Paavai College of Engineering (PCE ) is one of the college's under Paavai Institutions which was established in 2006 in Namakkal, Tamil Nadu, India. It is affiliated to Anna University, Chennai and a sister college of Paavai Engineering College.
Sybil and Blackhole attacks pose significant security threats in Wireless Sensor Networks (WSNs), as they disrupt communication, corrupt data, and reduce network efficiency. Traditional detection methods often suffer from high false positives and lack adaptability to dynamic network environments. A novel framework, Taylor-based Log Mean Squared Error loss function (TLMSE-MNet), is proposed to detect Sybil and Blackhole attacks in WSN, aiming to resolve these issues effectively. The framework optimizes energy-efficient routing and Cluster Head (CH) selection using the Fractional Caracal Optimization Algorithm (Fra-COA), which integrates the Caracal Optimization Algorithm (CAO) with fractional calculus. Furthermore, the input network log data is gathered from the databases and then normalized via Mean Normalization. The Fisher Score technique is utilized for feature selection. For detecting Sybil and black hole attacks, TLMSE-MNet is employed, which is established by updating the learning rule of MobileNetV2 with the TLMSE loss function, where TLMSE combines the Taylor series, Mean Squared Logarithmic Error (MSLE), and Logarithmic loss. The system is evaluated by simulating a WSN environment using SensorNetGuard, SmartCity Cybersecurity IoT, and Wireless Sensor Network Attack datasets, focusing on network behaviour, security events, and resource utilization. The results show significant improvements in attack detection, achieving an accuracy of 97.624%, a recall of 98.437%, and an F1-score of 97.426% at a feature dimension of 16, surpassing existing techniques in both detection accuracy and false positive reduction.
Carbon-filled polylactic acid produced by fused deposition modeling is a sustainable feedstock for lightweight structural parts, but performance depends on process-controlled porosity and interlayer fusion. The coupled effects of layer thickness, infill density, and print-path speed on simultaneous tensile, impact, and hardness performance in carbon–PLA systems remain insufficiently quantified. This study aims to define a multi-response analysis for carbon–PLA FDM. Eighteen coupon sets were fabricated in a factorial design spanning 0.15-0.25 mm layer thickness, 20-60
This study presents the development of an epoxy-based hybrid composite for prosthetic applications, reinforced with silane-treated basalt and bamboo fibers along with chitosan derived from Echinoidea spikes. The chitosan was extracted from marine waste and surface-modified using 3-aminopropyltrimethoxysilane (3-APTMS) to enhance interfacial bonding within the epoxy matrix. The composites were fabricated through a hand lay-up process followed by controlled curing and post-curing. Five formulations (M1–M5) with varying fiber and chitosan contents were evaluated for their mechanical and viscoelastic performance using tensile, dynamic mechanical analysis (DMA), and creep tests. Among them, specimen M4, containing 40 vol
Chitosan-doped CeO2-Ga2O3 nanocomposites were synthesised via a hydrothermal route and systematically characterised by X-Ray Diffraction (XRD), Fourier Transform Infrared (FTIR), UV–Vis, Photoluminescence (PL), Scanning Electron Microscopy (SEM), Transmission Electron Microscopy (TEM) and N2 adsorption-desorption analyses. The integration of chitosan improved surface area, particle distribution, and adsorption capability, whereas the CeO2-Ga2O3 heterojunction promoted effective charge separation. This unique integration of a biodegradable biopolymer with binary metal oxides provides a sustainable platform that combines environmental compatibility with high functional performance. The resultant composite demonstrated significant photocatalytic destruction of Methylene Blue (95.9
This study benchmarks four lightweight YOLO nano-variant object detection models, namely YOLOv5n, YOLOv8n, YOLOv11n, and YOLOv12n, using 701 thermal images of okra (Abelmoschus esculentus), comprising 355 adequately mature and 346 overripe samples. The models were evaluated using mean average precision (mAP@0.5−0.95), precision, recall, and inference latency across heterogeneous computing platforms, including GPU-accelerated parallel inference using an NVIDIA T4 GPU with TensorRT and sequential CPU execution using ONNX Runtime. This benchmarking addresses a critical gap in high-performance computing (HPC) deployment for non-RGB modalities, where real-time throughput exceeding 600 FPS requires GPU-level parallelism and optimized operator fusion. Under a 10-epoch training protocol designed for resource-efficient deployment, YOLOv8n achieved the highest detection accuracy with a mean mAP of 66.3 ± 0.3
