Search Results (607)

Electroencephalography (EEG) and artificial intelligence (AI) are central tools in affective computing, yet fear-related inference remains constrained by a gap between experimental performance and deployable evidence: high accuracy on controlled datasets guarantees neither neurobiological specificity, nor robustness across users, nor traceable inference. EEG does not measure fear directly; it captures cortical dynamics compatible with threat processing, arousal or vigilance. This paper introduces the Fear-EEG-AIFramework, a technology-independent architecture for designing, validating and deploying EEG–AI systems targeting fear-compatible neurophysiological states. Building on a prior systematic review, it organizes development into three hierarchical levels—neurobiological foundation, neuroelectric correlates and computational modeling—governed by six cross-cutting principles, and makes them evaluable through eight observable dimensions, the Fear-EEG-AI Alignment Index (FAI) and five validation evidence levels. Applied to the eight eligible studies of that corpus by two raters who had not drafted the instrument, with adjudication by a third, the rubric reached 67.2% exact agreement (quadratic-weighted $\kappa =0.741$; ICC(2,1) $=0.744$). Under 10,000 admissible weightings the equal-weight ordering survived in 10.1% of draws, and one of 56 ordered pairs met a stability criterion fixed before the pairwise results were tabulated. The index is therefore reported as a descriptor of methodological documentation and not as a ranking of implementations. Full article

Biological surfaces have unique characteristics and properties due to the presence of well-organized micro- and nano-architectures. Fish scales, lotus leaves, shark skin, gecko feet, and insect wings are prime examples of how hierarchical and directional features can control the interfacial interactions which regulate the friction, wettability, fluid transport, adhesion, and fouling properties. However, translation of these surface textures onto engineered surfaces requires precision manufacturing methods which are capable of reproducing well-defined textures while also maintaining surface integrity. The present review examines vibration-assisted machining (VAM) as a precision manufacturing approach for producing functional and bio-inspired surface textures and discusses their potential use in tribological and electrochemical biosensing interfaces. VAM can generate deterministic dimples, grooves, crosshatch patterns, fish-scale structures, and hierarchical micro-textures through controlled tool–workpiece kinematics. Particular attention is given to the relationship between texture geometry and function, including lubricant retention, hydrodynamic pressure generation, contact-area reduction, debris entrapment, wettability, and fluid transport. These mechanisms are then coupled to electrochemical sensing where structured interfaces may increase the electrochemically active surface area, facilitate analyte transport and charge transfer, and provide more sites for bioreceptor immobilization. Hierarchical micro–nano-textures are especially attractive as they combine the mechanical functions of micro-scale structures with the interfacial and electrochemical advantages of nano-scale features. The present manuscript further discusses tribochemical stability, biofouling, physiological interfaces, and emerging biosensing applications. Finally, future opportunities are identified in quantitative bio-inspired texture design, AI-assisted inverse design, hierarchical functionalization, scalable manufacturing, and in vivo validation. This perspective positions VAM as a manufacturing bridge between biological surface design principles and multifunctional engineered interfaces for sensing and biomedical applications. Full article

Dynamic covalent biomacromaterials are redefining the design of adaptive soft materials by integrating the structural diversity and biological functionality of biomacromolecules with reversible covalent bond exchange. In contrast to permanently cross-linked networks, these materials can continuously reorganize their internal architecture in response to mechanical, chemical, and biological stimuli, enabling self-healing, stress relaxation, injectability, shape adaptation, programmable degradation, and spatiotemporally controlled molecular transport. This review critically examines the molecular and network-level principles that govern such adaptive behavior, emphasizing how bond-exchange chemistry, reaction kinetics, cross-link density, network topology, and biomacromolecular architecture collectively dictate macroscopic performance. Major dynamic covalent motifs, including imine and hydrazone bonds, boronic esters, disulfides, reversible Diels–Alder adducts, and related exchangeable linkages, are analyzed in the context of polysaccharides, proteins, peptides, nucleic acids, and other biologically derived macromolecules. Rather than treating dynamic chemistry solely as a cross-linking strategy, we highlight how molecular exchange propagates across hierarchical length scales to regulate mechanics, molecular diffusion, cell–material interactions, tissue integration, degradation, and therapeutic release. Recent developments in injectable hydrogels, regenerative matrices, wound-healing platforms, drug and gene delivery systems, bioadhesives, and responsive biomedical interfaces are critically assessed to identify both emerging opportunities and persistent limitations. The remaining challenges lie in reconciling exchange kinetics with mechanical robustness, physiological stability, cytocompatibility, reproducibility, and manufacturability, all of which remain central to successful biomedical translation. By establishing explicit relationships between molecular bond dynamics, network adaptation, and biological function, this review provides a unifying framework for engineering next-generation biomacromaterials capable of dynamically interacting with, responding to, and integrating within complex biological environments. Full article

