International Journal of Electrical and Computer Engineering (IJECE)
Not a member yet
6355 research outputs found
Sort by
Tilted fiber Bragg grating based optical sensor for simultaneous measurement of vital signs: a novel approach
Get PDFVital signal monitoring acts as a highly significant diagnostic method. Continuous monitoring of vital signs like temperature, heart rate, and blood pressure aids in easy and fast disease diagnosis. Existing methods that enable continuous monitoring, such as smart watches, are not accurate enough to be used as a benchmark for diagnosing diseases. In this paper, we propose simultaneous temperature and heart rate measurement using tilted fiber Bragg grating (TFBG) which enables concurrent measurement of measurands due to its intrinsic optical property. Temperature and heart rate were considered for measurement. The proposed TFBG-based optical sensor has a sensitivity of 11.81 pm/°C and 1.73 pm/µε towards variation in temperature and strain, respectively. The sensitivity was determined by a shift in the wavelength of core mode resonance for temperature and a differential change in the wavelength of cladding mode resonance for heart rate. The average response time of the proposed TFBG sensor was found to be 3 seconds. Accuracy of 99.68% and 98.42% were achieved in temperature and heart rate measurements by the proposed TFBG sensor. The acceptability of the proposed TFBG sensor was analyzed using the Bland- Altman plot
IC-CGAN: Imbalanced class-conditional generative adversarial network with weighted loss function
Get PDFThis research proposes an advanced deep learning model that deals with the over-distribution of plant leaf disease classes by using an imbalanced class-conditional generative adversarial network (IC-CGAN) that is coupled with a weighted loss function. IC-CGAN model provides a solution to class imbalance through the synthesis of tomato leaf disease images and adding them to the dataset which as a consequence, improves the accuracy of disease detection. The weighted loss function essentially does a crucial job of solving the problem of imbalance in class during the training stage. Mixing of these models leads to the generation of realistic leaf disease synthetic images and balancing class distribution in the dataset, hence improving of tomato disease detection model’s accuracy. This study is another step toward the development of effective disease detection systems for agricultural purposes by addressing the concern of class imbalance with IC-CGAN through the vector-weighted loss function. The proposed IC-CGAN has a high chance of enhancing the disease detection at its early stage with a much higher level of accuracy (99.95%), precision (99.98%), recall (99.98%) and F1-score (99.98%) in tomato plant leaf disease detection
Efficient power optimized very-large-scale integration architecture of proportionate least mean square adaptive filter
Get PDFThe focus on power optimization in embedded systems is especially important for embedded applications since it has brought in many methods and factors that are necessary for developing systems that are both power- and area-efficient. In contrast to the current delayed wavelet μ-law proportionate least mean square (DWMPLMS) and delayed least mean square (DLMS) algorithms, this work offers the development of adaptive filters based on the least mean square (LMS) method, which improves power and timing performance. In order to improve area and time efficiency, the proportionate least mean square (PLMS) algorithm's architecture has been modified to remove delay, add a proportionate gain block, design for a fixed length, include an approximate multiplier block, and swap out standard blocks for floating-point adder and divider blocks. According to a power and temporal comparison with the DWMPLMS and DLMS algorithms, field-programmable gate array (FPGA) synthesis reduces power usage by 95% for a 32-bit filter length in PLMS when compared to the above methods
Towards a standardized enterprise architecture: enhancing decision-making in oncology multidisciplinary team meetings
Get PDFThis study proposes a novel enterprise architecture (EA) designed to enhance the efficiency and decision-making processes of multidisciplinary team meetings (MDTMs) in oncology by integrating advanced artificial intelligence (AI) technologies. The architecture addresses current inefficiencies in MDTMs, particularly the lack of real-time data integration and limited decision support, by providing a structured framework that improves interoperability and standardizes clinical workflows. Developed using the open group architecture framework (TOGAF) framework and the ArchiMate modelling language, this conceptual architecture lays the groundwork for future empirical research, offering a scalable solution that can be adapted to various healthcare settings. The AI component, centered on generative pretrained transformer (GPT) models, is designed to support oncologists by providing evidence-based treatment recommendations tailored to individual patient cases. Although the study focusses on the theoretical development of this architecture, it opens the door for subsequent empirical testing and validation, with the aim of ultimately improving patient outcomes and streamlined oncology care through enhanced decision support systems
Control of an aquaponic system to improve the yield of gray tilapia and lettuce cultivation
Get PDFWater quality assessment presents challenges, primarily the paucity of available data and ongoing system maintenance. This research develops an automated monitoring and control of water quality parameters in aquaponic systems with internet of thing (IoT) technology. Proper fish feeding management is important, which is why the fish were fed at 12:00, 16:00 and 07:00. The most significant relative error recorded during the validation of the DS18B20, PH-4502C, SEN0244, SEN0237-A, SEN0189 and DFR0300 sensors is 5.0%. The maximum standard deviation between the mentioned sensors was 1.96, and the highest coefficient of variation reached 7.24%. Before the installation of the aquaponic system, the specific growth rate (SGR) of fish was 4.89±0.17% and after implementing the automated aquaponics system, the SGR of fish increased to 6.21±0.24%. The feed conversion ratio values of the fish, both before and after the installation of the control system, were 1.98±0.14% and 1.53±0.09%, respectively. In addition, an improvement in plant growth was observed, evidenced by the difference in the values of height, number of leaves, leaf length, and weight of the plants before and after the installation of the control system, which was 7.74 cm, 5 leaves, 5.6 cm, and 41.6 g respectively
