Journal of Robotics and Control (JRC)

Optimizing Input Shaping for Flexible Beam Vibration Control Using Self-Adaptive Differential Evolution

Phuong-Tung Pham, Thanh Huy Phung, Quoc Chi Nguyen — Sat, 03 May 2025 00:00:00 +0700

This study develops a control strategy for the flexible beam linked to a moving hub utilizing input shaping control. The input shaping control technique is an open-loop control approach that employs a shaped command to suppress the undesired vibration. This command is formed by convolving the original command with input shapers (a sequence of impulses with amplitude and temporal location). Unlike the conventional input shaping control, which calculates the input shapers based on the system's natural frequencies and attenuation ratios, a metaheuristic input shaper searcher based on the self-adaptive differential evolution algorithm is employed in this paper to identify the optimal input shapers. Using this algorithm, the specifications of input shapers, including the amplitudes and time locations, can be optimized to ensure that the cost function corresponding to the position error and beam’s vibration approaches the global minimum value. The control performance is proved via the numerical simulation. The simulation results demonstrate that input shaping control utilizing optimized input shapers can significantly reduce residual vibrations in the beam. While this control strategy requires substantial computational resources and longer computation times to develop the optimal input shapers compared to traditional techniques, the effectiveness of the optimal input shapers in attenuating vibrations is remarkable.

Heart Disease Prediction Using Ensemble Methods, Genetic Algorithms, and Data Augmentation: A Preliminary Study

Deepali Yewale, Swati Patil, Archana Rajesh Date, Aziz Nanthaamornphong — Tue, 06 May 2025 00:00:00 +0700

Statistically speaking, heart disease (HD) accounted for 1 in 5 fatalities in 2022, demanding affordable and accurate diagnosis. Traditional methods of prediction are accurate but expensive, creating a demand for sophisticated and efficient technologies. One of the most popular methods that researchers employ to forecast diseases is machine learning (ML). The goal of this effort is to improve HD prognosis accuracy through the use of ensemble approaches, specifically Random Forest (RF), XGBoost, Voting, and Stacking methods, which improve prediction accuracy by combining multiple models to capture complex patterns. Genetic algorithms (GA) are used to prioritize features. Incorporating data balancing, outlier removal techniques, and data augmentation, creates a model that delivers performance comparable to state-of-the-art research. Methods like random oversampling address data imbalance, while an isolation forest is employed to identify anomalies. To increase the dataset size and improve model performance, random noise is added after anomaly removal. Performed the cross-validation and robustness checks to assess the model's performance on both augmented and non-augmented datasets, ensuring that the inclusion of random noise did not excessively affect generalizability or result in overfitting. The proposed model’s effectiveness is evaluated using various performance metrics. Achieving 99.36% accuracy, 98% sensitivity, 100% specificity, 100% PPV, 97% NPV, 0.99 F-score, and an AUC of 1, the methodology shows great promise as a cost-effective, accurate, and highly efficient diagnostic tool for heart disease. The model's short training time and high performance suggest its potential for practical implementation in clinical settings, offering a reliable and affordable solution for early heart disease detection.

Deep Learning-Based Continuous Sign Language Recognition

Thu, 08 May 2025 00:00:00 +0700

This study focuses on the development of a continuous sign language recognition system based on deep neural network models. A new Kazakh Sign Language (QazSL) dataset is created. DL models for continuous KazSL are developed, their accuracy and robustness under different environmental conditions are analyzed, and an optimized model algorithm to improve sign recognition processes are proposed. The main goal is to improve gesture recognition accuracy, account for gesture variability and environmental conditions, and promote the development of adaptive technologies for low-resource languages. This paper proposes a QazSL recognition system using an YOLOv8n and optimized 2DCNN models to improve accessibility for the hearing impaired. The optimized 2DCNN method includes optimal data preprocessing techniques and new training architecture, followed by model training and testing with precision, recall, and accuracy metrics. The proposed systems were trained using an opencourse K-RSL dataset with 5 signers and a newly created QazSL dataset, recorded by 7 signers. The test accuracy of gesture recognition are 98.12% for Yolov8n and 98, 57% for 2DCNN, indicating the robustness and capability of the models for realtime application. Certain issues, such as background variation and gesture consistency, were found to affect recognition under different conditions. This research contributes to the development of AI-based assistive technology to facilitate social inclusion and access to communication for deaf and hard-of-hearing people. By addressing the challenges identified in gesture recognition, this study paves the way for more reliable interactions between users and technology. Future work will focus on optimizing the model further to enhance its performance in varied environments and to expand its applicability across different languages and sign systems.

