Search Results (431)

With the accelerated digital transformation of the power industry, critical infrastructures such as power grids are increasingly migrating to cloud-native architectures, leading to unprecedented growth in service scale and complexity. Traditional operation and maintenance (O&M) methods struggle to meet the demands for real-time monitoring, accuracy, and scalability in such environments. This paper proposes a novel service performance anomaly detection system based on large language models (LLMs) and multi-agent systems (MAS). By integrating the semantic understanding capabilities of LLMs with the distributed collaboration advantages of MAS, we construct a high-precision and robust anomaly detection framework. The system adopts a hierarchical architecture, where lower-layer agents are responsible for tasks such as log parsing and metric monitoring, while an upper-layer coordinating agent performs multimodal feature fusion and global anomaly decision-making. Additionally, the LLM enhances the semantic analysis and causal reasoning capabilities for logs. Experiments conducted on real-world data from the State Grid Corporation of China, covering 1289 service combinations, demonstrate that our proposed system significantly outperforms traditional methods in terms of the F1-score across four platforms, including customer services and grid resources (achieving up to a 10.3% improvement). Notably, the system excels in composite anomaly detection and root cause analysis. This study provides an industrial-grade, scalable, and interpretable solution for intelligent power grid O&M, offering a valuable reference for the practical implementation of AIOps in critical infrastructures. Evaluated on real-world data from the State Grid Corporation of China (SGCC), our system achieves a maximum F1-score of 88.78%, with a precision of 92.16% and recall of 85.63%, outperforming five baseline methods. Full article

Model predictive control (MPC) has become a popular method in motor control due to its high adaptability to multivariate control. However, one issue for this control system is constructing a reasonable cost function (CF) and obtaining appropriate weighting factors (WFs) within it. This paper addresses the issue of effectively reducing torque ripple and current harmonic content in permanent magnet synchronous motors (PMSM). Within the three-vector model predictive current control (TV-MPCC) strategy for PMSM, a new CF including current error and switching frequency terms is constructed. Combined with the technique for order preference by similarity to ideal solution (TOPSIS), the optimal control vector is obtained. Compared with traditional methods, this method reduces the complexity of adjusting WFs in the CF. Simulation results show that the motor’s torque ripple and current harmonic content are effectively reduced. Both the steady state and dynamic performance of the PMSM are also improved by means of the proposed multi-objective MPC for current error and switching frequency. Full article

A longstanding ambiguity surrounds the operationalization of consciousness in artificial systems, complicated by the philosophical and cultural weight of subjective experience. This work examines whether cognitive architectures may be designed to support a functionally explicit form of artificial consciousness, focusing not on the replication of phenomenology, but rather on measurable, technically realizable introspective mechanisms. Drawing on a critical review of foundational and contemporary literature, this study articulates a conceptual and methodological shift: from investigating the experiential perspective of agents (“what it is like to be a bat”) to analyzing the informational, self-regulatory, and adaptive structures that enable purposive behavior. The approach combines theoretical analysis with a comparative review of major cognitive architectures, evaluating their capacity to implement access consciousness and internal monitoring. Findings indicate that several state-of-the-art systems already display core features associated with functional consciousness—such as self-explanation, context-sensitive adaptation, and performance evaluation—without invoking subjective states. These results support the thesis that cognitive engineering may progress more effectively by focusing on operational definitions of consciousness that are amenable to implementation and empirical validation. In conclusion, this perspective enables the development of artificial agents capable of autonomous reasoning and self-assessment, grounded in technical clarity rather than speculative constructs. Full article

Land–sea segmentation is a fundamental task in remote sensing image analysis, and plays a vital role in dynamic coastline monitoring. The complex morphology and blurred boundaries of coastlines in remote sensing imagery make fast and accurate segmentation challenging. Recent deep learning approaches lack the ability to model spatial continuity effectively, thereby limiting a comprehensive understanding of coastline features in remote sensing imagery. To address this issue, we have developed VMMT-Net, a novel dual-branch semantic segmentation framework. By constructing a parallel heterogeneous dual-branch encoder, VMMT-Net integrates the complementary strengths of the Mix Transformer and the Visual State Space Model, enabling comprehensive modeling of local details, global semantics, and spatial continuity. We design a Cross-Branch Fusion Module to facilitate deep feature interaction and collaborative representation across branches, and implement a customized decoder module that enhances the integration of multiscale features and improves boundary refinement of coastlines. Extensive experiments conducted on two benchmark remote sensing datasets, GF-HNCD and BSD, demonstrate that the proposed VMMT-Net outperforms existing state-of-the-art methods in both quantitative metrics and visual quality. Specifically, the model achieves mean F1-scores of 98.48% (GF-HNCD) and 98.53% (BSD) and mean intersection-over-union values of 97.02% (GF-HNCD) and 97.11% (BSD). The model maintains reasonable computational complexity, with only 28.24 M parameters and 25.21 GFLOPs, striking a favorable balance between accuracy and efficiency. These results indicate the strong generalization ability and practical applicability of VMMT-Net in real-world remote sensing segmentation tasks. Full article