Traditional reinforcement learning for autonomous driving faces challenges such as slow convergence, control instability, and limited adaptability. To address these issues, this paper proposes a hybrid decision-making framework that integrates V2X communication, human–machine interaction experience, and deep reinforcement learning. Based on the Actor–Critic architecture, the framework incorporates a human-like driving tendency network to constrain the policy search space and integrates a Kalman filter-based V2X communication anomaly handling module. The model constructs the vehicle’s action space and state representation. Comparative experiments with a method lacking the driving tendency module and the hierarchical planning method DDPG-RRT* show that DDPG-HP achieves faster convergence and better scenario adaptability than these baselines. It effectively avoids static and dynamic obstacles, demonstrates anticipatory and gradual deceleration, and initiates braking 1.7 s earlier than the DDPG-RRT* method, thereby reducing collision risks and enhancing pedestrian safety. Furthermore, the communication anomaly compensation mechanism effectively improves the system’s adaptability and overall robustness under unstable communication conditions. Full article

Auxiliary power supplies in voltage source converter-based high-voltage direct current (VSC-HVDC) converter stations are exposed to severe bus voltage sags, abrupt load changes, and wide-range operating conditions that seriously challenge conventional control strategies. This paper proposes an event-triggered multi-mode adaptive control architecture which combines linear active disturbance rejection control (LADRC) with model predictive control (MPC) in a hierarchical manner. The novelty of this work lies in embedding mode-adaptive prediction and disturbance feedforward into an event-triggered LADRC-MPC cascade, a combination that has not been reported for high-frequency auxiliary power converters in HVDC stations. The scheme features an adaptive event-triggering mechanism that cuts the MPC computational burden by over 80% in steady-state, a multi-mode LADRC outer loop handling continuous/discontinuous conduction mode (CCM/DCM) transitions smoothly, and a disturbance-feedforward MPC inner loop embedding the estimated disturbance into the cost function for proactive disturbance rejection. In addition, the input voltage is reconstructed from the inductor voltage balance using linear extended state observer (LESO) estimates, eliminating the dedicated input-voltage sensor. Simulations on a 50 kHz Boost converter show that the proposed method confines the voltage fluctuation during CCM-to-DCM transitions from ±28 V to within ±2.5 V; shortens the recovery time under input-voltage steps from 15 ms to 4.5 ms; and keeps the steady-state error within ±0.05% of the 600 V output, well inside the ±0.5% specification over the full load range. The event-triggered mechanism lowers the average CPU occupancy from 58% under periodic execution to below 15%, enabling high-frequency auxiliary power applications in HVDC systems. Full article