Efficiency Enhancement in SynRMs Using MTPW Control and Seven-Level NPC Inverter

Sara Elbadaoui, Ahmed Abbou, Yassine Zahraoui, Farid Oufqir — Thu, 08 May 2025 00:00:00 +0700

Multilevel inverters have emerged as a key research focus in power electronics due to their increasing importance in renewable energy systems and rotating machinery applications. These devices produce output voltages that closely approximate sinusoidal waveforms, significantly improving signal quality. Among available topologies, the Neutral Point Clamped (NPC) inverter proves particularly suitable for such applications, especially when numerous voltage levels are required. Our study examines a Synchronous Reluctance Motor (SynRM) driven by a sevenlevel multilevel inverter employing Maximum Torque Per Weber (MTPW) control. This approach achieves outstanding dynamic performance by directly linking torque control to current control. The selection of control methodology depends fundamentally on how reference current values are determined. Through comprehensive MATLAB/Simulink simulations, we performed a comparative analysis of conventional inverter characteristics. The results conclusively demonstrate superior performance in response time, torque ripple reduction, and current waveform quality enhancement.

A Hybrid Deep Learning Approach for Adaptive Cloud Threat Detection with Integrated CNNs and RNNs in Cloud Access Security Brokers

Israa Basim, Ahmed Fakhfakh, Amel Meddeb Makhlouf — Thu, 08 May 2025 00:00:00 +0700

Cloud computing offers on-demand, scalable, and cost-effective deployment models but also struggles with sophisticated and rapidly-evolving cybersecurity threats. Static, rule-based approaches to data moved by traditional Cloud Access Security Brokers (CASBs) are seldom able to detect these threats. In this work, we introduce Adaptive CASB a new framework built on a new hybrid deep learning architecture combining Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). CNNs learn spatial features in network traffic and RNNs find temporal dependencies, leading to robust static and dynamic threat detection. The system combines behavior-based anomaly detection with real-time threat intelligence applied to the Internet, providing adaptability to new attacks such as zero-day attacks. Experiments on benchmark datasets (e.g. NSL-KDD, UNSW-NB15) prove that our model outperforms the others with accuracy of 95%, precision of 92% and recall of 94%, which is significantly better than CASBs based on traditional techniques and machine learning models. Moreover, the automated threat response capabilities of the system send alerts and implement containment measures that mitigate threats in real-time. Such an Adaptive CASB framework signifies a scalable and cost-effective response to contemporary cloud security challenges, whilst also paving the way for future advancements, such as XAI integration and edge-computing optimization.

Smart Innovations in Food Spoilage Detection: A Focus on Electronic Nose, Machine Learning and IoT for Perishable Foods

Fri, 09 May 2025 00:00:00 +0700

This review article provides a comprehensive analysis of advanced technologies for detecting, analyzing, and controlling food spoilage, with a focus on perishable foods such as fruits, vegetables, and meats. Although traditional methods such as microbiological testing and sensory evaluation remain fundamental, emerging technologies such as machine learning (ML), computer vision, and electronic noses (enoses) offer transformative potential for real-time monitoring and predictive analytics. However, practical implementation of these technologies faces significant challenges, including heterogeneity in data, computational constraints, and environmental variability. For example, ML models, particularly deep learning architectures, require extensive labeled datasets and high-performance computing resources, which are often inaccessible in resource-constrained settings. Similarly, electronic noses, while effective in detecting volatile organic compounds (VOCs) associated with spoilage, suffer from sensor drift and cross-sensitivity issues, necessitating frequent recalibration. Blockchain technology, though promising for improving traceability and transparency in the food supply chain, struggles with scalability and energy efficiency. This review critically evaluates these limitations, highlighting gaps in current methodologies, such as the overreliance on external spoilage indicators in computer vision systems and the lack of standardized protocols for data collection and model evaluation. By addressing these challenges, future research can advance the development of robust, scalable and cost-effective solutions for food spoilage detection, ultimately contributing to improved food safety, reduced waste, and enhanced supply chain efficiency.