The high-fidelity lithium-ion battery (LIB) models are crucial for realizing an accurate state estimation in battery management systems (BMSs). Recently, the fractional-order equivalent circuit models (FOMs), as a frequency-domain modeling approach, offer distinct advantages for constructing high-precision battery models in field of electric vehicles. However, the quantitative evaluations and adaptability of these models under different driving cycle datasets are still lacking and challenging. For this reason, comparative evaluations of different FOMs using a novel drive cycle dataset of a battery was carried out in this paper. First, three typical FOMs were initially established and the particle swarm optimization algorithm was then employed to identify model parameters. Complementarily, the efficiency and accuracy of the offline identification for three typical FOMs are also discussed. Subsequently, the terminal voltages of these different FOMs were investigated and evaluated under dynamic operating conditions. Results demonstrate that the FOM-W model exhibits the highest superiority in simulation accuracy, achieving a mean absolute error (MAE) of 9.2 mV and root mean square error (RMSE) of 19.1 mV under Highway Fuel Economy Test conditions. Finally, the accuracy verification of the FOM-W model under two other different dynamic operating conditions has also been thoroughly investigated, and it could still maintain a RMSE and MAE below 21 mV, which indicates its strong adaptability and generalization compared with other FOMs. Conclusions drawn from this paper can further guide the selection of battery models to achieve reliable state estimations of BMS. Full article

A conflict is a situation where multiple stakeholders have different evaluations over possible scenarios or states. Conflict analysis is an essential tool for understanding and resolving complex conflicts, especially in scenarios involving multiple stakeholders and uncertainties. Confrontation analysis (ConAna) and graph model for conflict resolution (GMCR) have been integrated for dilemma reasoning and conflict resolution in region crisis analysis. This paper discusses the application of decision-theoretic rough sets (DTRS) to three-way decisions (3WD) in dilemma reasoning and conflict resolution. Three-way decisions are a strategy for making decisions under uncertain conditions, which compensates for the shortcomings of traditional two-way decisions (such as accept or reject) by introducing a “delayed decision” option. In terms of dilemma reasoning, we try to address incomplete or conflicting information and provide a more reasonable decision path for decision-makers through comprehensive evaluation of multi-criteria. In terms of conflict resolution, the DTRS model seeks a compromising solution that is acceptable to all parties by analyzing the game relationship between different stakeholders. The DTRS model combines decision-making theory and rough set theory to determine the balanced decision region by constructing a game between multiple criteria. This dynamic integration is of great significance for the study of complex international conflicts, providing a cross-disciplinary perspective for related research. In this paper, we demonstrate the application of DTRS in 3WD and discuss the relationship between DTRS and probabilistic rough sets. The research shows that the DTRS model has significant advantages in dealing with complex decision problems and can effectively deal with the conflicts and uncertainties in multi-criteria decision-making. Full article

Pipe-jacking construction technology is favored in urban construction due to its advantages of high safety and being a non-excavation technique. However, instability of the tunnel face often occurs due to unfavorable conditions, such as pipe jacking across the river channel, shallow soil cover, and improper control of the support pressure. In this study, we made a use of the limit balance method and mud–water balance theory. At this moment of passive damage and active destruction occurring at the pipe-jacking tunnel face, the general mathematical expressions of the tunnel-face support pressure (with lower limit value P_min and upper limit value P_max) are derived. In the non-river impact area and river impact area, the optimal value P_o of support pressure at the tunnel face is thus derived. Then, based on the Y25-Y26 pipe-jacking project across the Chu River channel in Hefei North District, a numerical simulation method is used to support further discussion. The results indicate that, when the river overburden is 3 m, the ultimate support pressure calculated by means of numerical simulation is 881.786 kN, and the optimal support ratio $\mathsf{\lambda }$ is taken in the interval of 1.0~1.5. Secondly, the upper limit value P_max, lower limit value P_min, and optimum value P_o calculated using the theoretical equations are 2669.977 kN, 309.910 kN, and 1044.870 kN, respectively. These results leads us to recommend setting the support pressure of the tunnel face in a reasonable range between the upper limit value P_max and the lower limit value P_min, to ensure that the tunnel-face support pressure and resistance during pipe jacking always remain in a balanced state. The relevant research results from this study provide an important technical guarantee for the successful implementation of the examined project and, at the same time, can serve as a reference example for similar projects. Full article

