting in structures with acoustic black holes.

摘要 : In the context of addressing the multi-objective vehicle routing problem, a hybrid time window multi-objective vehicle model was established using integer programming and the intelligent water drop algorithm. To overcome the limitation of the intelligent water drop algorithm potentially converging to local optimal solutions, enhancements were proposed through genetic algorithms, particularly by introducing genetic crossover and single-point recombination operators. Subsequently, the intelligent water drop algorithm was refined, and its effectiveness was evaluated through a real-world case study. Comparative analyses were conducted among three algorithms: IWD, GA, and SA. The results demonstrate that the improved algorithm effectively alleviates the common issue of traditional algorithms converging to local optimal solutions. Therefore, an enhanced solution is provided for the discrete hybrid time window problem, achieving superior optimization outcomes.

摘要 : This study addresses the challenge of designing an event-triggered observer for neural network-enhanced sliding mode control in nonlinear Takagi–Sugeno fuzzy Markov jump systems, where premise variables are not directly measurable. Firstly, for the purpose of state observer design, a dynamic event-triggered mechanism integrated with a neural network-based compensator is developed. Secondly, through the construction of an integral sliding surface, the dynamic behaviors of both the sliding mode and the error system are formulated, incorporating estimated premise parameters. Thirdly, rigorous stochastic stabilization criteria are established, incorporating H ∞ disturbance attenuation with a specified level γ, while accounting for transition rates with general uncertainty characteristics. Subsequently, a fuzzy adaptive sliding mode control scheme is synthesized to ensure finite-time convergence of the system states to the predefined sliding surface. Finally, the effectiveness of the proposed control strategy is thoroughly validated through high-fidelity numerical simulations on a practical example.

摘要 : Stimulus-responsive afterglow materials refer to a class of substances whose afterglow characteristics alter under external stimuli, showing considerable potential for advanced applications in anti-counterfeiting, optoelectronic displays, chemical sensing, and bioimaging. Carbon dots (CDs), as an emerging category of afterglow materials, have garnered significant attention due to their stable photophysical and chemical properties, low toxicity, and tunable luminescent energy bands. In recent years, significant progress has been made in the development of stimulus-responsive afterglow CDs, underscoring the need for a systematic summary of this rapidly advancing field. This review summarizes recent advances in CD-based afterglow, encompassing luminescence mechanisms and synthesis strategies. A particular focus is placed on the types of stimulus-responsive afterglow behaviors in CDs, their influence on afterglow performance, and the underlying response mechanisms. The potential applications of these stimulus-responsive afterglow CDs in sensing and information encryption are also discussed in detail. Finally, current challenges and future prospects are outlined, aiming to guide the rational design and development of next-generation stimulus-responsive afterglow CDs.

摘要 : Few-shot action recognition is a challenging yet practically significant problem that involves developing a model capable of learning discriminative features from a small number of labeled samples to recognize new action categories. Current methods typically infer spatial relationships either within or across skeletons to learn action representations, but this often results in features with insufficient discriminability and ineffective attention to critical body parts. To address these limitations, we propose DAF-Net, a novel framework that employs focal attention to jointly model intra-skeleton and inter-skeleton relationships, enhancing discriminative feature learning in few-shot skeleton-based action recognition. Unlike traditional methods that focus solely on intra-skeleton dependencies or inter-skeleton structures, DAF-Net dynamically integrates both components via focal attention, enhancing key body part representation and refining features, particularly in data-scarce conditions. Furthermore, DAF-Net incorporates an enhanced prototype generation strategy, optimizing class prototype formation via cosine similarity weighting to further improve feature discriminability in multi-shot scenarios. In temporal matching, cosine similarity evaluates local feature similarity within skeleton sequences, capturing directional variations of specific joints over time. Extensive experiments on three benchmark datasets (NTU-T, NTU-S, and Kinetics-skeleton) confirm significant performance gains, validating the effectiveness of DAF-Net.