Effect of Liquid Height on Sloshing Dynamics in Cylindrical Containers Using H-Infinity Control with Smooth Velocity Input

Sat, 10 May 2025 00:00:00 +0700

The research analyzes the dynamics of liquid sloshing in cylindrical containers, emphasizing the influence of liquid height on system stability and motion via H-Infinity Control methodologies. The main goal is to analyze how changes in liquid height affect the dynamics and stability of fluid transport systems and to assess the effectiveness of H-Infinity control in reducing sloshing effects. A simulation investigation was performed at different liquid levels using trapezoidal velocity profiles, including step inputs and gradual transitions. Performance was assessed using the Integral of Absolute Error (IAE) and Root Mean Square Error (RMSE). The results demonstrate that heightened liquid levels significantly improve sloshing dynamics, extend the settling time, and exacerbate inaccuracies in measurements due to increased fluid inertia. Smooth velocity profiles reduce sudden changes; yet, they cannot completely eliminate the destabilizing effects caused by large amounts of liquid. The study confirms a mechanical model for sloshing dynamics integrated with robust H-infinity control, providing significant insights for improving fluid management in robotics and automated systems. Subsequent research should encompass varied container designs, fluid characteristics, and sophisticated adaptive control methodologies.

Efficient Multimodal Biometric Identification via Gabor-Enhanced Attention Networks

Phan Minh Than, Hoanh Nguyen — Sat, 10 May 2025 00:00:00 +0700

Achieving robust multimodal biometric identification requires advanced feature extraction strategies and effective integration of diverse data modalities. Conventional methods often encounter limitations such as computational complexity and degradation of critical information during feature transformation. Although deep learning models address feature extraction challenges, their heavy architectures hinder real-world deployment. Moreover, traditional fusion strategies, based mainly on simple concatenation, overlook critical intermodal correlations, leading to suboptimal recognition accuracy. In this study, we propose a lightweight Gabor Attention Network framework designed for efficient multimodal biometric recognition. Our approach utilizes learnable Gabor filters to capture detailed local and directional features with enhanced precision and reduced computational burden compared to standard convolutions. We further introduce a convolutional attention mechanism that adaptively refines intermediate feature representations, and a novel attention-driven fusion architecture that dynamically models and exploits intermodal dependencies. Extensive experiments on two multimodal datasets demonstrate that our model achieves superior performance compared to several state-of-the-art methods, attaining up to 99.49% accuracy and 0.35% Equal Error Rate, while maintaining high efficiency with only 10.6M parameters, 0.85 GFLOPs, and 60 FPS inference speed. These results highlight the effectiveness of our biologically inspired and attention-enhanced design for achieving high-accuracy, low-complexity multimodal biometric identification.

Stability Control of Multi-Quadcopter Formation Based on Virtual Leader and Flocking Algorithm

Sat, 10 May 2025 00:00:00 +0700

This study aims to develop an efficient and stable formation strategy for multi-quadcopter systems, focusing on formation stability based on the number of flying quadcopter members. The formation strategy combines a virtual leader approach and flocking-based behavior to achieve consistent formation movement. The formations are designed as basic circular and elliptical patterns based on bearing measurement. Formation control in multi-quadcopter systems presents a complex challenge, as it requires coordination among autonomously flying UAVs while maintaining overall formation stability and reliability. A Twisted Sliding Mode Control (TSMC) system is implemented to ensure formation stability and responsiveness to predefined trajectory missions. After integrating TSMC, the Root Mean Square Error (RMSE) of position errors in the x, y, and z coordinates decreased by 0.02.

An Intelligent Fertilizer Dosing System Using a Random Forest Model for Precision Agriculture

Roger Fernando Asto Bonifacio — Mon, 12 May 2025 00:00:00 +0700

The inefficient application of fertilizers in horticultural crops, particularly in rural areas of Peru, leads to significant economic losses, soil degradation, and environmental risks. In response to this issue, this paper proposes an intelligent fertilizer dosing system that integrates solid and liquid fertilization applications through a predictive machine learning model. The main contribution of this research is the development and partial validation of an embedded system that dynamically adapts nutrient (NPK) doses based on real-time soil conditions, crop type, and phenological stage. The predictive model, based on Random Forest (RF), was trained using 10000 synthetic data points generated via Sobol-LHS sampling and validated with 1000 real field measurements. The method incorporates thirteen agronomic variables, including soil moisture, pH, temperature, and nutrient content, enabling adaptive control of the dosing mechanisms. The system achieved promising results, with root mean square errors (RMSE) of 2.81 kg/ha for nitrogen, 1.42 kg/ha for phosphorus, and 0.94 kg/ha for potassium. These results demonstrate the model’s ability to deliver accurate crop-specific fertilization recommendations, reducing input waste and improving nutrient use efficiency. Although full field trials are planned for future phases, the proposed system offers a scalable and low-cost solution for precision agriculture in resource-constrained settings, promoting more sustainable farming practices and enhancing the productivity of smallholder farmers.