Gesture recognition based on conventional machine learning is the main control approach for advanced prosthetic hand systems. Its primary limitation is the need for feature extraction, which must meet real-time control requirements. On the other hand, deep learning models could potentially overfit when trained on small datasets. For these reasons, we propose a hybrid Linear Discriminant Analysis–convolutional neural network (LDA-CNN) framework to improve the gesture recognition performance of sEMG-based prosthetic hand control systems. Within this framework, 1D-CNN filters are trained to generate latent representation that closely approximates Fisher’s (LDA’s) discriminant subspace, constructed from handcrafted features. Under the train-one-test-all evaluation scheme, our proposed hybrid framework consistently outperformed the 1D-CNN trained with cross-entropy loss only, showing improvements from 4% to 11% across two public datasets featuring hand gestures recorded under various limb positions and arm muscle contraction levels. Furthermore, our framework exhibited advantages in terms of induced spectral regularization, which led to a state-of-the-art recognition error of 22.79% with the extended 23 feature set when tested on the multi-limb position dataset. The main novelty of our hybrid framework is that it decouples feature extraction in regard to the inference time, enabling the future incorporation of a more extensive set of features, while keeping the inference computation time minimal. Full article

Current research on intelligent question answering mainly focuses on high-resource languages such as Chinese and English, with limited studies on question understanding and reasoning in low-resource languages. In addition, during the joint modeling of question understanding tasks, the interdependence among subtasks can lead to error accumulation during the interaction phase, thereby affecting the prediction performance of the individual subtasks. To address the issue of error propagation caused by sentence-level intent encoding in the joint modeling of intent recognition and slot filling, this paper proposes a Cross-lingual Token-level Bi-Interactive Model (Bi-XTM). The model introduces a novel subtask interaction method that leverages the token-level intent output distribution as additional information for slot vector representation, effectively reducing error propagation and enhancing the information exchange between intent and slot vectors. Meanwhile, to address the scarcity of Kazakh (Arabic alphabet) language corpora, this paper constructs a cross-lingual joint question understanding dataset for the Xinjiang tourism domain, named JISD, which includes 16,548 Chinese samples and 1399 Kazakh samples. This dataset provides a new resource for cross-lingual intent recognition and slot filling joint tasks. Experimental results on the publicly available multi-lingual question understanding dataset MTOD and the newly constructed dataset demonstrate that the proposed Bi-XTM achieves state-of-the-art performance in both monolingual and cross-lingual settings. Full article

This study constructs an index system based on provincial data from 2012 to 2023 for forestry industry agglomeration and eco-efficiency. Using methods such as the Coupling Coordination Degree and Relative Development Degree, the study explores the relationship between the coupled and coordinated development of forestry industry agglomeration and eco-efficiency at the provincial level, and introduces a balanced interval to regulate the coupling and coordination process between forestry industry agglomeration and eco-efficiency. The results indicate that: (1) During the study period, the overall coupled and coordinated development of China’s forestry industry agglomeration and eco-efficiency was in an antagonistic stage, with the development of forestry industry agglomeration lagging behind the level of eco-efficiency. (2) The Relative Development Degree of forestry industry agglomeration and eco-efficiency shifted from a “ladder” pattern to an “hourglass” pattern. (3) The process of coupling and coordinating the development of forestry industry agglomeration and eco-efficiency exhibited fluctuations, indicating that future efforts should focus on improving the quality of both forestry industry agglomeration and eco-efficiency to promote coordinated development. (4) During the period from 2012 to 2023, China’s forestry industry agglomeration and eco-efficiency generally failed to simultaneously reach a reasonably balanced state, with notable regional differences. Factors such as the number of non-forestry employees, geographic location, and environmental conditions significantly impacted the balance between forestry industry agglomeration and eco-efficiency. Full article

The value network of the high-tech park constitutes a value co-creation system where multiple entities facilitate knowledge transformation through interaction, thereby achieving collaborative innovation. The reasonable distribution of collaborative innovation benefits among various innovation entities is a critical factor in maintaining the motivation for innovation within the value network. This study examines the co-creation mechanism of the value network in high-tech parks from the perspective of knowledge innovation, with the aim of enhancing the efficiency of knowledge transfer and spillover among entities. Additionally, it seeks to establish a fairer and more rational benefit distribution framework to promote collaborative innovation and ensure the stable operation of the value network. Firstly, we identify the entities involved in value co-creation within the high-tech park. Subsequently, we analyze the roles and interrelationships of these entities within the value co-creation network. We determine the knowledge flow pathways by employing the shortest path method, and innovatively construct an MMPP/M/C queuing model to depict the processes of knowledge transfer and spillover among the entities engaged in value co-creation. We optimize and solve the queuing model using the matrix geometric method, deriving metrics such as the average queue length, average arrival rate, average waiting time, and service intensity under the steady state of the system, and verify the applicability and effectiveness of the model in the application of the high-tech park through empirical data. Finally, by integrating the improved Shapley value method, a benefit distribution model is constructed that incorporates five types of factors: contribution level, resource input, knowledge spillover effect, effort level, and risk undertaking. The rationality and operability of this model are validated through computational examples. Research findings indicate that the optimized queuing model enhances the efficiency of knowledge transfer and spillover among entities, while the refined benefit distribution mechanism effectively compensates entities with high contribution levels, substantial resource inputs, significant knowledge spillover effects, elevated effort levels, and high risk assumption levels. This provides both theoretical support and practical guidance for sustaining the long-term stable operation of the value network. Full article

