ABSTRACT The problem of root rot in hydroponic lettuce production is increasing and poses a threat to crop productivity and food security in controlled environments. Precise prediction is also problematic, as genotype-environment interactions follow a nonlinear, time-dependent model that includes nutrient concentration, pH, dissolved oxygen, temperature, and root-zone conditions. This paper seeks to overcome such shortcomings by developing the best genotype-environment interaction model, the Circular Motif Dilated Graph Frilled Lizard Heterogeneous Convolutional Attention Network (Cir-MDGF-LHCAN). Lactuca sativa (Rex variety) was experimentally tested by growing under a Nutrient Film Technique (NFT) hydroponic system. Signal stabilization via Cumulative Curve Fitting Approximation (CCFA) and feature modeling via Adaptive Causal Decision Transformer (AdaCred) were used as data preprocessing approaches to sequence the genotype-environment dependencies of a data set. The scheme will also combine a Circular Dilated CNN (CDIL-CNN) as a temporal modeling network, a Motif-Based Heterogeneous Graph Attention Network (MBHAN) as a relational learning network, and Frilled Lizard Optimization (FLO) as a parameter optimizer. The experimental scores are shown to be predictive, with 99.94% accuracy, superior to current machine learning and deep learning baselines, and yielding lower error measures (MSE, MAE, and RMSE). The statistical test confirms that performance is significantly improved (p < 0.001). The framework proposed offers an effective, scalable system for early detection of root rot in hydroponic environments and introduces a new concept for modeling intricate genotype-environment interactions in precision agriculture.

ABSTRACT Early detection of heart disease is essential for minimizing mortality, particularly in resource-constrained health-care environments. Recent models have shown accurate performance for medical diagnosis; however, they often struggles from limited interpretability and sensitivity to missing clinical data. To address these challenges, this paper proposes A Label-Enhanced Graph Neural Network Optimized by Elk Herd Optimization for Heart Disease Detection (HDD–LEGNN–EHOA). The input data is taken from Cleveland Heart disease data collection. Initially, data are pre-processed using Generalized Correntropy Sparse Gauss–Hermite Quadrature Filter (GCSGQF) to impute missing values. Then, the processed data is given to Secretary Bird Optimization Algorithm (SBOA) to select optimal characteristics. Then, the selected attributes fed to Label-Enhanced Graph Neural Network (LEGNN) to detect heart disease as heart disease present and heart disease absent. Elk Herd Optimization Algorithm (EHOA) to fine-tune the weight parameter of LEGNN. The newly suggested HDD–LEGNN–EHOA framework is assessed based on some metrics like accuracy, sensitivity, and computational time. Finally, the performance of HDD–LEGNN–EHOA method provides attains 25.68%, 22.54%, and 31.24% higher accuracy when compared to the existing models, respectively.

ABSTRACT This paper presents a reliability-based design optimization (RBDO) specifically applied to the design of low-speed wind turbines (LS-WT). The research explores the impact of uncertainty on optimum blade design by using metaheuristic (MH), integrating blade element momentum (BEM) theory for aerodynamic analysis and finite element analysis (FEA) for structural evaluation. Uncertainties, including tip speed ratio and allowable stresses, are modeled as truncated normal distribution. The multi-objective optimization problem is minimized total mass while maximized power output and the reliability index. Constraints include buckling factor, displacement, and the probability of failure. The optimum latin hypercube sampling (OLHS) is utilized to quantify uncertainty, compute the probability of failure, and reliability index. The distribution of uncertainties was investigated to assess its influence on optimum results. Findings show that reducing uncertainty levels enhances Pareto fronts while maintaining reliability index.

ABSTRACT The Internet of Things (IoT) is based on building an environment to interconnect a vast number of heterogeneous physical objects or things to the internet for enhancing the efficiencies of various types of applications through high-level computing. These types of ecosystems can be visualized as integrations of platforms and services managing the flowing of real-time data streams from the low-level stand-alone IoT devices and taking intelligent decisions. As the prime technology involved in these systems is the internet, most of the time it becomes difficult for the designers to figure out the suitable hardware, software, communication and networking protocols, and computing platforms for different applications. Nowadays, IoT systems suffer from additional challenges like bandwidth, latency, uninterrupted services, location constraints, climatic constraints, resource constraints, and security. To overcome some of the major problems, the authors have tried to provide a concise idea about different types of frameworks, architecture, services, and platforms that have been utilized in various IoT use cases for the last decade. This article contains the description of IoT systems from the viewpoint of different network models to various wireless standards, platform integration tools, and cloud platforms which have been successfully utilized for different IoT applications. The prime objective of this extensive review is that it can act as a reference for the designers while selecting conventional as well as contemporary technologies suitable for any type of complex IoT system design.

