Search Results (1,443)

Existing unsupervised anomaly detection methods suffer from insufficient parameter precision, poor robustness to noise, and limited generalization capability. To address these issues, this paper proposes an Adaptive Diffusion Adversarial Evolutionary Network (ADAEN) for unsupervised anomaly detection in tabular data. The proposed network employs an adaptive hierarchical feature evolution generator that captures multi-scale feature representations at different abstraction levels through learnable attribute encoding and a three-layer Transformer encoder, effectively mitigating the gradient vanishing problem and the difficulty of modeling complex feature relationships that are commonly observed in conventional generators. ADAEN incorporates a multi-scale adaptive diffusion-augmented discriminator, which preserves scale-specific features across different diffusion stages via cosine-scheduled adaptive noise injection, thereby endowing the discriminator with diffusion-stage awareness. Furthermore, ADAEN introduces a multi-scale robust adversarial gradient loss function that ensures training stability through a diffusion-step-conditional Wasserstein loss combined with gradient penalty. The method has been evaluated on 14 UCI benchmark datasets and achieves state-of-the-art performance in anomaly detection compared to existing advanced algorithms, with an average improvement of 8.3% in AUC, an 11.2% increase in F1-Score, and a 15.7% reduction in false positive rate. Full article

Connected and autonomous vehicle (CAV) platoons face the dual challenge of maintaining longitudinal formation stability while ensuring lateral safety in dynamic traffic environments, yet existing control approaches often address these objectives in isolation. This paper proposes a hierarchical cooperative control framework that integrates a differential game-based longitudinal controller with a risk potential field-driven model predictive controller (MPC) for lateral motion. At the coordination control layer, a differential game formulation models inter-vehicle interactions, with analytical solutions derived for both open-loop Nash equilibrium under predecessor-following (PF) topology and an estimated Nash equilibrium under two-predecessor-following (TPF) topology. The motion control layer employs a risk potential field model that quantifies collision threats from surrounding obstacles and road boundaries, guiding the MPC to perform real-time trajectory optimization. A comprehensive co-simulation platform integrating MATLAB/Simulink, Prescan, and CarSim validates the proposed framework across three representative scenarios: ramp merging with aggressive cut-in maneuvers, emergency braking by a preceding obstacle vehicle, and multi-lane cooperative obstacle avoidance involving multiple dynamic obstacles. Across all scenarios, the CAV platoon achieves safe obstacle avoidance through autonomous decision-making, with spacing errors converging to zero and smooth velocity adjustments that ensure both formation stability and ride comfort. The results demonstrate that the proposed framework effectively adapts to diverse and complex traffic conditions. Full article

Whether learners can correctly complete exercises is influenced by multiple factors, including their mastery of relevant knowledge concepts and the interdependencies among these concepts. To investigate how the structure of the knowledge space—particularly the complex relationships among learners, exercises, and knowledge points—affects learning outcomes, this study proposes the Hierarchical Heterogeneous Graph Knowledge Tracing model (HHGKT). A hierarchical heterogeneous graph was constructed to capture two types of interactions—“learner–knowledge concept” and “exercise–knowledge concept”—and incorporate the interdependencies among knowledge concepts into the graph structure. By leveraging this hierarchical representation, the model’s ability to characterize learners and exercises was enhanced. A hierarchical heterogeneous graph encompassing users, exercises, and knowledge concepts was built based on the ASSISTments dataset, and simulation experiments were conducted. The results indicate that the proposed structure effectively represents the complexity of the knowledge space. Incorporating knowledge concept interdependencies improves prediction accuracy by 1.79%, while the hierarchical heterogeneous graph outperforms traditional bipartite graphs by approximately 1.5 percentage points in accuracy. These findings demonstrate that the model better integrates node and relational information, offering valuable insights for knowledge space modeling and its application in educational contexts. Full article