Comparative Study of Linear and Nonlinear Controllers for DFIG-Based Wind Power Systems Under Different Operating Conditions

Mon, 12 May 2025 00:00:00 +0700

When doubly-fed induction generators (DFIGs) powered by wind energy are connected to the grid, unstable grid voltage causes distortion in the control of statoric active and reactive powers, especially if the controller uses this grid value for efficiency control, as well as parameter variation. Accordingly, this study focuses on evaluating the DFIG dynamics using different control topologies. The study presents a comparative analysis of linear and nonlinear control techniques for the DFIG, including both classical and robust controllers. A voltage converter based on Pulse Width Modulation (PWM) is employed to interface with the rotor, enabling independent control of active and reactive powers. Active and reactive powers are controlled using a linear proportional-integral (PI) controller and two types of nonlinear controllers: Backstepping (BSC) and Sliding Mode (SMC). This comparative study seeks to identify the most effective controller for tracking power reference, response to speed variations, sensitivity to external disturbances, and resilience against fluctuations in machine parameters. Three sets of evaluation tests are considered: normal operation with constant rotational speed while varying power references, robustness test under DFIG's parameters variation and rotor speed perturbation. The obtained results confirm the superiority of the nonlinear BSC and SMC approaches in comparison with the FOC, giving the BSC more priority than the SMC.

A Robust Speed and Torque Control of DC Motor with Cuk Converter Using PI and SMC

Tue, 13 May 2025 00:00:00 +0700

Robust speed and torque control of a DC motor powered by a DC/DC converter has become widespread attention recently. This research examines the Cuk converter's effectiveness in powering the DC motor using proportional-integral (PI) and sliding mode control (SMC) under three operating scenarios: variable speed (1600-500 rpm) and constant torque (20 N.m), constant speed(600 rpm) and variable torque (20-40 N.m), and variable speed (500-100 rpm) and variable torque (20-40 N.m). The research aims to provide accurate motor speed and torque control to enhance motor operations. PI and SMC controllers were constructed to investigate how the system operated in different scenarios, mathematical models were made, and Matlab/Simulink modeling was used. The performance parameters measurements are the speed and torque tracking response, armature current, and the output voltage from the Cuk converter with their total harmonic distortions (THDs). The results showed that SMC performed PI in speed and torque tracking and had fewer fluctuations under all scenarios. The SMC controller had a lower overshoot of 0.05 while PI was 0.75, and a settling time of SMC 0.5 seconds is less than the PI controller's 25 seconds in tracking speed and torque. For output converter voltage and armature current, the THD of the PI controller was 0.2441 and 0.3857, respectively, but the THD of SMC was reduced to 0.0833 and 0.0921. lower THD in SMC leads to smoother waveforms and less electromagnetic interference, resulting in faster responses, fewer overshoots, and improved speed and torque. The SMC with Cuk converter was the best control method for the DC motor drive applications, providing increased performance, efficiency, and decreased system losses.

Sensorless Hybrid Control System for Boost Converter in Presence of Uncertain Dynamics

Ibrahim K. Mohammed, Hajar K. Ibrahim, Salih M. Attya, Damain Giaouris — Tue, 13 May 2025 00:00:00 +0700

