Accurate real-time coral reef bleaching classification is challenging due to the high computational cost and limited interpretability of conventional convolutional neural networks (CNNs), which restrict deployment on resource-constrained edge devices. To address this, we propose Fibonacci-Net (F-Net), a lightweight and interpretable CNN that integrates Fibonacci-based filter scaling, a patch-based hybrid area-attention mechanism to enhance fine-grained coral features, and a particle swarm optimization-Adam hybrid optimizer for stable learning on small, imbalanced datasets. Evaluated on 7384 coral images, F-Net achieves 97.6% accuracy, better than some well-studied CNN models in the literature. The novel gradient-weighted class activation mapping and filter discriminability analyses further enhance interpretability, demonstrating F-Net's effectiveness and deployment readiness for large-scale autonomous coral reef monitoring.

Integrating processes, commonly found in industrial systems like liquid level and thermal control, present significant control challenges due to their non-self-regulating nature and sensitivity to input disturbances. This paper proposes a cascade control strategy tailored for such processes, focusing on effective disturbance rejection. The approach employs a modified dual Smith predictor (SP) architecture, integrating Internal Model Control (IMC) principles in both inner- and outer-loop designs. The inner loop employs a standard IMC method to address dominant dynamics, while the outer loop utilises a target-loop formulation to shape the desired open-loop behaviour. The key design parameters, the IMC filter, and the target loop gain are tuned by using the maximum sensitivity selection to balance performance and robustness. In comparison with other dual-loops and recent schemes, the new control structure and tuning approach have proven to be more robust and easier to tune. The experimental test of a two-tank level system demonstrates superior disturbance rejection and set-point tracking in real-time.

This article presents a novel approach to magnetic resonance coupling wireless power transfer (WPT). The proposed design implements a capacitive element with a fractional impedance (CEFI) of the order of less than unity in the receiving circuit. Existing research has primarily explored fractional-orders greater than one. However, experimental results demonstrate significant benefits from subunity fractional elements. Mathematical analysis and experimental validation of a series-series compensated system show promising results. With a CEFI of the order of 0.98, the system achieves a 169% increase in dc power output and a 137% improvement in ac power output compared to a similar classical system. The design maintains dc-dc efficiency and experiences only a 5% reduction in ac-ac efficiency at high coupling coefficients. These findings establish subunity fractional impedance components as viable solutions for WPT performance improvement.

This work proposes a straightforward fractional-order control scheme for industrial time-delay systems, where these systems can be approximated as an integrating plus time-delay model. The proposed scheme can improve performance and robustness without requiring the solution of complex equations or the execution of extensive parameter searches. Controller synthesis is conducted analytically, ensuring compliance with performance and robustness criteria in both the time and frequency domains. The presented method is simplified analytically with only two parameters for selection, and other parameters can be calculated using explicit relations. After validating the scheme on nonlinear and higher-order plants, it was further validated on two-area time-delayed cyber-physical and multi-energy standalone microgrid systems. In these studies, the method has effectively responded to cyberattacks and communication time delay issues. After numerical comparisons, the new controller is also validated on a two-tank system to verify its ability towards real-time implementation.

A complex-order internal-model-control (CoIMC) method is developed in this paper. The CoIMC has only three tuning parameters, which are readily solved using well-adopted tuning approaches with phase margin (ϕm) and gain margin (gm). The new structure is verified together with fractional order IMC and normal IMC filters. Significant improvements are observed in step, ramp characteristics, internal and external regulatory performances and control actions.

This study introduces and evaluates the novel Riemann-Liouville (RL) conformable fractional derivative-based Adaptable-Shifted-Fractional-Rectified-Linear-Unit ((RL)ASFReLU) activation function within multi-layer perceptron models, showcasing its superior performance across three distinct nonlinear benchmark systems. (RL)ASFReLU enhances identification accuracy by integrating a robust bias shift via the input standard deviation and extending the fractional order to alpha is an element of [0, 2], offering greater flexibility than existing fractional ReLU variants. It demonstrates faster convergence and improved ability to capture complex nonlinear patterns. The proposed function is adaptable across multiple machine learning domains, providing a versatile tool for advancing neural network design.