Class imbalance remains a persistent challenge in supervised classification, particularly when the minority class represents rare but important observations. Common approaches address this problem through undersampling, synthetic oversampling, cost-sensitive learning, or decomposition of complex class distributions. This study introduces Hierarchical Majority Decomposition (HMD), an architecture in which the majority class is recursively partitioned into a binary tree of increasingly localized subclasses while the complete minority class remains intact. A supervised local classifier is associated with each internal majority node and distinguishes the minority class from the node’s two child majority subclasses. During top-down inference, a minority prediction terminates traversal, whereas a majority-subclass prediction simultaneously identifies the current observation as not yet recognized as minority and determines the branch in which the next minority comparison is performed. HMD was evaluated using six binary benchmark datasets with imbalance ratios ranging from approximately 3.2:1 to 42.0:1. An unrestricted decision tree was used as a controlled base learner, and all methods were evaluated using 5-fold stratified cross-validation repeated 10 times. HMD was compared with an untreated decision tree, a class-weighted decision tree, random undersampling, SMOTE, and a root-only flat majority decomposition. Increasing HMD inference depth increased minority recall relative to the untreated decision tree on all six datasets. The strongest effects were observed on Mammography, Climate Model Simulation Crashes, Oil Spill, and Ozone Level Detection. Importantly, the root-only decomposition generally remained close to the untreated decision tree, whereas deeper HMD inference produced substantially larger sensitivity increases, indicating that recursive minority comparison rather than majority decomposition alone is responsible for much of the observed effect. Random undersampling frequently achieved equal or greater minority recall and balanced accuracy, but often at a substantially greater loss of specificity and minority precision. The results demonstrate that a recursively decomposed majority class can be used as an active inference structure in which an unchanged minority class is repeatedly reconsidered against progressively localized majority contexts. Full article

Gold leaf electrodes offer a low-cost, substrate-versatile strategy for electrochemical sensing and microfluidic systems. Herein, we demonstrate a unified Parafilm^®-mediated platform in which the Parafilm^® layer serves as both the adhesive for gold-leaf electrode patterning and the bonding/channel-defining layer for microfluidic assembly via heat-activated bonding. Gold leaf electrodes were patterned on diverse substrates by laser ablation, achieving a minimum achieved linewidth of 53 ± 3 μm on polyethylene terephthalate (PET) sheets. Electrochemical characterization showed stable CV responses after 40 manual bending cycles and good batch reproducibility. Leveraging Parafilm’s hydrophobicity and thermoplasticity, we integrated electrodes into hierarchical microfluidic architectures on paper and PET. Paper-based hybrid devices with asymmetric self-driven channels enabled dual-zone measurements. Structured PET microchannels yielded microfluidic devices with sequential flow control and selective loading. In a proof-of-concept H₂O₂ assay, on-chip Prussian blue electrodeposition followed by amperometry gave a detection limit of 0.44 mg/L and ~85% recovery in spiked pond water. This Parafilm^®-mediated strategy provides a low-cost route for cross-substrate microfluidic–electrochemical integration, unifying electrode fabrication and device assembly on a single thermoplastic bonding platform. Full article

Bilateral robotic rehabilitation has emerged as a technological approach for promoting coordinated upper-limb training after stroke. This state-of-the-art review critically analyzes bilateral upper-limb rehabilitation robots with emphasis on mechanical architecture, actuation and transmission, bilateral interaction modalities, control strategies, assistance modes, and validation evidence. A structured literature search covering 2010 to 8 July 2026 identified 141 records; 23 technology-related publications were retained for the state-of-the-art analysis, comprising 18 primary bilateral robotic studies and 5 supporting technical/contextual publications. The reviewed systems were organized according to a hierarchical framework distinguishing end-effector, exoskeleton, and hybrid architectures from simultaneous bilateral, master–slave/mirror-based, and cooperative bimanual interaction modalities. The evidence indicates that end-effector systems favor mechanical simplicity and adaptable workspaces, whereas exoskeletons provide more direct joint-level control at the cost of greater alignment and mechanical complexity. Control approaches increasingly incorporate impedance, admittance, assist-as-needed, and bio-signal-based strategies to improve compliant interaction and adapt assistance to user contribution. However, many advanced systems remain supported primarily by engineering validation or experiments involving healthy participants, while direct post-stroke clinical validation is comparatively limited. Future development should therefore prioritize clinically validated adaptive assistance, control strategies capable of accommodating asymmetric bilateral contribution, and safe, usable, and affordable systems suitable for clinical and home-based rehabilitation. Full article

