Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
Business intelligence for decision-making in the collection area of a municipality
Get PDFThe large volume of data in systems in the collection area leads to the lack of adequate management of information, as well as dissatisfaction on the part of the user. The purpose of the study is to implement business intelligence (BI) technology to improve the effectiveness of the information and the satisfaction of the attention of the users of a municipality of Lima in the area of collection; therefore, the phases of the Ralph Kimball method with the following phases: project planning; definition of requirements; design of technological architecture; in the dimensional modeling a snowflake scheme was made with 9 dimensions and 1 table made, in the physical design it was implemented in the MySQL management system and in the extract, transform, and load (ETL) development the migration, transformation and cleaning of the data from the online transaction processing (OLTP) database to online analytical processing (OLAP) was executed; obtaining as results that BI managed to increase the level of information efficiency by 53.32%, as well as the level of user satisfaction (LUS) by 1.90%, concluding that BI allows to meet the needs of the user since it maintains a clean, secure and reliable data structure
A study on the impact of layout change to knowledge distilled indoor positioning systems
Get PDFConvolutional neural networks (CNN)-based indoor positioning systems (IPS) have gained significant attention over the past decade due to their ability to provide precise localization accuracy. However, the use of CNNs in these systems comes with a higher computational cost. To tackle this issue, recent studies have introduced knowledge distilled positioning schemes to mitigate the computational burden. Despite the clear possibility of performance degradation due to signal fluctuations, there remains a lack of investigation into the performance of knowledge distilled and CNN based indoor positioning schemes in dynamic indoor environment. To fill this research gap, this paper investigates the practicality of implementing knowledge distilled-based indoor positioning schemes in real-world by analyzing the impact of indoor layout change on these schemes. Results demonstrate that in the case of layout change, the knowledge distilled-based indoor positioning schemes without teaching assistant can still achieve good performance, with an improvement of 11.56% in average positioning error compared to simple CNN model, while taking only 49.05% of the complex CNN model’s execution time. However, the knowledge distilled-based indoor positioning scheme with teaching assistant fails under the same condition as the inclusion of teacher assistant leads to increased error in modeling the received signal strengths (RSS) and locations relationship
Improving Arabic handwritten text recognition through transfer learning with convolutional neural network-based models
Get PDFArabic handwritten text recognition is a complex and challenging research domain. This study proposes an offline Arabic handwritten word recognition system based on transfer learning. The system exploits four pre-trained convolutional neural network (CNN) architectures, namely VGG16, ResNet50, AlexNet, and InceptionV3. In addition, a specialized image recognition model derived from the ImageNet dataset is incorporated. A combination strategy is designed to combine transfer learning with specific fine-tuning techniques, aiming to improve recognition accuracy. The study is conducted on the IFN/ENIT dataset, which includes images of Tunisian City and village names. The results show that the proposed system achieves a recognition accuracy of 94.73%, which is significantly higher than the accuracy rates achieved by previous approaches. These results suggest that the proposed system is a promising approach for Arabic handwritten text recognition
Braille letter recognition in deep convolutional neural network with horizontal and vertical projection
Get PDFBrail is a written mode of communication utilized by individuals with visual impairments to engage in interpersonal exchanges. The braille writing system consists of patterns printed on specialized paper that feature embossed dots. Braille documents enable the visually impaired to acquire knowledge and information exclusively through the application of their sense of contact. Comprehending braille is not a simple undertaking, particularly for the general populace. Because braille is not a required subject in Indonesian education, the majority of the population lacks proficiency in the language. This may therefore result in a minor communication barrier between visually impaired individuals and non-impaired individuals. In order to address this challenge, the present study employs digital image processing via the deep convolutional neural network (DCNN) technique to facilitate comprehension of braille document contents by non-braille speakers. This study employs a deep learning technique that is highly accurate, effective at image processing, and capable of recognizing complex patterns. This study employed the following image processing methods: grayscaling, filtering, contrast enhancement, thresholding, morphological operation, and resizing. Following testing in this investigation, it was determined that the proposed method accurately identifies embossed braille images with a precision of 99.63%
Representative power distribution network: a review of available models
Get PDFIn recent decades, there has been an increasing penetration of new smart grids (SG) and distributed generations (DG) that are connected to the distribution network (DN). Thus, it is critical for utilities to analyze and assess their impact on the power system networks, which often necessitates major decisions about network operation and planning. Consequently, researchers are constantly developing new and improved methods of advanced control and operation to address these challenges. Unfortunately, there are a limited number of realistic DN models that are made publicly available by the utilities for the development, testing, and evaluation of such new methods. This is mainly caused by the utilities' concerns and reluctance to reveal the public's real and “sensitive†network information. Although international standard test systems such as IEEE and CIGRE are publicly available, these test network models are customized based on the US DN and are not representative of the other networks that operate under different network settings. This paper presents a brief literature survey of existing and prominent representative DNs with a special emphasis on identifying the general description, and application, as a comparison for future development of test network in Malaysia