Legged robot locomotion remains a critical challenge in robotics, demanding control strategies that are not only dynamically stable and robust but also capable of adapting to complex and changing environments. deep reinforcement learning (DRL) has recently emerged as a powerful approach to automatically generate motion control policies by learning from interactions with simulated or real environments. This study provides a systematic overview of DRL applications in legged robot control, emphasizing experimental platforms, measurement techniques, and benchmarking practices. Following PRISMA guidelines, 27 peer-reviewed studies published between 2018 and 2025 were analyzed, covering model-free, model-based, hierarchical, and hybrid DRL frameworks. Our findings reveal that reward shaping, policy representation, and training stability significantly influence control performance, while domain randomization and dynamic adaptation methods are essential for bridging the simulation-to-real-world gap. In addition, this review highlights instrumentation approaches for evaluating algorithm effectiveness, offering insights into sample efficiency, energy management, and safe deployment. The results aim to guide the development of reproducible and experimentally validated DRL-based control systems for legged robots. Full article

Model predictive control (MPC) has become one of the most promising control strategies in the field of electric vehicle energy management due to its rolling optimization and explicit constraint processing capabilities. This study analyzes the modeling mechanism and implementation path of MPC in power allocation, regenerative braking and energy collaborative control, which elaborates on the improvement principle of energy efficiency and system stability through predictive modeling and dynamic optimization. The evolution of MPC application in hybrid power systems, vehicle dynamic stability control, and hierarchical optimization control is discussed. The synergistic effect of multi-objective optimization and health-conscious control in energy efficiency improvement and service life extension is analyzed. With the development of artificial intelligence technology, MPC is expanding from model-based deterministic control to the directions of intelligent learning and distributed adaptation. Model uncertainty, computational complexity, and real-time solving efficiency are the main challenges faced by MPC. Future research will focus on the deep integration of model simplification, rapid solving, and intelligent learning to achieve a more efficient and reliable intelligent energy management system. Full article

Ship detection in optical images is critical for maritime supervision but faces challenges from scale variations and complex backgrounds. Existing detectors often struggle to balance global context modeling with computational efficiency. To address this, we propose Contextual Efficient Hierarchical DETR (CEH-DETR), an efficient framework for multi-scale ship detection. First, we introduce the Cross-stage Parallel State Space Hidden Mixer (CPSHM) backbone, integrating State Space Models with CNNs to capture global dependencies with linear complexity. Second, the Efficient Adaptive Feature Integration (EAFI) module reduces attention complexity to linear using Token Statistics-based Attention. Third, the Hierarchical Attention-guided Feature Pyramid Network (HAFPN) effectively fuses multi-scale features while preserving spatial details. Experiments on the ABOships dataset demonstrate that CEH-DETR achieves a superior balance between accuracy and efficiency. Relative to the baseline RT-DETR, our approach achieves a parameter reduction of 25.6% while increasing mAP@50 by 2.0 percentage points and boosting inference speed to 133.7 FPS (+112.1%), making it highly suitable for real-time maritime surveillance. Full article

Background and Objectives: Public health emergencies, such as the COVID-19 pandemic, significantly impact individuals and families, particularly in educational settings. School closures and changes in daily routines reduced students’ opportunities for learning and social interaction, affecting their mental health. Caregivers also faced increased responsibilities and stressors. This study aimed to evaluate a predictive model of mental health outcomes—specifically posttraumatic stress symptoms (PTSSs) and posttraumatic growth (PTG)—in Chilean schoolchildren and their caregivers. Materials and Methods: A total of 489 students (48% female sex; aged 10–17) from educational communities in various Chilean cities participated in the study, along with their caregivers (aged 21–69; 86.5% female), including mothers, fathers, and guardians. Mental health variables were assessed through self-report instruments. Hierarchical linear regression and path analyses were used to evaluate predictive models for PTSSs and PTG in students. Results: The model predicting PTSSs in students was significant. Key predictors included female sex, aggressive behavior, coping strategies such as keeping problems to oneself, cognitive avoidance, and intrusive rumination, and caregiver PTSSs. The model for PTG was also significant, with predictors including active problem-solving, communication, a positive attitude, and deliberate rumination. These results indicate distinct psychological processes underlying negative and positive outcomes following trauma. Conclusions: The findings underscore the complexity of mental health outcomes among school-aged children and the influence of caregiver well-being. The study highlights the importance of supporting both students and caregivers through targeted interventions. Multi-level strategies addressing emotional regulation, communication, and coping mechanisms may foster resilience and psychological growth in educational communities facing the aftermath of public health emergencies. Full article