This study presents a design method for a voltage regulation system for a Boost converter that can be used in a power distribution unit within a power generation system. The regulation system is based on a hybrid, sensorless control approach, the structure of the controller is built based on the combination of PI and LQR controllers. The role of the new structural controller in improving the transient and steady-state response as well as enhancing the stability of the Boost converter output signal is studied. The states of the converter are estimated by Luenberger observer system, which is designed using pole placement (PP) technique. Mathematical model of the Boost converter with the hybrid LQR-PI controller is formulated. The gain parameters of the LQR-PI controller are obtained effectively by using Grey Wolf Optimizer (GWO) algorithm. In optimization process the GWO with an effective fitness function is used to tune the state and input weighting matrices of LQR controller. To validate the proposed control system a comparison between the performance of the LQR-PI controller and LQR controller with integral action (I) is achieved. The Boost converter circuit with feedback LQR-I/PI controllers are simulated utilizing Simulink software and their responses are assessed based on rise time, settling time overshoot and steady state error performance parameters. To verify the robustness of the control system, the performance of the converter is evaluated in five working scenarios under hard uncertainties in source voltage, reference voltage and resistive load. The simulation results demonstrate the effectiveness of the presented LQR-I/PI controllers in rejecting the effect of disturbances in the system response. However, the LQR-PI controller showed more accurate and stable output voltage compared to the LQR-I controller.

Bipedal Robots: A Systematic Review of Dynamical Models, Balance Control Strategies, and Locomotion Methods

Ibrahim Al-Tameemi, Oger Amanuel — Tue, 13 May 2025 00:00:00 +0700

Bipedal robots, designed to replicate human locomotion, face significant balance challenges due to instability and high degrees of freedom. This study examines dynamical models, balance control strategies, and locomotion methodologies. Dynamical models are categorized into simplified, centroidal dynamics, and whole-body dynamics models. Simplified models, such as the Linear Inverted Pendulum Model (LIPM), approximate the robot as a point mass at the Center of Mass (CoM) but neglect upper-body dynamics and complex terrain interactions. Centroidal dynamics models incorporate CoM motion, contact forces, and angular momentum for improved disturbance rejection but require extensive computational resources. Whole-body models achieve high fidelity by integrating joint torques and external forces but are constrained by computational complexity. Balance control methods for standing bipedal robots are classified into joint-specific and whole-body approaches. Ankle and hip strategies address small perturbations but are insufficient for real-world disturbances. Whole-body control utilizes all body segments to modulate contact forces and regulate momentum, enhancing stability against external disturbances, though challenges remain in force modeling and state estimation. Locomotion control is divided into model-based and learning-based approaches. Model-based strategies include LIPM and its extensions-based methods, Zero Moment Point (ZMP)-based methods, which ensure dynamic stability by maintaining moments within the support polygon; Capture Point (CP)-based methods, which predict foot placement to prevent falls; and Divergent Component of Motion (DCM)-based approaches, which adjust footsteps based on CoM divergence. While learning-based methods leverage Reinforcement Learning (RL) and human motion data for adaptive and energy-efficient gait generation. This study highlights challenges in energy efficiency, terrain adaptation, and scalability, proposing sensor fusion, energy-aware RL reward functions, and hierarchical control architectures as potential solutions.

Utilization of Convolutional Neural Network for Effective Recognition of Complex and Common Facial Emotions

Ammar Ibrahim Majeed, Suhad Qasim Naeem, Elaf A. Saeed — Wed, 14 May 2025 00:00:00 +0700

Facial expression recognition is an important area of computer vision used for human-computer interaction. The convolutional neural network model in this work was tested on the Fer-2013 dataset, and the experimental results demonstrated the superiority of the recognition rate. It is known that the Fer-2013 dataset contains data collected in an experimental environment, and to verify the generalization capability of model recognition, a self-made facial expression data set in a natural state was created, and the models are trained using this dataset to identify emotions from face photos, however, it has biases and limitations, including poor resolution (48 x 48 pixels) and class imbalance, which causes some emotions to be overrepresented. Additionally, it is devoid of demographic data, which may cause some groups to do poorly, furthermore, even though emotions are frequently mixed and context-dependent, it assumes that they are entirely distinct. More varied datasets, better class balance, the addition of demographic data, context, and sophisticated deep learning might all be employed to boost performance. also performed a series of pre-processing on the face images such as cropping, and pixel adjustment. The cropping is used to increase processing efficiency by removing extraneous portions of the image to highlight the crucial area. Normalization and contrast enhancement are examples of pixel manipulation that improves analysis and make the image more readable. The expression recognition results indicate that the model achieved an overall accuracy rate of 85.10% on the self-made natural expression dataset. Recognition accuracy was high for happy, neutral, and surprised expressions, while it was lower for disgust and fear expressions due to their variability and similarity in features. Because they have recognizable facial traits that are simple for models to identify—such as a grin for happiness or an open mouth for surprise—they are more accurate at identifying emotions of pleasure, neutrality, and surprise. On the other hand, the model's accuracy is lower for disgust and fear expressions since some of their characteristics are comparable to those of other emotions (for example, the resemblance between the expressions of fear and surprise) and differ from person to person, making it challenging to tell them apart. The confusion matrix highlights that fear expressions were often misidentified as a surprise, primarily due to pupil dilation in both expressions. The study concludes that the developed pre-training CNN model effectively recognizes facial expressions, demonstrating significant accuracy, particularly with certain emotions. Future work may focus on improving recognition rates for less distinct expressions and expanding the dataset for better generalization.

