International Journal of Electrical and Computer Engineering Systems

Measurement of State of Charge of Lithium-Nickel Manganese Cobalt Battery using Artificial Neural Network and NARX Algorithm

2024-01-10T11:52:22+01:00

The battery's SoC is a crucial variable since it reflects its performance. An accurate estimation of SoC protects the battery, prevents overcharging or discharge, and extends its life time. Since most of the traditional methods use complex equations, ANN has been implemented to reduce the complications and provide better accuracy. In this research, Li-NMC with capacity rating of 2000mAh is used for the estimation of SoC. In this paper, Feedforward Neural Network (FNN) algorithm and Nonlinear Auto-Regressive network with exogenous inputs (NARX) have been used for designing a neural network model. Here, the performance matrixes of both neural network models have been compared and analyzed with the same dataset.

Boosting Reliability: A Comparative Study of Silicon Carbide (Sic) and Silicon (Si) in Boost Converter Design Using MIL-HDBK-217 Standards

2024-01-19T10:39:15+01:00

Reliability is very important in the world of electronic device design and production, particularly in applications where continuous and flawless performance is a necessity. This directs our attention to the boost converter, which forms the foundation of power electronics, renewable energy systems, and electric vehicles. However, as technology progresses, the choice of materials for these converters is a big challenge. For that, in this paper, the impact of using Silicon Carbide (SiC) devices, with their promising material properties, on the reliability of boost converters is presented. Because the results showed that more than 80% of boost converter failures are caused by semiconductors, the use of SiC materials is assessed by determining its reliability using MIL-HDBK-217 standard. In addition, a comparative study with the use of traditional Silicon (Si) is conducted. The results showed that the failure rate of boost converters based on SiC devices reduced from 8.335 failure/10-6h to 6.243 failure/10-6h. This notable shift in failure rates establishes SiC as a pivotal material in the evolution of boost converter technology, offering a compelling solution to address the persistent challenges associated with semiconductor-related failures.

Design of Regenerative Braking System and Energy Storage with Supercapacitors as Energy Buffers

2024-01-26T13:44:06+01:00

Vehicles are part of urban area transport and are subjected to variable loads as they traverse the city with varying slopes and stop-and-go traffic. Electric Vehicles (EVs) can be a good option because of their high efficiency under stop-and-go conditions and ability to gain energy from braking. However, limited battery energy makes EVs less efficient and degrades their lifetime. In contrast to a Li-Ion battery, supercapacitors work well under high power charge and discharge cycles. However, their high cost and low energy density prevent them from being viable replacements for batteries. Due to the slow charging and discharging process of batteries, they have a low power density, but a high energy density compared to the supercapacitor. In this paper, we discussed our system design consisting of both a battery and a supercapacitor. The main aim is to design and develop a scheduling algorithm to optimize energy flow between the battery, supercapacitor, and motor. We further described an analogue-based control methodology and algorithm for the supercapacitor, augmented battery-powered motoring process. This is in addition to a charge controller designed to optimize the supercapacitor bank's current-based charge-discharge profile. The system design and tests are developed on PSPICE and a hardware platform.

Optimizing Enhanced Extended Topological Active Nets Model Using Parallel Processing

2024-02-16T13:27:07+01:00

In numerous clinical applications that support the diagnosis and treatment planning of a broad variety of disorders, medical image segmentation is essential. Medical picture segmentation using the Enhanced Extended Topological Active Net (EETAN) model has proven to be successful in correctly identifying structures. This study suggests a novel way to combine the best clustering techniques and parallel processing approaches to maximize the segmentation performance of the EETAN model. The Probabilistic Depth Search Optimization (PDSO) Algorithm, which makes the parallel searching technique to find the ideal contour set, is responsible for this. This work implements parallel processing and ideal clustering to improve the EETAN model's performance in medical image segmentation. Performance metrics like accuracy, precision, recall, dice similarity, and computational time are used for a comparison study. The results demonstrate the notable enhancements attained by employing parallel processing and effective clustering.

DHM-OCR: A Deep Hybrid Model for Online Course Recommendation and Sustainable Development of Education

2024-01-05T09:50:36+01:00

In recent years, there has been an increase in online education resources to help learners improve their skills. However, it is difficult to select the right course from available online education resources due to the demands and needs of learners with different knowledge domains. To solve this problem, an online course recommendation model has the important factor of enhancing learner's knowledge. Many existing recommendation systems (RS) use collaborative filtering (CF) to recommend courses to learners. The major problems with the Collaborative Filtering Recommendation System (CFRS) are the sparse preferences and the scalability of the data. According to the similarity of items, many recommendation models are proposed and developed, but none of these provide suggestions to users without their associations or preferences. We propose a deep hybrid model-online course recommendation (DHM-OCR) that uses high-level learner behavior and course objective features. We demonstrate the improvements and efficiency of the model for suggesting online e-learning courses. According to the analysis and evaluation results, it seems that our DHM-OCR outperforms the parallel research recommendation system. Experimental findings from online course data reveal that the suggested model and approach significantly improve classification accuracy and training efficiency, particularly limited available data.