This paper introduces an innovative hybrid method for the classification of Gleason Scores (GS) in prostate cancer using Whole Slide Images. The researchers combine Deep Learning and Machine Learning (ML) techniques to automate the precise classification of GS. They employ a specially designed Variational Autoencoder for feature extraction, utilizing a pre-trained VGG16 Convolutional Neural Network to build the encoder. Principal Component Analysis is then used to reduce the dimensionality of the feature vector to 50 significant features for further Gleason Grade classification. The study uses the SICAPv2 database and evaluates feature importance with Shapley Additive explanations (SHAP). Comparative analysis of five ML techniques and two custom-designed Deep Neural Network (DNN) architectures shows that the Support Vector Machine algorithm with hyperparameter tuning and the custom-designed five-layer DNN architecture achieved accuracies of 84

Thermodynamic and surface properties of Ag–Al–Au–Cu liquid alloy were computed and analysed in the temperature range 1573–1873 K using different geometrical models. The energy interaction parameters of constituent binary subsystems of Ag–Al–Au–Cu were optimised in the framework of the Redlich–Kister polynomial using experimental data of enthalpy of mixing and excess entropy of mixing. The thermodynamic and surface properties of binary and ternary subsystems of the Ag–Al–Au–Cu system were also calculated to explore their effects in quaternary complex formation. The excess Gibbs free energy of mixing of the quaternary system was computed using General Solution Model, Kohler model, and Toop model at different cross-sections. These values calculated were found to be consistent with each other. The mixing tendency of the system was further validated by the computing activities of its monomers. The surface properties of the systems were computed using Butler model.

A stealthy cyberattack (SCA) aims to gradually drive the system toward instability. Although research is continuing in attack detection algorithms, a resilient control strategy is vital if SCAs go undetected for a considerable time. Especially for a controller involved in automatic generation control (AGC), resiliency becomes more crucial as the controller has to sometimes work under abnormal conditions considering that shutdown of operation is not an option. This paper introduces an innovative approach to support dual-area thermal power systems in the face of network-related challenges such as transference latency (TL) and SCAs. We propose a new tri-parametric fractional controller (TFC) that merges the advantages of proportional-integral and proportional-derivative operations, eliminating the need for additional control loops and enhancing system efficiency. The TFC's efficacy in mitigating random and step load disturbances is established, outperforming existing controllers. We employ the complex root boundary method to define the optimal parameter search space, with an improved equilibrium optimizer algorithm determining the best controller settings. The inclusion of renewable energy sources, including solar and wind, is considered, and the robustness of the system is evaluated using uncertainty norms and complementary sensitivity functions. Our approach is validated through hardware-in-loop real-time verification on the OPAL-RT platform, demonstrating superior performance in maintaining frequency stability under SCAs and TL challenges.

In the framework of Redlich-Kister (R-K) polynomials, both linear and exponential temperature-dependent energy interaction parameters for constituent binary subsystems of liquid Al-Ga-Mg alloys were optimised. These parameters were then used to estimate the excess Gibbs free energy of mixing, enthalpy of mixing and activity of the ternary system at 923 K, 973 K, 1023 K, 1073 K and 1123 K employing General Solution Model (GSM), Toop and Kohler models. The surface tensions of binary systems were calculated using Butler's equations. The viscosity of binary subsystems was calculated employing Kaptay's equations with the aid of exponentially optimised coefficients of R-K polynomials. Furthermore, the surface tension and viscosity of the ternary mixing were computed using aforementioned frameworks within the temperature range 923 K-1123 K. Theoretical investigation showed that the excess Gibbs free energy of mixing for the Al-Ga-Mg system was found to be negative, indicating a compound forming tendency.

In many real-world applications, sequential data exhibit additive and multiplicative dependencies between features and their temporal context. Traditional Long Short-Term Memory (LSTM) networks update their state through additive interactions modulated by multiplicative gates, which can limit flexibility when stronger feature interactions are present. Conversely, architecture such as the Multiplicative-integration Recurrent Neural Network (Mi-RNN) relies purely on multiplicative fusion, often sacrificing stability and interpretability. We introduce FlexGate-LSTM, a recurrent architecture that adaptively blends additive and multiplicative operations inside each gate. A learnable parameter vector in each gate continuously tunes the trade-off between the two interaction modes, allowing the network to select the most suitable integration strategy for the current task or time step. The proposed FlexGate-LSTM is evaluated in five controlled synthetic scenarios: additive, multiplicative, conditional, noisy, and non-stationary, where it matches or exceeds the performance of both vanilla LSTM and Mi-RNN baselines. To demonstrate real-world usefulness, we further test the model on the ETTh1 electricity-transformer temperature dataset. The investigation shows that the FlexGate-LSTM performs competitively in specialized cases and significantly outperforms other architectures when the data exhibits mixed or uncertain dynamics. In addition, the analysis of the learnt parameters provides insight into the internal adaptation strategies of the model, improving the interpretability.

Analyzing fractional-order wireless power transfer (WPT) systems through scattering parameters reveals several advantages over their classical counterparts. Compared to the fixed value in classical systems, the most notable improvement is the variable nature of the critical coupling coefficient in fractional systems. This contrasts with the classical system, which shows optimal performance only under near-perfect coupling conditions and within a narrow frequency band around the resonant frequency. The efficiency analysis reveals that the fractional system maintains a higher transfer efficiency at lower coupling coefficients by leveraging higher frequency operations, a capability not present in classical WPT systems. The ability to optimise performance through frequency adjustment rather than physical coupling modifications represents a significant advancement in wireless power transfer technology.

