Search Results (2,567)

This study proposes a Cooperative Traffic Controller System (CTS), an urban eco-network control system that leverages Multi-Agent Reinforcement Learning (MARL), to address urban road congestion and environmental pollution. The proposed system synergizes traffic signal timing optimization and speed guidance control, simultaneously enhancing network efficiency, reducing carbon emissions, and minimizing energy consumption. A Beta-enhanced Soft Actor-Critic (SAC) algorithm is applied to achieve the joint optimization of the traffic signal phasing and vehicle speed coordination. Experimental results show that in large-scale networks, the improved SAC reduces the average delay time per vehicle by approximately one minute, reduces CO₂ emissions by more than double, and reduces fuel consumption by 56%. Comparative analyses of the algorithm versus the PPO and standard SAC demonstrate its superior performance in complex traffic scenarios—specifically in congestion mitigation and emissions reduction. Full article

Early wildfire detection in remote areas remains a critical challenge due to limited connectivity, intermittent solar energy, and the need for autonomous, long-term operation. Existing systems often rely on fixed sensing schedules or cloud connectivity, making them impractical for energy-constrained deployments. We introduce EcoWild, a reinforcement learning-driven cyber-physical system for energy-adaptive wildfire detection on solar-powered edge devices. EcoWild combines a decision tree-based fire risk estimator, lightweight on-device smoke detection, and a reinforcement learning agent that dynamically adjusts sensing and communication strategies based on battery levels, solar input, and estimated fire risk. The system models realistic solar harvesting, battery dynamics, and communication costs to ensure sustainable operation on embedded platforms. We evaluate EcoWild using real-world solar, weather, and fire image datasets in a high-fidelity simulation environment. Results show that EcoWild consistently maintains responsiveness while avoiding battery depletion under diverse conditions. Compared to static baselines, it achieves 2.4× to 7.7× faster detection, maintains moderate energy consumption, and avoids system failure due to battery depletion across 125 deployment scenarios. Full article

Waste materials have emerged as attractive low-cost feedstocks for adsorbent development in environmental remediation and materials engineering. Organic wastes are particularly rich in cellulose, hemicellulose, lignin, and pectin, which provide reactive oxygenated groups such as hydroxyls and carboxyls. While inorganic wastes offer stability, lower water retention makes them promising candidates. This study explores the functionalization of waste-derived organic and inorganic matrices using two amine-based agents: 3-aminopropyltrimethoxysilane (APTMS) and polyethylenimine (PEI). The materials were categorized as organic (orange peel, corn cob) or inorganic (silica gel, eggshell) and subjected to a pretreatment process involving drying, grinding, and sieving; inorganic substrates additionally underwent acid activation with citric acid. Surface modification was carried out in ethanolic (APTMS) or aqueous (PEI) media. To assess their suitability and processability as particulate sorbents, drying kinetics, physicochemical properties (FTIR, ζ-potential, pH, conductivity, Boehm titration), and flow characteristics (Carr and Hausner indices) were evaluated. The findings enable a comparative analysis of the functionalization efficiency and elucidate the relationship between substrate type (organic vs. inorganic) and its performance as a modified adsorbent. This approach advances the development of novel sorbent matrices for greenhouse gas mitigation while reinforcing circular economy principles through the valorization of low-cost, readily available waste materials. Full article

This study presents, for the first time, a systematic investigation of the use of micronized post-consumer silica gel as a reinforcing filler in natural rubber composites, in direct comparison with commercial silicas. Thirteen formulations were prepared using three types of silica (recycled, Copasil, and ZC-185P) and three coupling agents (TESPT, VTMO, and Chartwell C-515.71HR^®). The recycled silica exhibited high purity (97.33% Si) and irregular morphology but resulted in lower crosslink density (0.47–0.59 × 10⁻⁴ mol·cm⁻³) and inferior mechanical performance, with tensile strength up to 7.9 MPa and high abrasion loss (878–888 mm³). In contrast, ZC-185P silica combined with TESPT achieved the best results, with a tensile strength of 18.5 MPa, tear resistance of 99.36 N·mm⁻¹, and minimum abrasion loss of 170 mm³. Although less efficient in reinforcement, composites containing recycled silica were successfully applied in the production of a functional rubber mat, demonstrating their practical viability. The results confirm the potential for valorization of spent silica gel as an alternative raw material for sustainable composites, contributing to the circular economy. Full article