Reduced-Order Dynamic Modeling for a DC Motor Coupled with Flywheel and Torsion Shaft Using the Eigenmode Truncation Method

Ngo Manh Tung, Vu Thi Kim Nhi, Duong Quoc Tuan — Fri, 16 May 2025 00:00:00 +0700

This study assesses the effectiveness of model order reduction for the DC Motor Coupled with Flywheel and Torsion Shaft Mechanism (DCM‑FTSM) by minimizing the number of state variables from nine to three while preserving essential dynamic behavior. Unlike balanced truncation, the Eigenmode Truncation (ET) algorithm prioritizes modal dominance rather than energy‑based approximations, selecting modes that most significantly influencing the system’s response. By transforming the system into modal coordinates and extracting the critical submatrix of eigenvectors, the original ninth‑order model is reduced to third order without compromising stability or performance in either the time or frequency domains. MATLAB simulations demonstrate that the reduced‑order model achieves an H∞ norm error of 11.9456, a mean step response error of 0, and average phase and magnitude errors of 14.8858 deg and 0.0055 dB, respectively. Key time‑domain metrics (rise time, overshoot, peak value, peak time) and frequency‑domain parameters (gain margin, phase margin, phase crossover frequency) align closely with those of the full‑order model within the typical operating range. Moreover, by reducing the state dimension by 67%, ET yields significant computational savings, facilitating faster simulation and real‑time controller computation. The ET method thus enables real‑time control of complex electromechanical systems by balancing accuracy and computational efficiency.

Wavelet Neural Network-Based Controller Design for Magnetic Levitation System

Abdulla Ibrahim Abdulla, Mohammed Qasim, Mohammed Almaged — Sat, 17 May 2025 00:00:00 +0700

The magnetic levitation system (MLS) poses a substantial control challenge owing to its intrinsic instability and pronounced nonlinear dynamics. The implementation of robust control methodologies is imperative to guarantee stable operational performance, particularly in environments characterized by external disturbances and parametric uncertainties. This study investigates the development of a PID-like control strategy for a magnetic levitation system (MLS), employing WNN architecture. The parameters of the proposed controller are optimized by employing Fick's Law Algorithm (FLA). The optimization process utilizes a cost function that comprises a weighted sum of the Integral Time-weighted Square Error (ITSE), Integral Time-weighted Absolute Error (ITAE), maximum overshoot (MO), and minimum undershoot (MU). This multi-objective cost function enables a comprehensive evaluation of the controller's performance across various criteria. A square wave reference signal is employed to conduct the optimization process, presenting a challenging test case for control system performance due to its abrupt transitions. The efficacy of the proposed controller is evaluated through a comparative analysis with a conventional PID controller. Comparative simulations are conducted employing three distinct reference trajectories: step, sinusoidal, and square waves. These diverse trajectories provide a comprehensive evaluation of the controller's performance. To assess the robustness of the proposed controller, simulations are conducted within the MATLAB/Simulink environment, subjecting the MLS model to both external disturbances and parametric uncertainties. The developed controller exhibits superior performance and robustness characteristics in comparison to the conventional PID controller. It effectively attenuates the detrimental impact of both parametric uncertainties and external disturbances, while concurrently maintaining a high degree of performance accuracy in terms of overshoot, steady-state error, and energy consumption.