LNG serves as a pivotal element within integrated energy systems, especially in coastal regions where the implementation of a stable and reliable LNG cold energy power generation system significantly elevates energy efficiency. This system can effectively meet concurrent demands for cold energy utilization and electricity supply while contributing to the mitigation of carbon emissions. However, the inherent complexity of the system coupled with the scarcity of historical operational data for the novel dual-Rankine cycle process renders conventional reliability assessment methodologies inadequate. This study proposes an integrated framework utilizing fuzzy Bayesian methods to address data scarcity during the early stages of equipment deployment. A hierarchical risk factor model, incorporating process decomposition, expert evaluations, and triangular fuzzy numbers, is developed to quantify uncertainties in failure probabilities. The Bayesian network models the causal relationships among equipment failure factors, allowing for the inference of overall system reliability from individual equipment performance. Through a case study of a LNG terminal in Zhoushan, this approach integrates sensitivity analysis with forward-backward reasoning methodologies to rigorously evaluate and quantify system reliability under operational conditions. The results show that under high load conditions within the 1000 h prior to overhaul, following long-term accumulated operation, the probability of complete system shutdown in the power generation system is 3.30%, while the probability of the LNG cold energy power generation system failing to operate fully due to aging-related faults is 8.24%, demonstrating the system’s strong reliability under extreme conditions. Critical risks identified through backward inference include the seawater pump SWP1, with a posterior failure probability of 59.92% during complete shutdown, and the propane-side pump SWP3, with a posterior failure probability of 32.29% when the cold energy power generation system can only operate in a single-cycle mode. This study provides an advanced methodological framework for risk management in newly constructed LNG cold energy power generation systems, playing a crucial role in promoting sustainable, low-carbon technologies in the energy sector. Full article

Transformers have recently been applied to change detection (CD) of multitemporal remote sensing images because of their ability to model global information. However, the rigid patch partitioning in vanilla self-attention destroys spatial structures and consistency in observed scenes, leading to limited CD performance. In this paper, we propose a novel dual-scale pixel aggregation transformer (DSPA-Former) to mitigate this issue. The core of DSPA-Former lies in a dynamic superpixel tokenization strategy and bidirectional dual-scale interaction within the learned feature space, which preserves semantic integrity while capturing long-range dependencies. Specifically, we design a hierarchical decoder that integrates multiscale features through specialized mechanisms for pixel superpixel dialogue, guided feature enhancement, and adaptive multiscale fusion. By modeling the homogeneous properties of spatial information via superpixel segmentation, DSPA-Former effectively maintains structural consistency and sharpens change boundaries. Comprehensive experiments on the LEVIR-CD, WHU-CD, and CLCD datasets demonstrate that DSPA-Former achieves superior performance compared to state-of-the-art methods, particularly in preserving the structural integrity of complex change regions. Full article

Accurate and energy-efficient localization of autonomous underwater vehicles (AUVs) remains a fundamental challenge due to the complex, bandwidth-limited, and highly dynamic nature of underwater acoustic environments. This paper proposes a fully adaptive deep reinforcement learning (DRL)-driven localization framework for AUVs operating in Underwater Acoustic Sensor Networks (UAWSNs). The localization problem is formulated as a Markov Decision Process (MDP) in which an intelligent agent jointly optimizes beacon selection and transmit power allocation to minimize long-term localization error and energy consumption. A hierarchical learning architecture is developed by integrating four actor–critic algorithms, which are (i) Twin Delayed Deep Deterministic Policy Gradient (TD3), (ii) Soft Actor–Critic (SAC), (iii) Multi-Agent Deep Deterministic Policy Gradient (MADDPG), and (iv) Distributed DDPG (D2DPG), enabling robust learning under non-stationary channels, cooperative multi-AUV scenarios, and large-scale deployments. A round-trip time (RTT)-based geometric localization model incorporating a depth-dependent sound speed gradient is employed to accurately capture realistic underwater acoustic propagation effects. A multi-objective reward function jointly balances localization accuracy, energy efficiency, and ranging reliability through a risk-aware metric. Furthermore, the Cramér–Rao Lower Bound (CRLB) is derived to characterize the theoretical performance limits, and a comprehensive complexity analysis is performed to demonstrate the scalability of the proposed framework. Extensive Monte Carlo simulations show that the proposed DRL-based methods achieve significantly lower localization error, lower energy consumption, faster convergence, and higher overall system utility than classical TD3. These results confirm the effectiveness and robustness of DRL for next-generation adaptive underwater localization systems. Full article