Handwritten text recognition system using Raspberry Pi with OpenCV TensorFlow
Get PDFHandwritten text recognition (HTR) technology has brought about a revolution in the way handwritten data is converted and analyzed. This proposed work focuses on developing a HTR system using deep learning through advanced deep learning architecture and techniques. The aim is to create a model for real-time analysis and detection of handwritten texts. The proposed deep learning architecture that is convolutional neural networks (CNNs), is investigated and implemented with tools like OpenCV and TensorFlow. The model is trained on large handwritten datasets to enhance recognition accuracy. The system’s performance is evaluated based on accuracy, precision, real-time capabilities, and potential for deployment on platforms like Raspberry Pi. The actual outcome is a robust HTR system that can convert handwritten text to digital formats accurately. The developed system has achieved a high accuracy rate of 91.58% in recognizing English alphabets and digits and outperformed other models with 81.77% mAP, 78.85% precision, 79.32% recall, 79.46% F1-Score, and 82.4% receiver operating characteristic (ROC). This research contributes to the advancement of HTR technology by enhancing its precision and utility
Estimation of harmonic impedance and resonance in power systems
Get PDFSince power systems are designed to work at the fundamental frequency, the presence of other frequencies from various sources may induce series and parallel resonances, leading to damage. The behavior of the power system in the presence of harmonics becomes evident with knowledge of harmonic impedance. Measurement offers the most accurate means of estimating harmonic impedance. However, when precise data of the power system parameters are available, highly satisfactory results can be achieved through calculation methods, particularly regarding loads, which are unknown and always change. This paper presents a study on estimating harmonic impedance using the Electromagnetic Transients Program Alternative Transient Program Draw (EMTP-ATPDraw) program, applied to an authentic network of Petrovice line 67, 22/0.4 kV, located in the Czech Republic. Hypothetically, the network was subjected to harmonic injection from a source (3rd, 5th, 7th, 9th, and 11th harmonics), and the harmonic impedance was calculated for three different variants: individual harmonics, all harmonics, and all except the 9th harmonic. The results show that the presence of the 9th harmonic can lead to a parallel resonance. This study is the first to employ EMTP-ATPDraw for programming this network. It gives the possibility to create a network database for different operating conditions, offering an asset for future project planning
Amharic event text classification from social media using hybrid deep learning
Get PDFThis study aims to develop a hybrid deep-learning model for detecting and classifying Amharic text. Various natural language applications, such as information extraction, event extraction, conversation, text summarization, and require an automatic event classification. However, existing studies focused on classification, giving little attention to the preprocessing and feature extraction techniques. To address this problem, this work proposed a hybridized deep learning-based Amharic social media text event classification model. The model consists of word-to-vector (Word2vecv) word embedding techniques to capture the semantic and syntactic representation. Convolutional neural network (CNN) is used to extract short-length text features. Additionally, bidirectional long-short memory (Bi-LSTM) is used to extract features from long Amharic sentences and classify those events based on their classes. The dataset used for training and testing consists of 6,740 labeled Amharic text sentences, collected from social media. The result shows an accuracy of 94.8% in detecting and classifying Amharic text events
Development of a fuzzy logic-based greenhouse system for optimizing bio-fertigation
Get PDFModern agriculture faces growing challenges in meeting food and resource demands, particularly with increasing pressure on water and fertilizer usage. This study proposes a fuzzy logic-based algorithm to optimize bio-fertigation by managing key greenhouse parameters—temperature, humidity, soil pH, and soil moisture. Implemented in MATLAB, the system automates the control of actuators (fan, heater, irrigation, fertilization and fertigation pumps) based on sensor data and fuzzy rules. Results show a 27.58% reduction in water use, 58.82% decrease in fertilizer consumption, and a 47.5% increase in tomato yield. Additionally, statistical error metrics mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), and mean absolute percentage error (MAPE) were reduced to zero, confirming the system’s high precision and effectiveness in promoting sustainable agricultural practices
Low complexity human fall detection using body location and posture geometry
Get PDFThis paper presents the human fall detection using body location (HFBL) and posture geometry. The main contribution of the proposed HFBL system is to reduce the computational complexity of fall detection system while maintaining accuracy, as most fall detection techniques rely on computationally complex algorithms from machine learning or deep learning. This approach examines the human posture by applying the image segmentation and ratio by posture geometry. Then, the distance transform is used to calculate the high brightness points on the human body. These points are the maximum values compared with the edge values. Afterward, one of these points is selected as a center point. A line is formed by this center point aligned horizontally to separate the upper area and lower area, then an intersection line is drawn through this center point vertically that can separate the four quadrants of body location. With the help of posture geometry, the angles are employed for prediction “Fall” or “NotFall” actions at each frame of video sequence. Referring to the dynamic balance, the ratio between the distance vectors from the center point to the right and left legs is calculated to confirm fall and non-fall activities, utilizing the Pythagorean trigonometric identity. For experiments, 2,542 images from the UR fall detection dataset, with dimensions of 640×480×3 were prepared through image segmentation to find the human body shape for analysis using the proposed HFBL system. Results demonstrate that the low computational HFBL approach can provide 91.23% accuracy, the precision value is 99.14%, the recall value is 84.48%, and the F1-score value is 91.22%