Aflatoxin contamination in stored grains due to its severe health effects, economic losses, and persistence under post-harvest conditions remains a critical global food safety challenge. Aflatoxin B1 (AFB1), classified as a Group 1 carcinogen, poses significant mutagenic, hepatotoxic, and immunosuppressive risks. Conventional detection techniques including HPLC, LC-MS, and ELISA offer high sensitivity but are limited by complex instrumentation, lack of field applicability and high cost. Recent advancements in nanotechnology for rapid detection and effective mitigation of aflatoxins provide transformative solutions. The present review comprehensively discusses nanoparticle-based biosensors including metallic nanoparticles, carbon and graphene quantum dots, up-conversion nanoparticles, and surface-enhanced Raman spectroscopy (SERS) techniques for ultra-sensitive and onsite detection of AFB1. Detection limits ranging from pg/mL to ng/mL levels demonstrate the excellent analytical performance of nanotechnology-driven systems. Furthermore, nanocomposites, nano-encapsulated antifungal agents, green-synthesized nanoparticles, and nanoparticle-mediated RNA interference approaches for their role in suppressing fungal growth and inhibiting aflatoxin biosynthesis are emphasized. Mechanistic insights reveal that nanoparticles induce reactive oxygen species generation, gene downregulation in aflatoxin biosynthetic pathways, and structural toxin degradation. Additionally, nanocomposite-based grain storage materials improve barrier properties, reducing moisture and fungal proliferation. Despite promising advancements, concerns regarding nanoparticle toxicity, environmental accumulation, regulatory compliance, and large-scale implementation remain critical challenges. Overall, nanotechnology offers a multifunctional, sensitive, and sustainable strategy for strengthening aflatoxin detection, detoxification, and post-harvest management systems, thereby enhancing global grain safety and food security.
The growing global demand for sustainable energy and the environmental limitations of fossil-fuel-based power generation have intensified interest in next-generation photovoltaic technologies. Among emerging approaches, oxide-based solar cells—particularly dye-sensitized solar cells (DSSCs) and perovskite–oxide hybrid systems—have gained considerable attention due to their low fabrication cost, material versatility, and compatibility with flexible or building-integrated photovoltaic applications. Recent studies demonstrate that nanocarbon materials and room temperature ionic liquids (RTILs) can significantly enhance the performance of these devices by improving charge transport, catalytic activity, and interfacial stability. For example, graphene- or ionic liquid-modified electrolytes have been reported to increase ionic conductivity by up to 93
This article addresses a collocation method based on shifted Chebyshev cardinal functions defined on the interval [0, T], T>0 , to solve stochastic delay differential equations involving a constant delay. In this method, the stochastic delay differential equation is transformed into the stochastic Itô - Volterra integral equation with constant delay, then shifted Chebyshev cardinal functions are used as basis functions to approximate the obtained stochastic Itô - Volterra integral equation with constant delay, and the obtained equation is collocated at the suitable collocation points. Then, the M+1 Gauss-Legendre quadrature rule and the Itô approximation are used to approximate integral parts. Newton’s method is used to solve a generated system of algebraic equations to get the desired approximate solution. Moreover, the convergence analysis of the presented method is also established in detail. Additionally, the applicability of the proposed method is demonstrated by solving some numerical examples.
Wi-Fi Channel State Information (CSI) has become a widely studied modality for device-free sensing as it captures fine-grained wireless channel variations that can be mapped to human motion and presence while avoiding the explicit visual disclosure typical of vision-based systems. CSI-based pipelines have been explored for human activity and gesture recognition, fall detection, gait analysis, pose-related inference, and indoor localization. Despite strong results in controlled settings, practical deployment remains difficult due to measurement noise, sensitivity to environmental dynamics, multi-user interference, and system-level constraints in data acquisition and real-time processing. This article surveys machine learning methods forWi-Fi CSI sensing and analyzes more than 65 representative models, connecting algorithmic design choices with implementable end-to-end system design. We introduce a hierarchical taxonomy that organizes the literature into classical machine learning approaches, deep learning architectures, and hybrid strategies. Beyond modeling, we describe the full sensing pipeline- from hardware and network interface card (NIC) selection to software tools, antenna configuration, and signal conditioning- highlighting the design trade-offs that affect robustness and reproducibility. We further compare methods across major application domains and summarize open challenges in generalization to dynamic environments, multi-user separation, and resource-efficient inference. Finally, we outline research directions toward robust generalization, scalable deployment, and privacy-aware learning to support broader real-world adoption.
Landslides represent one of the most pervasive and hazardous geohazards globally, particularly in tectonically active and monsoon-dominated regions such as the Himalayas. The inner lesser and higher Himalayas are extremely susceptible to slope instability due to a fragile geological setup and erratic monsoon precipitation. The study focuses on the eastern Ramganga River basin, located in the Uttarakhand Himalaya, which experiences frequent occurrences of landslides. Therefore, identification and monitoring of hazard-prone zones are essential for risk mitigation and sustainable development planning. This study undertakes a comparative assessment of landslide susceptibility using two distinct approaches, the Analytical Hierarchy Process (AHP), a knowledge-driven multi-criteria decision-making method, and the Random Forest (RF) machine learning (ML) algorithms, a data-driven ensemble technique. A spatial database is developed, incorporating multiple Landslide Explanatory Variables (LEV) such as slope, aspect, curvature (plan and profile), distance to streams, topographic wetness index (TWI), geological units, structural features, lineament density, land use/land cover (LULC), soil type, and normalized difference vegetation index (NDVI). The model performance is evaluated using the area under the curve (AUC) metric. The landslide susceptible zonation analysis indicates a similar to 431.93 km(2) area under high (similar to 32 %) and 150.68 km(2) very high (similar to 12 %) in the RF. However, similar to 569.16 km(2) area under high (similar to 42 %) to 101.32 km(2) very high (similar to 8 %) in AHP. The RF model demonstrated a superior predictive capability (AUC 86 %), which outperformed to AHP model (AUC 70 %). Variable importance analysis revealed that slope, geology, and distance to streams are the most influential parameters controlling landslide dynamics. The comparative analysis between the two aforementioned techniques shows that the outcomes of RF are more reliable for landslide susceptibility analysis, owing to no subjectivity and bias compared to AHP. Overall, this study attributes the effectiveness of ML techniques for accurately marking the susceptible zones of the landslide and related hazards.