Comparative Analysis of Banana Detection Models: Deep Learning and Darknet Algorithm

2024-02-08T10:27:28+01:00

This study aims to compare and evaluate the performance of banana detection models utilizing deep learning techniques and the Darknet algorithm. The objective is to identify the most effective approach for accurately detecting bananas in various real- world scenarios. The analysis involves training and testing multiple models using different datasets and evaluating their performance based on precision, recall, and overall accuracy. The results provide valuable insights into the strengths and weaknesses of each approach, enabling researchers and practitioners to make informed decisions when implementing banana detection systems. To detect banana objects, several convolutional neural network (CNN) models were used, including MobileNetV2, YOLOv3-Nano, YOLO Fastest 1.1, YOLOv3-tiny-PRN, YOLOv4-tiny, YOLOv7, and DenseNet121-YOLOv3. The training process utilizes the Darknet algorithm to facilitate the identification of banana types/classes captured by a camera, resulting in an MP4 film file. In this research, various experiments were carried out using different CNN models. However, these six models achieve optimal accuracy above 80%. Among them, the YOLOv7 model shows the highest average accuracy (MAP) at 100%, followed by the small model YOLOv4 at 92%. Meanwhile, for performance measurements, the accuracy of the YOLOv4-tiny model was 87%, followed by the YOLOv7 model at 84%. In the banana fruit experiment, several models showed very good performance, such as recognition of the Ambon, Kepok, and Emas banana classes up to 100% using the YOLOv7 and YOLOv4-tiny models. The YOLOv7 model itself can recognize other banana classes up to 100% in the Barangan, Rjbulu, Uli, and Tanduk classes. Furthermore, theYOLOv4-tiny model can recognize other banana classes, up to 90% of the Barangan, Rjbulu, Rjsereh, and Uli banana types. Thus, this experiment provides very good average accuracy results on 2 CNN models, namely YOLOv7 and YOLOv4-tiny. Future research will involve grouping pictures of bananas, which produces different image shapes, so it requires a different way to recognize them. It is hoped that this research can become a basis for further research in this field.

Gray Level Co-occurrence Matrix based Fully Convolutional Neural Network Model for Pneumonia Detection

2024-02-05T15:24:51+01:00

This study presents a new method to improve the detection ability of a convolutional neural network (CNN) in pneumonia detection using chest X-ray images. Using Gray-Level Co-occurrence Matrix (GLCM) analysis, additional channels are added to the original image data provided by Guangzhou Children's Hospital in Guangzhou, China. The main goal is to design a lightweight, fully convolution network and increase its available information using GLCM. Performance analysis is performed on the new CNN model and GLCM-enhanced CNN model, and results are compared with Transfer Learning approaches. Various evaluation metrics, including accuracy, precision, recall, F1 score, and AUC-ROC, are used to evaluate the improved analysis performance of CNN. The results showed a significant increase in the ability of the model to detect pneumonia, with an accuracy of 99.57%. In addition, the study evaluates the descriptive properties of the CNN model by analyzing its decision process using Grad-CAM.

Data-driven Gait based Severity Classification for Parkinson's Disease using Duo Spatiotemporal Convoluted Kernel Boosted ResNet model

2024-02-25T19:18:01+01:00

Parkinson’s disease (PD) is one of the reformed brain syndromes that results in unintended stiffness and difficulty with balance and dexterity. To detect PD in medical scenery, physicians commonly use experimental indicators like motorized and non-motor symptoms and the severity rating depends on the unified PD Rating Scale (UPDRS). However, these medical assessments highly rely on expertized clinicians and lead to inter-variability discrepancies. Nowadays, gait sensor data assists doctors in diagnosing PD and estimates the severity level of gait abnormalities in patients. However, the gait sensor data increases the dimensionality issues and is subjected to high non-linear complexity. Hence, this study suggests an innovative deep learning (DL) technique for accurate PD analysis using gait patterns. Initially, the gait sensor data is preprocessed by performing data cleaning, and decimal scaling normalization (DS- Norm) to enhance the data quality. The Hoehn and Yahr (H&Y) scale is a commonly used rating scale for measuring the progression of Parkinson's disease symptoms. It's typically used to assess motor symptoms like tremors, rigidity, and bradykinesia. The scale ranges from 0 to 5, with higher numbers indicating more severe symptoms and disability. The preprocessed data are then fed into the proposed Duo spatiotemporal convoluted kernel boosted ResNet (DSCK-RNet) model for classifying the PD severity rating by learning the gait spatiotemporal features. The developed method is processed and scrutinized via the Python platform and a publicly available Physio- Net dataset is utilized for the simulation process. Various assessment measures like accuracy, precision, sensitivity, specificity, PPV, FPR, and MCC are examined and compared with traditional studies. In the experimental section, the developed DSCK-RNet model achieved an accuracy of 100%, 99.6%, 99%, and 99.64% for different classes like healthy, severity-2, severity-2.5, and severity-3 respectively. Compared to the conventional techniques, our suggested approach performs better. The experimental findings demonstrate the clinical significance of the suggested approach for the impartial evaluation of gait motor impairment in PD patients.