International Journal of Electrical and Computer Engineering Systems

Robust and Adaptive Position Control of Pneumatic Artificial Muscles Using a Fuzzy PD+I Controller

2025-07-13T07:52:25+02:00

This study evaluates the effectiveness of a fuzzy PD+I (FPD+I) controller for robust and adaptive position control of pneumatic artificial muscles (PAMs), addressing the challenges arising from system nonlinearity and hysteresis. Experiments were conducted under varying loads, setpoints, and actuated distances to assess the robustness and adaptability of the controller under diverse conditions. As part of the evaluation, the results were compared with those obtained using a conventional PID controller. The FPD+I controller consistently demonstrated superior transient response characteristics, improved trajectory-tracking accuracy, and greater adaptability to dynamic operational changes. Notable improvements include a 21% reduction in settling time and a 22% reduction in rise time under constant loads, as well as a 49% improvement in root mean square error and a 24% reduction in rise time during trajectory-tracking tasks. The controller also exhibited enhanced resilience to continuous load disturbances and maintained stable performance under varying signal amplitudes. These findings suggest that the FPD+I controller is a promising solution for precision control applications in robotics and industrial systems employing PAMs, particularly in dynamic and uncertain environments, where both robustness and adaptability are critical.

Privacy-First Mental Health Solutions: Federated Learning for Depression Detection in Marathi Speech and Text

2025-07-06T14:25:48+02:00

Federated Learning (FL) is a cutting-edge approach that allows machines to learn from data without compromising privacy, making it especially valuable in sensitive areas like mental health. This research focuses on using FL to detect depression through speech and text data from a Marathi-speaking population. Depression, a widespread mental health issue, often leaves subtle clues in the way people speak and write, making speech and text analysis a powerful tool for early identification. However, mental health data is highly personal, and protecting it is crucial. FL addresses this by enabling training across multiple devices without ever sharing the raw data. In this study, we introduce a federated learning framework designed specifically to detect depression in Marathi speakers. The framework combines Natural Language Processing (NLP) for analyzing text and audio processing techniques for studying speech patterns. Using a Marathi dataset that includes both speech and text samples from individuals with and without depression, we train local models on individual devices. These models are then combined into a global model, which is continuously improved through a process called federated averaging. Our findings show that this FL-based approach performs well in detecting depression while keeping the data private and secure. This highlights the potential of FL in mental health applications, especially for languages like Marathi, where gathering and processing data centrally can be difficult. By prioritizing privacy, this work opens the door for future research into using federated learning for other regional languages and mental health challenges. The Federated Learning model outperforms the non-FL model, achieving around 97.9% across accuracy, precision, recall, and F1-score, compared to 97.4% without FL. with speech dataset the model demonstrates high parameter values of above 96.0%., This demonstrates FL’s effectiveness in improving performance on the text and speech based depression detection task.

Offloading of mobile – cloud computing based on Lion Optimization Algorithm (LOA)

2025-07-04T15:56:57+02:00

Many problems and limitations appeared with the spread of mobile cloud computing (MCC) technology, due to the battery life of the mobile, limited resources, storage space and processors. In addition, the biggest challenge is to make decisions about executing tasks between local devices and cloud servers (edge) to save energy and reduce delay time by reducing the energy consumption (or cost) of the system concerned. In this paper, an optimization method based on the Lion Optimization Algorithm (LOA) is proposed for task- offloading of mobile cloud computing. Task- offloading is the transfer of tasks from mobile devices to servers that handle high-level computational operations, such as cloud servers or Mobile-Edge Computing (MEC) servers, to reduce task execution time and energy consumption. The proposed algorithm has a high performance in reducing the cost for offloading tasks. This method can be used to improve the performance of 5G, 6G and later mobile technologies to implement applications that require a large amount of computing, also, multi-objectives could be applied for more cost reduction. The results showed that the proposed algorithm LOA is better than other traditional algorithms such as the genetic algorithm in terms of calculating the costs to find the best solutions.