Recently, cognitive training programs using digital content with visuoperceptual stimulation have been developed and commercialized. In particular, digital intervention content for children with attention deficit hyperactivity disorder (ADHD) has been developed as games, enhancing motivation and accessibility for the target population. Active stimulation is required to elicit positive effects on self-regulation training, including attention control and impulse inhibition, through task-based content. Common forms of stimulation include emotional stimuli, such as praise and encouragement, and economic stimuli based on a self-directed token economy system. Economic stimulation can serve as active reinforcement because the child directly engages as the primary agent within the task content. This study applied and validated a token economy intervention using digital therapeutic content in children with ADHD. Behavioral assessments were conducted using the Comprehensive Attention Test (CAT) and the Korean version of the Child Behavior Checklist (K-CBCL). The developed digital intervention content implemented a user-centered token economy based on points within the program. In the CAT Flanker Task, the experimental group (0.84 ± 0.40) showed significantly higher sensitivity factor scores than the control group (0.72 ± 0.59) after 4 weeks, with a large effect size (F = 4.76, p = 0.038, partial η² = 0.150). Additionally, the rate of change in externalizing behavior scores on the K-CBCL showed a significant difference between the two groups (t = 2.35, p = 0.026, Cohen’s d = 0.860), demonstrating greater improvement in externalizing symptoms in the experimental group than in the control group. Therefore, this study suggests that the participant-centered implementation model using token economy mechanisms in digital intervention content may serve as a novel and effective therapeutic approach for children with ADHD. Full article

Border surveillance, as a critical component of national security, places increasingly stringent demands on the target perception capabilities of video monitoring systems, especially in wide-area and complex environments. To address the limitations of existing systems in low-confidence target detection and multi-camera collaboration, this paper proposes a novel visual enhancement method for cooperative control of multiple PTZ (Pan–Tilt–Zoom) cameras based on hierarchical reinforcement learning. The proposed approach establishes a hierarchical framework composed of a Global Planner Agent (GPA) and multiple Local Executor Agents (LEAs). The GPA is responsible for global target assignment, while the LEAs perform fine-grained visual enhancement operations based on the assigned targets. To effectively model the spatial relationships among multiple targets and the perceptual topology of the cameras, a graph-based joint state space is constructed. Furthermore, a graph neural network is employed to extract high-level features, enabling efficient information sharing and collaborative decision-making among cameras. Experimental results in simulation environments demonstrate the superiority of the proposed method in terms of target coverage and visual enhancement performance. Hardware experiments further validate the feasibility and robustness of the approach in real-world scenarios. This study provides an effective solution for multi-camera cooperative surveillance in complex environments. Full article

This work presents a comprehensive study on the fabrication, microstructural evolution, and mechanical performance of hybrid aluminum matrix composites based on Al6060 alloy reinforced with ~2 wt.% TiB₂ and ~1 wt.% multi-walled carbon nanotubes (MWCNTs). The composites were produced via ultrasonically assisted stir casting followed by radial-shear rolling (RSR). The combined processing route enabled a uniform distribution of reinforcing phases and significant grain refinement in the aluminum matrix. SEM, EDS, XRD, and EBSD analyses revealed that TiB₂ particles acted as nucleation centers and grain boundary pinning agents, while MWCNTs provided a network structure that suppressed agglomeration of ceramic particles and enhanced interfacial load transfer. As a result, hybrid composites demonstrated a submicron-grained structure with reduced anisotropy. Mechanical testing confirmed that yield strength (YS) and ultimate tensile strength (UTS) increased by 67% and 38%, respectively, in the cast state compared to unreinforced Al6060, while after RSR processing, YS exceeded 115 MPa and UTS reached 164 MPa, with elongation preserved at 14%. Microhardness increased from 50.2 HV0.2 (base alloy) to 82.2 HV0.2 (hybrid composite after RSR). The combination of ultrasonic melt treatment and RSR thus provided a synergistic effect, enabling simultaneous strengthening and ductility retention. These findings highlight the potential of hybrid Al6060/TiB₂–MWCNT composites for structural applications requiring a balance of strength, ductility, and wear resistance. Full article