Modified multicarrier sinusoidal pulse-width modulation for three-phase open-load five-level inverter
Get PDFFive-level voltage source inverter (VSI) is a power inverter topology generating a five-level output voltage waveform. This inverter topology can reduce harmonics distortion to be lower compared to a conventional two-level inverter. In practical, delay of gating signals is unavoidable during switching operation of power semiconductor switches. Adding dead time in the gating signals of VSI’s power switches is mandatory to avoid short circuit during switching operation. However, the dead time of the inverter’s switching signals causes low frequency harmonics and distortion of inverter’s output waveforms. In this paper, a different multicarrier sinusoidal pulse-width modulation (SPWM) method with harmonics suppression capability was proposed and applied in the three-phase open-connection load five-level inverter. The proposed modified SPWM was tested using computer simulation of Powersim (PSIM) software. The measured output waveforms of the five-level VSI at different power factor conditions are presented and analyzed. The total harmonics distortion (THD) values of inverter’s output current were suppressed using the proposed SPWM method to be less than 1%. The test results showed that the proposed modified SPWM method was able to reduce the distortion (THD) of alternating current (AC) waveform, and increase the quality of the inverter’s output power
Combining dual attention mechanism and efficient feature aggregation for road and vehicle segmentation from UAV imagery
Get PDFUnmanned aerial vehicles (UAVs) have gained significant popularity in recent years due to their ability to capture high-resolution aerial imagery for various applications, including traffic monitoring, urban planning, and disaster management. Accurate road and vehicle segmentation from UAV imagery plays a crucial role in these applications. In this paper, we propose a novel approach combining dual attention mechanisms and efficient multi-layer feature aggregation to enhance the performance of road and vehicle segmentation from UAV imagery. Our approach integrates a spatial attention mechanism and a channel-wise attention mechanism to enable the model to selectively focus on relevant features for segmentation tasks. In conjunction with these attention mechanisms, we introduce an efficient multi-layer feature aggregation method that synthesizes and integrates multi-scale features at different levels of the network, resulting in a more robust and informative feature representation. Our proposed method is evaluated on the UAVid semantic segmentation dataset, showcasing its exceptional performance in comparison to renowned approaches such as U-Net, DeepLabv3+, and SegNet. The experimental results affirm that our approach surpasses these state-of-the-art methods in terms of segmentation accuracy
An extended sensor fault tolerant control method applied to three-phase induction motor drives
Get PDFThis research presents a fault tolerant (FT) control method for three-phase induction motor drives (IMDs) against sensor failures in the operating process. In this paper, an IMD applied the field oriented (FO) control for the speed and torque control is used to study the operation under sensor fault conditions. A fault detection isolation function is integrated into the FO control loop as an intermediary component to evaluate the quality of the measured signals of the sensors and provide proper signals for speed control of the drive system. A combined method of a comparison algorithm and a third difference operator (TDO) is proposed for the fault diagnosis function to improve the sustainable operation of the drive. The reliability of the proposed method will be verified through the operation mechanism of the FT function corresponding to three sensor fault states and a random noise state in the simulation environment by MATLAB/Simulink software
Toward enhanced skin disease classification using a hybrid RF-DNN system leveraging data balancing and augmentation techniques
Get PDFSignificant health concerns are associated with skin diseases, and accurate and timely diagnosis is essential for effective treatment and patient management. To improve the classification of cutaneous diseases, we propose a novel hybrid system that incorporates the strengths of random forest (RF) and deep neural network (DNN) algorithms. The system employs data augmentation and balancing techniques to enhance model performance and generalizability. The HAM10000 dataset of diverse dermatoscopic images is used for training and evaluation in this study. In the hybrid system proposed, the RF model provides an initial diagnosis based on patient-reported symptoms, while the DNN analyzes images of skin lesions, resulting in more precise and efficient diagnoses. Using hyper-parameter optimization, we fine-tune the system for optimal performance. The evaluation demonstrates the accuracy of the hybrid model, which achieves a classification accuracy of 96.8% overall. According to our findings, the hybrid system demonstrates exceptional efficacy in six of seven skin disease classes. Variations in sensitivity and reliance on data quality and quantity are however cited as limitations. Nevertheless, this hybrid system has the potential to revolutionize skin disease diagnosis and treatment
Gaussian filter and CNN based framework for accurate detection of brain tumor by analyzing MRI images
Get PDFThe diagnosis of cancer can be challenging and time-consuming due to the complex characteristics of tumors and inherent noise in medical imaging. The significance of early detection and localization of tumors must be considered. Radiological imaging techniques can detect and potentially forecast the presence of neoplastic growths at various phases. The expeditiousness of the diagnosis process can be notably enhanced by amalgamating these images with algorithms designed for segmentation and relegation. Early detection of tumors and accurate localization of their position are critical factors. Medical scans, when used with segmentation and relegation procedures, enable the prompt and precise detection of cancerous tumor regions. The identification of malignant tumors enables this achievement. The present article introduces a framework for detecting brain tumors based on a convolutional neural network (CNN). The initial step in processing brain magnetic resonance imaging (MRI) images involves the application of a Gaussian filter to eliminate any noise present. Subsequently, CNN and long short-term memory (LSTM) deep learning methodologies are employed to classify images. CNN has demonstrated improved accuracy in the classification and detection of brain tumors. CNN has achieved an accuracy of 99.25% in cancer image classification. The sensitivity and specificity of CNN are also 98.75% and 99.25%, respectively