ABSTRACT Cognitive radio for vehicular ad hoc networks (CR-VANET) plays a key role in managing the spectrum dynamically and provides data communication in smart transportation. However, the security aspect is threatened by Sybil attacks where the adversary creates multiple fake identities in the network to hinder performance, damage topology, and mount denial of service attacks. To address these challenges, we present Vision-Augmented Split-Attention Neural Architectures for Sybil Resilience via Chaos-Driven Secure Elliptic Key Synthesis to Assured Data Exchange in CR-VANETs (Fuzz-CViAt_DuBe), a new approach. The proposed comprises of (i) Cluster Head selection with the Sooty Tern Maximizer for efficient communication; (ii) Sybil attack detection using Convolutional Neural Networks Augmented by Vision Transformers with split attention enhanced by the Dung Beetle Adaptive Optimizer; and (iii) Cryptographic security with the Fuzz-Resilient Chaotic Elliptic Curve Cryptographic Infrastructure. In simulations it is seen that there is a very much enhancement in the network performance. The proposed system increases the packet delivery ratio by 97.4%, improves throughput by 95.8%, and reduces latency by 88.3%. Additionally, the security rate is enhanced by 98.5%, while encryption time for 100KB of data is reduced to 15.2 s, demonstrating its superior performance over existing models. These findings highlight the benefits of the Fuzz-CViAt_DuBe framework in protecting CR-VANETs against Sybil attacks and enabling safe communication, providing solid grounds for improved future intelligent transportation systems.

ABSTRACT Micro-coordinate measuring machines (micro-CMMs) are essential for high-precision dimensional metrology in micro-device manufacturing, but their tactile probes often suffer from dynamic instability due to small size and high compliance. This paper proposes a two-stage optimization framework for a kirigami-inspired micro-CMM probe that integrates the Linearized Inverse Eigenvalue (LINE) method with Grey Relational Analysis (GRA). In the first stage, LINE is used as an inverse eigenvalue tool to tune three key geometric parameters (sheet thickness, hinge width, hinge length) so that the fundamental natural frequency is shifted to a higher, target range without resorting to trial-and-error design sweeps. In the second stage, GRA aggregates three competing performance metrics – natural frequency, tip compliance, and maximum von Mises stress – into a single Grey Relational Grade to identify the best compromise design. Finite element analysis (FEA) shows that the optimized probe achieves a 41% increase in fundamental frequency (from 512 Hz to 725 Hz) and a 22% reduction in maximum stress (from 425 MPa to 330 MPa under a 5 mN load), while maintaining adequate compliance for low-force probing. Quasi-static strain-gauge experiments on the fabricated probe confirm the predicted deformation patterns and directional compliance, supporting the proposed design methodology. The two-stage LINE–GRA framework provides a systematic and generalizable approach for designing high-performance compliant mechanisms in precision micro-metrology.

ABSTRACT The ever-increasing power demands have led to the need to integrate conventional and renewable power sources to meet the requirements. The key concern during such integration is the frequency fluctuations during varying loads. This paper proposes an optimized cascade proportional integral and proportional derivative (PI-PD) algorithm based on the flower pollination algorithm (FPA) for automatic generation control of a six-source two-area hybrid power system. The various power generation sources include thermal, hydro, wind, gas, diesel, and solar power generators. PI controllers provide improved system dynamic response, low cost, and design simplicity, however, it might result in slower response times and to combat this, a PD controller is cascaded with the PI controller. The gains for the cascade PI-PD controller are tuned by the FPA. This technique follows the principle of keeping the fittest product alive. The superiority of the dynamic behavior of the suggested technique is inspected by comparing it with the various controllers that are optimized using different algorithms. In addition, the robustness of the suggested controller is also examined by subjecting the system to disturbances like communication time delay (CTD), generation rate constraint (GRC) and governor dead band (GDB). It is observed that the cascaded PI-PD controller based on FPA achieves control over the frequency in the optimum settling time as well as remains stable and consistent over external disturbances.