In response to the escalating plastic pollution crisis, the development of high-performance biodegradable materials is critical. Poly(butylene succinate) (PBS) is an important biodegradable polymer as it possesses excellent biodegradability and processability. But it suffers from limitations such as low mechanical strength, poor thermal stability, and high production costs. In this study, taxus residue (TF), a waste by-product, was utilized as a reinforcing filler to reduce PBS costs while enhancing its overall performance. To address the interfacial incompatibility between TF and PBS, branched PBS (T-PBS) was introduced as a compatibilizer. The TF was surface-modified via alkali treatment and silane coupling (KH550), and a series of PBS/TF/T-PBS composites with varying T-PBS viscosity grades were prepared by melt blending. The compatibilization mechanism of T-PBS and its influence on the composite structure, crystallization behavior, thermal stability, rheological, and mechanical properties were systematically investigated. Results show that the branched structure significantly enhanced T-PBS melt strength and reactivity. The introduction of T-PBS effectively improved interfacial compatibility between TF and PBS matrix, reducing phase separation and interfacial defects. Compared to uncompatibilized PBS/TF composites, those with appropriately viscous T-PBS exhibited improved tensile strength (increased by 19.7%) and elongation at break (increased by 78.8%), while flexural strength was also maintained at an enhanced level. The branched points acted as nucleating agents, increasing the onset temperature and degree of crystallinity. In the high-temperature region, the synergistic barrier effect from TF and char residue improved thermal stability (T_85% reached 408.19 °C). Rheological analysis revealed enhanced viscosity and elasticity of the system. This study provides a promising strategy and theoretical foundation for the high-value utilization of taxus waste and the development of high-performance biodegradable PBS-based composites. Full article

In many real-world control tasks, agents operate under partial observability, where access to complete state information is limited or corrupted by noise. This poses significant challenges for reinforcement learning algorithms, as methods relying on full states or long observation histories can be computationally expensive and less robust. Four-Phase Rapid Motor Adaptation (FRMA) is a reinforcement learning framework designed to address these challenges in high-frequency control tasks under partial observability. FRMA proceeds through four sequential stages: (i) full-state pretraining to establish a strong initial policy, (ii) auxiliary hidden-state prediction for LSTM memory initialization, (iii) aligned latent representation learning to bridge partial observations with full-state dynamics, and (iv) latent-state policy fine-tuning for robust deployment. Notably, FRMA leverages full-state information ($s_t$) only during training to supervise latent representation learning, while at deployment it requires only short sequences of recent observations and actions. This allows agents to infer compact and informative latent states, achieving performance comparable to policies with full-state access. Extensive experiments on continuous control benchmarks show that FRMA attains near-optimal performance even with minimal observation–action histories, reducing reliance on long-term memory and computational resources. Moreover, FRMA demonstrates strong robustness to observation noise, maintaining high control accuracy under substantial sensory corruption. These results indicate that FRMA provides an effective and generalizable solution for partially observable control tasks, enabling efficient and reliable agent operation when full state information is unavailable or noisy. Full article