This paper introduces a novel Riemann–Liouville (RL) conformable fractional derivative based Adaptable-Shifted-Fractional-Rectified-Linear-Unit, briefly called RLASFReLU, and evaluates its efficacy in enhancing the performance of convolutional neural network (CNN) models for pneumonia and skin cancer detection. The study conducts a comprehensive comparative analysis against traditional activation functions and state-of-the-art CNN architectures. The results show that RLASFReLU consistently outperforms other functions, achieving higher accuracy. Comparative evaluations with various neural network architectures reveal that the model equipped with RLASFReLU exhibits superior performance despite its simplicity and fewer trainable parameters, highlighting its efficiency and effectiveness. The findings suggest that RLASFReLU holds promise in improving diagnostic accuracy and efficiency in medical imaging applications, contributing to advancements in healthcare technology and facilitating better patient care. The proposed fractional nonlinear transformation can offer high performance with reduced computational cost, making it practical for deployment in healthcare settings.

This research improves the output of the integrating processes. The technique regulates a double-integrating process with time delay and non-minimal phase while handling high parametric uncertainties and load disturbances. A Smith predictor (SP) architecture based on a fractional-order internal model controller is developed (FOSPIMC). The gain and phase margins are applied to tune the fractional-order parameter. The fractional filter time constant is calculated using the intended performance limitation. Numerical studies and comparisons reveal better performance with a hybrid structure.

One major concern in control engineering is the problem of introducing an unstable system. Such systems are even more sensitive to ramp input changes, either set-point or disturbance. We have proposed an extended fractional-order IMC (FOIMC) control for an unstable system exhibiting a time delay. The complete design involves only three adjustable parameters, such as PID tuning. In the proposed structure, the inner loop control stabilises the system, whereas the fractional IMC filter improves the overall performance, together to tackle the ramp inputs. A systematic approach is developed to tune the required design parameters to obtain the desired peak of the sensitivity function and stability margins. The proposed control is simple and can easily calculate the FOIMC parameters from the explicit formulae. The method works under practical considerations, such as process parameter perturbations and load disturbances. The developed scheme is also tested on the nonlinear continuous stirred tank reactor system. The proposed control method results in a percentage enhancement of 61.9% under the perfectly ideal condition (when the process model is equal to the actual plant) whereas 81.3% enhancement is obtained when the process parameter variations are considered for the CSTR system.

In the case of a minimum-phase system, output reacts to changes in the input as quickly as feasible. In contrast, a non-minimum phase (NmP) system is dynamic in which the output reacts more slowly to input changes. In the literature, it is seen that such processes are difficult to handle with classical control structures. This paper proposes a fractional order controller with the modified Smith predictor structure. The inner loop is designed with a proportional derivative, making the plant more stable. The outer loop is then developed with a fractional-order tilt integral derivative type. This new control structure can be designed using the well-known phase-margin and maximum sensitivity specifications. A fractional-order target loop uses simple explicit relations according to the stability margins and plant parameters. Using just two tunable parameters, the inner and outer controllers are designed. In addition, the suggested design is capable of controlling various types of industrial processes dominant in delays, such as stable, integrating, unstable, and non-minimum phase systems. After a numerical investigation, the method is also practically verified in the two-tank plant. The performance analysis indicates that the suggested scheme can provide a balanced control between setpoint tracking with robustness, even with large parameter perturbations.

A theoretical analysis of the thermodynamic, structural, and surface properties of liquid Fe–Si alloys has been conducted at various elevated temperatures using R–K polynomials. The thermodynamic parameters, including the free energy of mixing, heat of mixing, entropy of mixing, and activity of the components in Fe–Si alloys, were found to be in excellent agreement with the relevant experimental values. The structural behavior of the alloys indicates that the alloy has an ordering nature. Strongly ordered alloys produce more stable and homogeneous joints, which is desirable in welding. In addition, it was found that the surface tension of the alloy reduces as the temperature increases. The surface concentration of silicon was determined to be higher than the surface concentration of iron.

Thermodynamic functions like excess Gibbs free energy of mixing (∆GxsM) and activity (a) of liquid Al-Sr alloy were studied in the temperature range 1323-1623 K in the frame work of Redlich-Kister (R-K) polynomial. The compositional and temperature dependence of structural functions like concentration fluctuation in long wavelength limit (Scc(0)), Warren-Cowley short range order parameter (α) and ratio of mutual to intrinsic diffusion coefficients (Dm/Did) were computed and analysed using the same approach. The surface tension (σ) and surface segregation (xsi) tendencies of the atoms in the liquid mixture were computed and studied using Butler’s model. Present investigations revealed that association tendency among the atoms of metallic mixture gradually decreased with increase in temperature.