Large Language Model (LLM) agents depend heavily on multiple external tools such as APIs, databases and computational services to perform complex tasks. However, these tool executions create latency and introduce costs, particularly when agents handle similar queries or workflows. Most current caching methods focus on LLM prompt–response pairs or execution plans and overlook redundancies at the tool level. To address this, we designed a multi-level caching architecture that captures redundancy at both the workflow and tool level. The proposed system integrates four key components: (1) hierarchical caching that operates at both the workflow and tool level to capture coarse and fine-grained redundancies; (2) dependency-aware invalidation using graph-based techniques to maintain consistency when write operations affect cached reads across execution contexts; (3) category-specific time-to-live (TTL) policies tailored to different data types, e.g., weather APIs, user location, database queries and filesystem and computational tasks; and (4) session isolation to ensure multi-tenant cache safety through automatic session scoping. We evaluated the system using synthetic data with 2.25 million queries across ten configurations in fifteen runs. In addition, we conducted four targeted evaluations—write intensity robustness from 4 to 30% writes, personalized memory effects under isolated vs. shared cache modes, workflow-level caching comparison and workload sensitivity across five access distributions—on an additional 2.565 million queries, bringing the total experimental scope to 4.815 million executed queries. The architecture achieved 76.5% caching efficiency, reducing query processing time by 13.3× and lowering estimated costs by 73.3% compared to a no-cache baseline. Multi-tenant testing with fifteen concurrent tenants confirmed robust session isolation and 74.1% efficiency under concurrent workloads. Our evaluation used controlled synthetic workloads following Zipfian distributions, which are commonly used in caching research. While absolute hit rates vary by deployment domain, the architectural principles of hierarchical caching, dependency tracking and session isolation remain broadly applicable. Full article

This comprehensive review examines hierarchical control principles and frameworks for grid-connected microgrids operating in environments prone to load shedding and under demand response. The particular emphasis is on South Africa’s current electricity grid issues, experiencing regular planned and unplanned outages, due to numerous factors including ageing and underspecified infrastructure, and the decommissioning of traditional power plants. The study employs a systematic literature review methodology following PRISMA guidelines, analysing 127 peer-reviewed publications from 2018–2025. The investigation reveals that conventional microgrid controls require significant adaptation to address the unique challenges brought about by scheduled power outages, including the need for predictive–proactive strategies that leverage known load-shedding schedules. The paper identifies three critical control layers of primary, secondary, and tertiary and their modifications for resilient operation in environments with frequent, planned grid disconnections alongside renewables integration, regular supply–demand balancing and dispatch requirements. Hybrid optimisation approaches combining model predictive control with artificial intelligence show good promise for managing the complex coordination of solar–storage–diesel systems in these contexts. The review highlights significant research gaps in standardised evaluation metrics for microgrid resilience in load-shedding contexts and proposes a novel framework integrating predictive grid availability data with hierarchical control structures. South African case studies demonstrate techno-economic advantages of adapted control strategies, with potential for 23–37% reduction in diesel consumption and 15–28% improvement in battery lifespan through optimal scheduling. The findings provide valuable insights for researchers, utilities, and policymakers working on energy resilience solutions in regions with unreliable grid infrastructure. Full article