To address the limitations of the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm in autonomous control tasks including low convergence efficiency, poor training stability, inadequate adaptability of confrontation strategies, and challenges in handling sparse reward tasks—this paper proposes an enhanced algorithm by integrating the Rainbow module. The proposed algorithm improves long-term reward optimization through prioritized experience replay (PER) and multi-step TD updating mechanisms. Additionally, a dynamic reward allocation strategy is introduced to enhance the collaborative and adaptive decision-making capabilities of agents in complex adversarial scenarios. Furthermore, behavioral cloning is employed to accelerate convergence during the initial training phase. Extensive experiments are conducted on the MaCA simulation platform for 5 vs. 5 to 10 vs. 10 UAV island capture missions. The results demonstrate that the Rainbow-MADDPG outperforms the original MADDPG in several key metrics: (1) The average reward value improves across all confrontation scales, with notable enhancements in 6 vs. 6 and 7 vs. 7 tasks, achieving reward values of 14, representing 6.05-fold and 2.5-fold improvements over the baseline, respectively. (2) The convergence speed increases by 40%. (3) The combat effectiveness preservation rate doubles that of the baseline. Moreover, the algorithm achieves the highest average reward value in quasi-rectangular island scenarios, demonstrating its strong adaptability to large-scale dynamic game environments. This study provides an innovative technical solution to address the challenges of strategy stability and efficiency imbalance in multi-agent autonomous control tasks, with significant application potential in UAV defense, cluster cooperative tasks, and related fields. Full article

Locusts represent a persistent global agricultural pest, responsible for significant crop losses and socio-economic repercussions. The initiation of chemical control measures dates back to the late 19th century, with the use of poisoned baits, before advancing in the mid-20th century with the introduction of organochlorines, such as dieldrin. Despite their efficacy, the associated environmental, ecological, and human health risks led to the prohibition of dieldrin by the United States and the FAO by 1988. The demand for insecticides with reduced persistence and toxicity prompted the establishment of international organizations to coordinate locust research and management. In recent decades, chemical control has transitioned towards compounds with diminished persistence and selective agents. Concurrently, research has progressed in the development of bioinsecticides, notably Metarhizium acridum, and has reinforced preventive strategies. Emerging technologies, including remote sensing and machine learning, have facilitated early monitoring and predictive modeling, thereby enhancing outbreak forecasting. These tools support proactive, targeted interventions and are consistent with Integrated Pest Management principles, promoting more sustainable and ecologically responsible locust control strategies. Full article

Reinforcement learning agents are highly susceptible to adversarial attacks that can severely compromise their performance. Although adversarial training is a common countermeasure, most existing research focuses on defending against single-type attacks targeting either observations or actions. This narrow focus overlooks the complexity of real-world mixed attacks, where an agent’s perceptions and resulting actions are perturbed simultaneously. To systematically study these threats, we introduce the Action and State-Adversarial Markov Decision Process (ASA-MDP), which models the interaction as a zero-sum game between the agent and an adversary attacking both states and actions. Using this framework, we show that agents trained conventionally or against single-type attacks remain highly vulnerable to mixed perturbations. Moreover, we identify a key challenge in this setting: a naive mixed-type adversary often fails to effectively balance its perturbations across modalities during training, limiting the agent’s robustness. To address this, we propose the Action and State-Adversarial Proximal Policy Optimization (ASA-PPO) algorithm, which enables the adversary to learn a balanced strategy, distributing its attack budget across both state and action spaces. This, in turn, enhances the robustness of the trained agent against a wide range of adversarial scenarios. Comprehensive experiments across diverse environments demonstrate that policies trained with ASA-PPO substantially outperform baselines—including standard PPO and single-type adversarial methods—under action-only, observation-only, and, most notably, mixed-attack conditions. Full article