Attribute-Based Authentication Based on Biometrics and RSA-Hyperchaotic Systems

2025-07-18T11:58:44+02:00

Attribute-based authentication is a security technique that allows access to resources according to characteristics of human biometrics. It confirms the security and increases the efficiency of applications, devices, and resources by controlling them and avoiding cyberattacks. Standard feature extraction approaches can also give erroneous results and limit processing efficiency when processing complex biometric data. This paper proposes a hybrid method combining the Rivest-Shamir-Adleman (RSA) method with 6D hyperchaotic systems to generate dynamic and sophisticated keys for key exchange in a more secure and effective authentication system. User attributes like ID and fingerprint are used for attribute-based authentication by extracting fingerprint features (Minutiae extraction) for biometric matching. The 6D-hyperchaotic systems generate dynamic values over time, influenced by the initial values and input constants. Three of the generated sequences were used on the sender side, and the other three sequences were used on the receiver side after processing them to satisfy the RSA condition. The time consumed for generation numbers is about 0.0717 msec. The results indicated that the system that produces the keys has robust resistance to statistical attacks. The average time for authentication is 0.307867 sec. Thus, the research presents an integrated solution that improves authentication system security and dependability, especially in critical areas that demand sophisticated data and user protection. The user identity verification reached 95.14% accuracy, and the proposed method could be integrated with the extended system by adding a node.

Automation of Surgical Instruments for Robotic Surgery: Automated Suturing Using the EndoStitch Forceps

2025-05-21T14:59:39+02:00

Suturing in minimally invasive robotic surgery poses significant challenges in terms of technique, du- ration, and tool usage. This study presents an innovative semi-autonomous robotic system for suture automation in minimally invasive surgery. The system integrates motorized EndoStitch forceps with a UR3 collaborative robot, combining robotic dexterity with the functionality of a proven surgical tool. The de- sign of the motorized gripper coupling, the development of a modular ROS-based control architecture, and the implementation of a library of parameterized movements optimized for suturing are described. Exper- imental results, obtained in tissue simulations, demonstrate micrometer accuracy in needle positioning. The variability of the positioning error and its relation to the characteristics of the surgical environment are analyzed. This system represents a representative advance towards reducing the surgeon’s cognitive load, improving accuracy and efficiency in robotic suturing, and opening new avenues for safer and more consistent surgical procedures. The semi-autonomous robotic suturing system demonstrated micrometer- level precision in tests. Mean positioning errors ranged from 0.5 × 10−5 m to 2.0 × 10−5 m, with low standard deviations (highest at 0.70 × 10−5 m) indicating high repeatability.

The Conditional Handover Parameter Optimization for 5G Networks

2025-06-05T18:57:43+02:00

Conditional Handover (CHO) is a state-of-the-art handover technique designed for 5G and beyond networks. It decouples the preparation and execution phases of the traditional handover process and aims to reduce wrong cell selection by utilizing a predefined list of target cells. Despite its advantages, the limitations of static parameter configuration compromise CHO performance. This paper proposes a self-optimization mechanism for CHO parameters in 5G networks. Our proposed mechanism is an automated method for estimating and optimizing CHO parameters, dynamically adjusting key parameters to fine-tune the conditions that trigger the execution phase of the handover process. In addition, we introduce a second handover trigger referred to as the cell outage condition. We compared the performance of our proposed mechanism with the baseline CHO, velocity-based, and cell-outage based mechanisms, using Ping-Pong Handovers (PPHO) and Radio Link Failures (RLF). The simulation results demonstrate reduction of up to 7% in RLF, a 0.15% decrease in handover errors, and an improvement of approximately 10% in handover performance at velocities of up to 200 km/h in high-mobility scenarios.

Experimental Study and Modeling of Radio Wave Propagation for IoT in Underground Wine Cellars

2025-11-11T16:30:59+01:00

This paper presents the results of a study of radio wave propagation in underground wine cellars in the context of the optimal use of wireless communication systems for the application of the Internet of Things (IoT) in wine production environments. Electric field strength measurements were carried out in two subterranean line-of-sight (LOS) and non-line-of-sight (NLOS) conditions at 860 MHz, 2400 MHz, and 3600 MHz. The measured results were compared with predictions from seven existing propagation models, including site-general models (Free-space, ITU-R P.1238-13, ITU-R P.1411-12, ETSI TR 138 901 V16.1.0) and site-specific models (ITU-R P.1411-12, tunnel and knife-edge diffraction). Statistical analysis determined that the ITU-R P.1238-13 model, which estimates path loss in corridors, and the tunnel model have the best agreement with measurements in LOS conditions, with average Root Mean Square Error (RMSE) values of 2.8 dB and 3.56 dB, respectively. For the NLOS regions, the knife-edge diffraction model achieved the highest accuracy (average RMSE = 2.77 dB). Furthermore, based on the measurement results, the coefficients of the general path-loss model were adjusted to the data using the least-squares method, yielding RMSEs of 2.02 dB for LOS and 2.65 dB for NLOS conditions. The analysis showed that combining the ITU-R P.1238-13 or tunnel model for LOS conditions with a diffraction-based model for NLOS conditions provides a good basis for modeling radio propagation in subterranean wine cellars. These findings support the design of efficient and robust IoT communication networks in winery environments.