A hierarchical cooperative control strategy guided by a virtual leader is proposed to enhance the speed regulation and robustness of dual permanent magnet synchronous motor (PMSM) systems. The upper layer employs a virtual leader with model predictive speed control (MPSC) to achieve coordinated tracking, while the lower layer utilizes model predictive current control (MPCC) for regulation. A theoretical complexity analysis demonstrates that this decoupled architecture reduces the computational burden by approximately 75% compared to centralized MPC. Furthermore, a load disturbance observer is designed to estimate and compensate for external torques. Simulation and experimental results, covering both forward and reverse rotations, validate the effectiveness of the proposed strategy. Comparative results show that, compared with a conventional PI controller, the proposed method reduces speed overshoot by approximately 20% under sudden load changes, exhibiting superior steady-state performance and strong robustness against load variations. Full article

This paper addresses the problem of cooperative coverage control for heterogeneous Autonomous Underwater Vehicle (AUV) swarms operating in complex underwater environments. The objective is to achieve optimal coverage of a target region while simultaneously ensuring collision avoidance—both among AUVs and with static obstacles—and satisfying the inherent dynamic constraints of the AUVs. To this end, we propose a hierarchical control framework that fuses Control Barrier Functions (CBFs) with consensus theory. First, addressing the heterogeneity and limited sensing ranges of the AUVs, a cooperative coverage model based on a modified Voronoi partition is constructed. A nominal controller based on consensus theory is designed to balance the ratio of task workload to individual capability for each AUV. By minimizing a Lyapunov-like function via gradient descent, the swarm achieves self-organized optimal coverage. Second, to guarantee system safety, multiple safety constraints are designed for the AUV double-integrator dynamics, utilizing Zeroing Control Barrier Functions (ZCBFs) and High-Order Control Barrier Functions (HOCBFs). This approach unifies the handling of collision avoidance and velocity limitations. Finally, the nominal coverage controller and safety constraints are integrated into a Quadratic Programming (QP) formulation. This constitutes a safety-critical layer that modifies the control commands in a minimally invasive manner. Theoretical analysis demonstrates the stability of the framework, the forward invariance of the safe set, and the convergence of the coverage task. Simulation experiments verify the effectiveness and robustness of the proposed method in navigating obstacles and efficiently completing heterogeneous cooperative coverage tasks in complex environments. Full article

Accurate lithologic maps derived from satellite imagery underpin structural interpretation, mineral exploration, and geohazard assessment. However, automated mapping in complex terranes remains challenging because spectrally similar units, narrow anisotropic bodies, and ambiguous contacts can degrade boundary fidelity. In this study, we propose SegNeXt-HFCA, a hierarchical multiscale fusion network with coordinate attention for lithologic segmentation from a Sentinel-2/DEM feature stack. The model builds on SegNeXt and introduces a hierarchical multiscale encoder with coordinate attention to jointly capture fine textures and scene-level structure. It further adopts a class-frequency-aware hybrid loss that combines boundary-weighted online hard-example mining cross-entropy with Lovász-Softmax to better handle long-tailed classes and ambiguous contacts. In addition, we employ a robust training and inference scheme, including entropy-guided patch sampling, exponential moving average of parameters, test-time augmentation, and a DenseCRF-based post-refinement. Two study areas in the Beishan orogen, northwestern China (Huitongshan and Xingxingxia), are used to evaluate the method with a unified 10-channel Sentinel-2/DEM feature stack. Compared with U-NetFormer, PSPNet, DeepLabV3+, DANet, LGMSFNet, SegFormer, BiSeNetV2, and the SegNeXt backbone, SegNeXt-HFCA improves mean intersection-over-union (mIoU) by about 3.8% in Huitongshan and 2.6% in Xingxingxia, respectively, and increases mean pixel accuracy by approximately 3–4%. Qualitative analyses show that the proposed framework better preserves thin-unit continuity, clarifies lithologic contacts, and reduces salt-and-pepper noise, yielding geologically more plausible maps. These results demonstrate that hierarchical multiscale fusion with coordinate attention, together with class- and boundary-aware optimization, provides a practical route to robust lithologic mapping in structurally complex regions. Full article