Design and Simulation of a Multi-Stage On-Site System for Hazardous Medical Waste Treatment in Low-Resource Healthcare Settings

Sat, 17 May 2025 00:00:00 +0700

The management of hazardous medical waste in rural settings with limited resources faces significant constraints due to the lack of specialized infrastructure and inefficiencies in collection systems. This study presents the design and theoretical validation of a compact multi-stage system that integrates five key processes: double shredding, thermal evaporation, gas purification, hydraulic compaction, and UV-C disinfection. The methodology involved finite element analysis (FEA) to verify the shredding subsystem's structural integrity and thermal simulations to assess the efficiency of the evaporation process and the thermal safety of the equipment. Results obtained using SimSolid showed safety factors greater than 2.5 in critical shaft and blade regions, with structural displacements below 0.21 mm. Thermal simulations indicated that the chamber reached operating temperatures between 400 and 600 °C within 20 to 25 minutes, while the external surface remained below 60 °C due to the use of refractory insulation. A consistent thermal response was observed even under extreme simulated conditions (700-1100 °C), reinforcing the design’s stability. The combined heat treatment and compaction stages enabled an estimated waste volume reduction of 70% to 75%. In addition, the microbiological neutralization potential of the system, based on advanced filtration and UV-C disinfection, was evaluated, acknowledging simulation limitations and the need for future experimental validation. The primary contribution of this work lies in demonstrating the feasibility of an autonomous, safe, and efficient system for on-site hazardous medical waste treatment. Future work will focus on building a functional prototype, conducting real-world testing, and analyzing energy consumption and adaptability in rural settings with variable infrastructure.

Enhanced Adaptive Neuro Sliding Mode Controller Parameter Optimization for Coupled Tank System

Nguyen Anh Tuan, Ho Pham Huy Anh — Tue, 20 May 2025 00:00:00 +0700

This paper proposes the EANSMC-MDE method for the coupled tank system (CTS) liquid level control, which consists of the Improved Difference Evolution (MDE) optimizing method optimized parameters for the Adaptive Neuro Sliding Mode Controller (ANSMC). The CTS system represents a nonlinear object with delay and uncertainties, including varying parameters, sensor and output valve noises, etc. The suggested controller contains a direct adaptive controller directly approximated by a Radial Basis Function (RBF) neural network combined with a sliding mode controller used to compensate for the approximation errors of the RBF network and ensure system stability. The stability Lyapunov criterion is used to construct the sliding-mode control system and adaptive rule. The proposed algorithm delivers good control performance right from the start-up phase thanks to the use of pre-optimized parameters, which is an advantage compared to conventional adaptive control algorithms. Simulations are conducted to demonstrate the effectiveness of the proposed optimization method compared to different optimization methods using identical beginning conditions and objective function values to establish equitable comparisons. Furthermore, to demonstrate the superiority of the suggested control method, it is contrasted with the optimal SMC and the traditional ANSMC method. Additionally, the simulations evaluate the response capability of the proposed algorithm under the influence of significantly varying sensor noise levels across different magnitudes, changes in the reference signal, and substantial variations in system parameters. The proposed algorithm has the potential to be applied to other uncertain nonlinear systems. However, it has not yet been validated on systems with fast dynamic responses.

Adaptive Sliding Mode Control for Structural Vibration Using Magnetorheological Damper

Alaa Al-Tamimi, T. Mohammad Ridha — Tue, 20 May 2025 00:00:00 +0700

This study presents the design of a new adaptive sliding mode controller to mitigate building vibrations induced by earthquakes, utilizing a semi-active magnetorheological damper (MRD) positioned on the top floor. This damper operates as a passive damper under low-intensity vibrations and transitions to an active damper during high-intensity vibrations, facilitating optimized performance according to vibration severity. The efficiency of the proposed controller was assessed by comparing it with two other robust controllers established in previous studies. This comparison was performed on a prototype three-story structure, subjected to a severe earthquake called El Centro 1940 with an acceleration of 3.9 m/s². The simulation results demonstrated that the proposed controller effectively reduced vibrations significantly. The proposed controller demonstrated enhancement in control effort 660N compared to 751N for first methodology and 722N for second methodology from the literature. The proposed method offers the advantage of reduced design requirements relative to the first and second methods from literatures. Moreover, the proposed method eliminates the necessity for filter adjustments, hence simplifying its implementation. All controllers utilized for comparison are robust and do not require prior knowledge of disturbance bounds. Moreover, the damper was positioned on the top floor in all the procedures analyzed. The results indicate that the proposed controller significantly reduces control effort relative to alternative controllers in addition The proposed method reduces the displacement of the upper floor by 89% compared to other methods, rendering it an effective choice for vibration control and enhancing building reactions to earthquakes.