Vertical farming requires precise monitoring and control of environmental conditions to support consistent year-round crop production and efficient resource utilization. However, multilayer cultivation creates spatial, vertical, and temporal variations in temperature, humidity, CO₂, airflow, and lighting, complicating the maintenance of uniform growing conditions. This review analyzed 70 recent studies on remote monitoring and control in smart vertical farms, focusing on environmental variability, sensing infrastructure, sensor-node placement, communication architectures, actuator operation, and control strategies. Studies were classified into five thematic areas: sensing and monitoring, automation and control, environmental and climate control, IoT and wireless communication, and CFD-based analysis. Recent research advances include the development of spatially distributed sensing networks for monitoring and the adoption of adaptive, data-driven, model-based, and hierarchical methods for control. Nevertheless, significant research gaps persist, such as the lack of standardized sensor-placement strategies, limited integration of actuator-level operational data, insufficient long-term commercial-scale validation, and limited transferability of monitoring and control frameworks across facility configurations. Many methods remain adapted from greenhouse systems rather than specifically designed for vertical farms. Addressing these gaps is essential for improving the reliability, energy efficiency, and scalability of smart vertical farm systems. Full article

Three-dimensional coverage deployment is a fundamental challenge for underwater mobile sensor networks (UMSNs), where the coupled requirements of high coverage, guaranteed connectivity, and low energy consumption must be simultaneously satisfied under dynamic and partially observable conditions. To address these issues, this paper proposes a coverage-prioritized constrained multi-agent proximal policy optimization (CP-CMAPPO) algorithm that decomposes the control process into three hierarchical layers: a coverage-frontier target allocation layer for explicit exploration guidance, a residual MAPPO control layer for local motion refinement, and a connectivity-aware action synthesis layer for topology integrity maintenance. The simulations demonstrate that CP-CMAPPO achieves nearly complete cumulative coverage while guaranteeing full network connectivity. Notably, the proposed method maintains stable and predictable energy expenditure despite the additional cost of ensuring coverage and connectivity. These results validate that the proposed hierarchical architecture effectively mitigates the exploration burden, prevents topology fragmentation, and enables robust, energy-efficient cooperative coverage in underwater acoustic mobile sensor networks. Full article

This paper presents a hierarchical estimator-assisted control architecture for a three-phase shunt active power filter (SAPF) for industrial power-quality improvement. The framework combines grid synchronization, harmonic-reference generation, cascaded DC-link and dq-current control, adaptive state–parameter estimation, modulation, and supervisory protection. Numerical and experimental validations are reported separately. In the final 15 kHz switching-model validation of a 400 V, 50 Hz system supplying a six-pulse rectifier load, grid-current THD is reduced from 29.50% to 0.7469%, corresponding to a 97.47% reduction, while the power factor reaches 0.99993. The DC-link voltage is regulated to 750.32 V with 0.0896% ripple. Current- and DC-link-voltage estimation RMSEs are 0.0590 A and 7.76 mV, respectively, and all defined Safe Operating Area criteria are satisfied. Joint online identification yields R_f = 0.03 ohm and L_f = 2.998 mH, with errors of +0.0164% and −0.0667%. On the 15 kHz laboratory prototype, grid-current THD decreases from 29.1% to 2.8%. The proposed architecture provides improved robustness against parameter variations and load disturbances, enhanced observability through online state reconstruction, superior energy regulation, and comprehensive diagnostic capabilities while preserving a computational structure suitable for real-time DSP and FPGA implementation in advanced industrial power-quality conditioning systems. A broader robustness and thermal-estimator validation remain subjects for future work. Full article