In many extended object tracking applications (e.g., tracking vehicles using a millimeter-wave radar), the shape of an extended object (EO) remains unchanged while the orientation angle varies over time. Thus, tracking the shape and the orientation angle as individual parameters is reasonable. Moreover, the tight coupling between the orientation angle and the heading angle contains information on improving estimation performance. Hence, this paper proposes a constrained filtering approach utilizing this information. First, an EO model is built using an orientation vector with a heading constraint. This constraint is formulated using the relation between the orientation vector and the velocity vector. Second, based on the proposed model, a variational Bayesian (VB) approach is proposed to estimate the kinematic, shape, and orientation vector states. A pseudo-measurement is constructed from the heading constraint and is incorporated into the VB framework. The proposed approach can also address the ambiguous issue in orientation angle estimation. Simulation and real-data results are presented to illustrate the effectiveness of the proposed model and estimation approach. Full article

Residual stress (RS) is one of the main reasons for component failure during an additive manufacturing (AM) process, especially using selective laser melting (SLM) technology. This paper reviews RS’s investigation methods, formation mechanisms and regularities of distribution. When considering recent research progress, studies indicate that the dominant stress is primarily attributed to thermal stress induced by significant laser temperature gradients during the rapid melting and forming process, which subsequently transforms into RS upon cooling to room temperature, as verified by simulation and experiments. Then, the distribution regularities of RS are analyzed. SLM RS gradually increases when it is measured from the surface layer to the substrate. In the plane direction, at the center and edge of the part, tensile stresses are found; as for the middle area, which is the transition area of compressive stress, the whole plane stress remains in an equilibrium state. Based on the forementioned conclusions, the three-dimensional distribution diagram of RS on the sample was constructed. Finally, the strategic approaches for stress mitigation are briefly discussed. The excessive stress in forming can be reduced by process parameter matching, and the RS can be greatly remitted by pre-treatment/post-treatment, so as to improve the quality of formed parts. This review provides a valuable theoretical basis for practical applications of SLM. Full article

Several researchers recently demonstrated that attackers can interfere with an inertial measurement unit (IMU) sensor’s normal function or take complete control of sensor measurements by physically injecting malicious signals into the sensor. Although there are existing methods for detecting such signal injection attacks, most do not provide resilience. Indeed, detection-only methods cannot respond when attacks have already occurred, which results in accidents such as crashes or falls. In this paper, we propose the first method that can detect signal injection attacks on IMU sensors based on the relation between the gyroscope and the magnetometer, and provide long-term resilience against these attacks. We construct a mathematical model to estimate one sensor’s data from the other’s data based on their relation. With this mathematical model, the device can detect signal injection attacks on the IMU sensor and continue to function in a near-normal state based on the estimated data. Our method can be easily adapted to deployed devices since it requires only estimation software and no additional hardware. We evaluated our method using a total of five open datasets and commercial devices. Our method has a resilience of 99.78% against signal injection attacks while consuming only reasonable computational costs. Full article

Red melilite structure CaY_1−xAl₃O₇: Eu_x (x = 0.04–0.24) phosphors for white LEDs were synthesized through a straightforward solid-state reaction process. These phosphors exhibit efficient excitation under near-ultraviolet light at 398 nm (⁷F₀ → ⁵L₆), producing the desired emission peak at 622 nm from the transitions of ⁵D₀ → ⁷F₂. The Eu doping concentration was also optimized as x = 0.16. The complete 3003 × 3003 energy matrix was constructed based on an effective Hamiltonian including both free-ion and crystal-field interactions within a complete diagonalization method (CDM). Eighteen experimental fluorescent spectra for Eu³⁺ ions at the Y³⁺ site of CaYAl₃O₇ crystal were quantitatively identified with high accuracy through fitting calculations. The fitting values are in reasonable agreement with the experimental results, thereby showcasing the efficacy of the CDM in probing luminescent phosphors for white LEDs. Full article