ABSTRACT This research investigates the optimization of density and porosity in AA8090/TiH2–Al2O3 composite foams fabricated via FSP. Five key parameters – heat treatment temperature, holding time, tool rotational speed, traverse speed, and tool tilt angle were optimized using the Taguchi L27 orthogonal array. The foam density and porosity were determined through Archimedes’ principle, revealing their strong interdependence and influence on the mechanical performance. Experimental results indicated that density varied significantly with processing conditions, decreasing from 1.362 g/cm3 at 560°C, 3 min, 600 rpm, 60 mm/min, and 0° TTA to 0.909 g/cm3 at 600°C, 6 min, 900 rpm, 60 mm/min, and 2° TTA, corresponding to an increase in porosity from 49.56% to 66.33%. The highest density (1.42 g/cm3) was achieved at 640°C, 4.5 min, 600 rpm, 100 mm/min, and 0° tilt, indicating a dense and stable cellular structure. Analysis of variance (ANOVA) confirmed TRS (F = 26.21, p < 0.001) as the most dominant factor influencing density, followed by TTS (F = 11.45, p = 0.001) and TTA (F = 6.61, p = 0.008). Microstructural examination using SEM revealed that samples reinforced with Al2O3 exhibited uniformly distributed pores, enhanced structural integrity, and reduced pore coalescence compared to samples without Al2O3. The alumina-stabilized foams demonstrated a porosity of 60.63% and reduced density of 1.01 g/cm3, compared to 41.70% porosity and 1.49 g/cm3 in foams without alumina, validating the effectiveness of Al2O3 as a stabilizing phase. The novelty of this work lies in developing AA8090/TiH2– Al2O3 composite foams via FSP using an optimized Taguchi L27 design. The combined addition of TiH2 and Al2O3 refines pore structure, enhances cell wall integrity, and achieves uniform, lightweight foams with controlled density for advanced structural and aerospace applications.

ABSTRACT This study presents the design and development of a novel lightweight bicycle brake disc that offers significant improvements in weight reduction, thermal performance, and aesthetic customization compared to conventional models. Traditional bicycle brake discs, typically made entirely of chromium-molybdenum steel, are relatively heavy and susceptible to excessive heat buildup during prolonged braking. The newly designed brake disc features a composite structure, incorporating a chromium-molybdenum steel outer frame for mechanical strength and a lightweight aluminum alloy core to reduce overall weight while maintaining structural integrity significantly. The brake disc design enhances aesthetics through an anodized aluminum alloy center frame, allowing for vibrant color customization tailored to fleet and brand preferences. More importantly, the riveted joint structure between the steel outer frame and the aluminum core acts as a thermal insulation, increasing thermal resistance and preventing excessive heat transfer to the hub. This design effectively reduces thermal deformation risk, improving braking reliability and rider safety. This study investigates the thermal properties and fatigue crack growth behavior of a proposed composite brake disc through a combination of experimental testing and numerical simulations. A comparison between the traditional brake disc (Model A) and the novel composite brake discs (Models B, C, and D) reveals that Model D demonstrates the most effective thermal isolation. In addition, structural evaluations indicate that Model D significantly outperforms Model A in terms of deformation, stress, fatigue life, damage, and safety factors. Model D demonstrates a significantly higher maximum fatigue life of 108 cycles, which is 100 times greater than that of Model A. These findings provide valuable insights for optimizing performance, enhancing safety, and improving the user experience in modern bicycle braking systems.

ABSTRACT Intrusion Detection System (IDS) is a major concern in network security. Deep learning-based intrusion schemes have emerged as a part of network security research. This paper proposes an Anti-Interference Dynamic Integral Neural Network for Intrusion Detection System in Sensor Networks (AIDINN-IDS-SN). The data are taken from KDD CUP99 dataset. The collected data is fed into the pre-processing stage. During pre-processing, the Regularized Bias-Aware Ensemble Kalman Filtering (RBEKF) is used to normalizing data. Then the pre-processed data is supplied into the feature extraction stage. In this stage, spatiotemporal features, like objects position, velocity, optical flow, crowd density, using Modified Spline-Kerneled Chirplet Transform (MSKCT) are extracted. Finally, Anti-Interference Dynamic Integral Neural Network (AIDINN) is employed to classify intrusion detection as normal, probe attack, U2R attack, DoS attack, R2L attack and uncertainty. The efficiency of the AIDINN-IDS-SN is evaluated under some performance metrics, like accuracy, precision and detection rate. The proposed AIDINN-IDS-SN attains 21.65%, 20.32%, 21.19% better accuracy, 23.54%, 21.24%, 21.57% better precision when compared with existing methods respectively.