The growing demand for high-quality motor fuels, combined with increasingly stringent environmental regulations, has stimulated research aimed at developing advanced catalytic methods for upgrading petroleum feedstocks. Conventional technologies for the deep purification and upgrading of petroleum fractions are based on hydrotreating processes that require an external H₂ supply, elevated pressure, and dedicated infrastructure, resulting in high capital and energy costs. In this context, hydrogen-free catalytic upgrading is considered a promising alternative approach. This review analyzes current approaches to the hydrogen-free catalytic upgrading of hydrocarbon feedstocks, with particular emphasis on zeolite-based catalysts. The main reaction pathways, including dehydrogenation, isomerization, dehydrocyclization, and intermolecular hydrogen transfer, are discussed in relation to the structural characteristics of the catalysts and their acidic and redox properties. Particular attention is given to microporous ZSM-5, Y, and BEA zeolites, as well as MCM-41-type mesoporous materials, hierarchical micro–mesoporous composites, and nanocatalysts. The effects of the distribution of Brønsted and Lewis acid sites, pore architecture, and modification with metals (Ni, Co, Zn, Mo, rare-earth elements, etc.) on the catalytic activity, selectivity, and operational performance of catalytic systems are examined in detail. The analysis demonstrates that the efficiency of hydrogen-free upgrading processes is governed by the synergistic interaction between the acidic and metallic functions of the catalyst surface, while also depending on the accessibility of active sites and mass transfer. The development of hierarchical porosity can reduce diffusion limitations; however, excessive formation of secondary porosity may lead to a partial loss of the zeolite crystalline structure and acid sites. Similarly, metal incorporation enhances catalyst activity along key reaction pathways, but its effects on selectivity, coke formation, and stability are determined by the nature, loading, and location of the metal sites. Therefore, the prospects for hydrogen-free upgrading are primarily associated with optimizing these interrelated characteristics. Analysis of pilot- and industrial-scale experience with early hydrogen-free processes indicates that current restrictions on sulfur, benzene, and total aromatic contents determine the target selectivity requirements for such catalytic systems. Overall, hydrogen-free catalytic upgrading represents a promising complement to conventional hydroprocessing technologies, offering potential advantages in terms of improved energy efficiency, reduced dependence on hydrogen, and lower environmental impact. Further advances in this field are expected to rely on the coordinated control of acidic and metallic functions, pore structure, and mass transfer, together with the integration of physics-based modeling, systematic experimental screening, and data-driven analytical methods. Full article

The integration of LiDAR sensors into quadcopter control systems is fundamental for autonomous navigation in cluttered environments, yet the precise performance trade-offs between different tracking architectures under perceptual uncertainty remain insufficiently quantified. This paper presents a comprehensive 3D computational study evaluating the impact of LiDAR-informed reactive control on quadcopter flight dynamics. A non-linear six-degree-of-freedom quadcopter model coupled with a ray-casting LiDAR simulator and a hierarchical proportional–integral–derivative (PID) controller was developed. To ensure a rigorous and fair comparison, two distinct navigation architectures were evaluated under identical conditions of stochastic perceptual noise and kinematic limits: a standard discrete waypoint-tracking algorithm (Scenario A) and an advanced stochastic hybrid system utilizing 3D Spline interpolation, Pure Pursuit tracking, and a noise-filtered LiDAR reactive field (Scenario B). While both approaches achieved 100% collision-free mission success across a multi-obstacle environment, the comparative analysis revealed spatial and temporal trade-offs. The discrete navigation of Scenario A proved highly efficient, completing the trajectory in 174.95 s over 230.25 m with a minimum safety clearance of 1.02 m. Conversely, the continuous hybrid architecture of Scenario B successfully avoided local minimum traps but induced massive orbital evasion maneuvers, which increased flight time to 287.10 s and traveled distance to 389.01 m. Furthermore, the Pure Pursuit controller’s corner-cutting behavior under stochastic conditions reduced the minimum safety distance to 0.78 m, while it penalized the average horizontal speed due to intermittent reactive braking. These findings provide strict quantitative benchmarks, demonstrating that while hybrid spline-based tracking offers robust deadlock evasion in noisy environments, it incurs substantial penalties in overall flight efficiency and safety margins compared to discrete routing. Full article