Brazilian red propolis (BRP) has emerged as a promising source of multifunctional phytochemicals with potent anti-inflammatory activity. This review provides a comprehensive analysis of the anti-inflammatory effects of BRP’s bioactive compounds, their molecular targets, and their mechanisms of action. Isolated compounds from BRP (such as formononetin, biochanin A, daidzein, calycosin, medicarpin, vestitol, and neovestitol) have demonstrated the ability to modulate critical pro-inflammatory signaling pathways, including NF-κB, TLR4, JAK/STAT, and PI3K/AKT, while concurrently activating antioxidant and cytoprotective responses via the Nrf2/HO-1 axis. These effects are further supported by the suppression of pro-inflammatory cytokines, regulation of immune cell infiltration and activation, inhibition of inflammasome components such as NLRP3, induction of autophagy, and polarization of macrophages and microglia from a pro-inflammatory (M1) to an anti-inflammatory (M2) phenotype. Collectively, these findings reinforce the potential of BRP as a rich source of multifunctional phytochemicals with broad therapeutic relevance for chronic inflammation and related pathologies. Future research should address the identified knowledge gaps by employing rigorous in vitro and in vivo toxicity assessments, exploring structure–activity relationships, and leveraging advanced delivery systems to optimize bioavailability. Such methodological approaches will be essential for translating the promising biological activities of BRP compounds into clinically viable therapeutic agents. Full article

Cooperative Multi-Agent Reinforcement Learning (MARL) relies on effective coordination among agents to maximize team performance in complex environments. However, existing coordination graph-based approaches often overlook dynamic group structures and struggle to accurately capture fine-grained inter-agent dependencies. In this paper, we introduce a novel method called the Group Attention Aware Coordination Graph (G2ACG), which builds upon the group modeling capabilities of the Group-Aware Coordination Graph (GACG). G2ACG incorporates a dynamic attention mechanism to dynamically compute edge weights in the coordination graph, enabling a more flexible and fine-grained representation of agent interactions. These learned edge weights guide a Graph Attention Network (GAT) to perform message passing and representation learning, and the resulting features are integrated into a global policy via QMIX for cooperative decision-making. Experimental results on the StarCraft II Multi-Agent Challenge (SMAC) benchmark show that G2ACG consistently outperforms strong baselines, including QMIX, DICG, and GACG, across various scenarios with diverse agent types and population sizes. Ablation studies further confirm the effectiveness of the proposed attention mechanism, demonstrating that both the number of attention heads and the number of GAT layers significantly affect performance, with a two-layer GAT and multi-head attention configuration yielding the best results. Full article

Background: The pathophysiology of Alzheimer’s disease (AD) is strongly linked to damage to the cholinergic systems of the central nervous system (CNS), mainly due to the formation of β-amyloid peptide plaques, which trigger intense inflammatory responses and are currently the main cause of the symptoms of the disease. Among the therapeutic strategies under investigation, classes of natural products with immunomodulatory properties, action on the CNS, and potent antioxidant activity, which contribute to neuroprotection, stand out. Methods: We aimed to evaluate the flavonoid quercetin using in silico, in vitro, and in vivo methods for the treatment of AD. Initially, the compounds were selected, and molecular dynamics simulations were performed. The in vitro assays included tests of antioxidant activity (DPPH), enzymatic inhibition of acetylcholinesterase (AChE), and prediction of oral toxicity. The in vivo studies investigated the effects on scopolamine-induced learning deficits and conducted histopathological analysis of the brain. Results: Quercetin showed structural stability in the complex with (AChE), with no significant alterations in the Root Mean Square Deviation (RMSD), SASA and radius of gyration (Rg) parameters. Through the same method it was possible to predict stability between the quercetin and inducible nitric oxide synthase (iNOS) complex, a possible mechanism for quercetin immunomodulation in the CNS. In the AChE inhibition test, the IC₅₀ obtained for quercetin was 59.15 μg mL⁻¹, while in the antioxidant test with DPPH, the concentration of 33.1 µM exhibited 50% of the scavenging of reactive oxygen species. This corroborates the perspective of quercetin having neuroprotective activity. This activity was also corroborated in vivo, in a zebrafish model, in which quercetin reduced the cognitive deficit induced by scopolamine. Histopathological analysis revealed its ability to prevent atrophy, caused by scopolamine, in the nervous tissue of animals, reinforcing the potential of quercetin as a neuroprotective agent. Conclusions: The results of the tests carried out with quercetin suggest that this molecule has antioxidant, AChE inhibitory, and neuroprotective activities, making it a good candidate for use in future clinical trials to ensure its efficacy and safety. Full article