In our earlier methodology paper, we introduced a hierarchical framework combining graph compression, Diffusion Convolutional Recurrent Neural Networks (DCRNNs), and Multi-Agent Reinforcement Learning (MARL) to approximate Bellman’s optimality principle for real-time energy system control, validated using Palm Springs, California weather data. This extension paper addresses three critical advances. First, we provide a rigorous theoretical proof demonstrating that the DCRNN’s Markovian reduction of strongly path-dependent dynamics yields a bounded approximation error, with the error decaying exponentially in the mixing time of the underlying graph diffusion process. Second, we extend the framework to diverse weather regimes—from stable Mediterranean climates (San Diego) to highly variable marine west coast (Seattle), monsoon (Mumbai), and typhoon-prone regions (Hong Kong)—quantifying how weather-induced path dependence affects the required hidden state dimension and forecasting accuracy. We present comprehensive Leave-One-Out Cross-Validation (LOOCV) results across six cities, demonstrating consistent generalization with R² drops of less than 0.1% under out-of-sample testing. A controlled baseline comparison under matched training protocols shows that the graph-free GRU achieves comparable or higher R² on the one-step prediction task, which we attribute to the near-cumulative structure of the target and the small evaluation graph. We frame the DCRNN’s contribution around its theoretical guarantee and its potential advantage on larger graphs and longer horizons. We also characterize conditions under which the model expects to fail, specifically when weather stochasticity violates the geometric mixing assumption or when the effective temporal correlation length exceeds the GRU’s memory capacity. Third, we outline physics-informed enhancements that are proposed as future development: CFD-integrated loss functions, differentiable Model Predictive Control (MPC) heads, and a modular design enabling alternative turbine configurations. We also propose a standardized rooftop solar thermal deployment architecture with 200 m × 100 m, 100 m × 100 m, and 100 m × 50 m modules designed for data center footprints with pre-allocated HVAC space. We conclude with a stage-gated validation roadmap progressing from unit tests to hardware-in-the-loop simulation to full-scale FEED-site deployment. The completed contributions of this paper are the theorem, its empirical assumption verification, the multi-climate LOOCV study, the matched-protocol baseline comparison, and the sensor-failure robustness analysis. The remaining components are described as proposed extensions. Full article

Agentic artificial intelligence requires safeguards that remain effective across trajectories rather than only at individual decisions. This study introduces MARIS-TRA, a hierarchical temporal runtime assurance extension of Controlled Agentic AI Systems. A compact, auditable fragment of Signal Temporal Logic (STL) provides quantitative robustness for speed, arena, pairwise separation, restricted-zone, and bounded-recovery requirements. The final architecture uses invariant hard-safety formulas to define feasibility, while recovery/liveness is monitored separately and may trigger escalation. Across a 5750-episode main campaign, MARIS-TRA achieved 100.00% realized hard-safety window satisfaction, 100.00% hard-safety episode satisfaction (450/450; Wilson 95% CI 99.15–100.00%), zero collision episodes, and 2.55 ms mean latency in the core comparison. Under the primary strict numerical semantics, an independent CBF-QP baseline achieved 88.89% hard-safety episode satisfaction; all 50 strict failures were very small arena-boundary overshoots in the boundary-stress scenario, with no collision or separation failures, and the post hoc tolerance sensitivity reached 100% at epsilon = 10⁻⁴ normalized simulator units. An additional 8640-episode targeted validation examined recovery hysteresis, model mismatch, AHO control flow, and safety–recovery conflicts. Under confirmatory high-density testing, safety-only shielding preserved hard safety in 810/810 episodes, whereas joint-hard enforcement produced 28/810 separation-safety failures (3.46%; Wilson 95% CI 2.40–4.95%) without collisions. Actuation-noise and delay experiments further show that the formal result is a conditional predicted-trace certification rather than a disturbance-robust guarantee on future receding-horizon execution. The empirical claims are, therefore, limited to the evaluated continuous-action multi-agent setting, while the architecture remains policy-separable and auditable. Full article