EMITTER - International Journal of Engineering Technology
Not a member yet
261 research outputs found
Sort by
Performance Analysis of MIMO-OFDM System Using Predistortion Neural Network with Convolutional Coding Addition to Reduce SDR-Based HPA Nonlinearity
Full text linkIn recent years, the development of communication technology has advanced at an accelerated rate. Communication technologies such as 4G, 5G, Wi-Fi 5 (802.11ac), and Wi-Fi 6 (802.11ax) are extensively used today due to their excellent system quality and extremely high data transfer rates. Some of these technologies incorporate MIMO-OFDM into their protocol. MIMO-OFDM is widely used in modern communication systems due to its benefits, which include high data rates, spectral efficiency, and fading resistance. Despite these benefits, MIMO-OFDM has disadvantages, with the use of a nonlinear HPA being one of them. Nonlinear HPA causes in-band and out-of-band distortions in MIMO-OFDM signals. Utilizing predistortion (PD) is one way of solving this issue. PD is a technique that uses the inverse distortion of the HPA to compensate for the nonlinear characteristics of the HPA. To enhance the quality of MIMO-OFDM systems that the use of HPA has degraded, the convolutional coding (CC) method can be combined with the help of PD. Convolutional coding is a type of channel coding that can be used for error detection and correction. This study will evaluate a combined technique of PD neural networks (PDNN) and CC on the MIMO-OFDM system using Software Defined Radio (SDR) devices. The evaluation of this system led to the use of a technique that combines PDNN and CC to improve SNR and minimise BER on MIMO-OFDM systems that HPA on SDR devices has degraded. In addition, at code rates 1/2, 2/3, and 3/4, using PDNN reduces the SNR value required to achieve BER equal to 0 by 12.037%, 37.8%, and 4.10% when compared to Digital Predistortion (DPD)
An Implementation of blood Glucose and cholesterol monitoring device using non-invasive technique
No full textInvasive testing of glucose and cholesterol levels in the blood is the most prevalent procedure, which is uncomfortable, expensive, and risky since it can spread infections and harm skin cells. Diabetes and cholesterol are two of the most common diseases in the world, and they require constant monitoring to avoid health issues and organ damage. As a result, a non-invasive approach will allow for more regular testing and painless monitoring. The blood glucose and cholesterol levels can be assessed using the principle of reflecting and refractive properties of NIR light source against blood components. The MAX30100 sensor circuit gives SPO2 (Saturated Peripheral Oxygen Level) and BPM (beats per minute, or heart rate) information to the regression model, which is used to forecast blood glucose and cholesterol levels. The polynomial regression model is trained using preset datasets, and the trained model yields regression co-efficient values. For the fresh sample inputs from the sensor, the co-efficient values are used to estimate the new needed parameter value. The projected blood glucose and cholesterol levels are displayed on the LCD Display and delivered through Bluetooth HC-05 module via Serial communication to the mobile application
IRAWNET: A Method for Transcribing Indonesian Classical Music Notes Directly from Multichannel Raw Audio
Full text linkA challenging task when developing real-time Automatic Music Transcription (AMT) methods is directly leveraging inputs from multichannel raw audio without any handcrafted signal transformation and feature extraction steps. The crucial problems are that raw audio only contains an amplitude in each timestamp, and the signals of the left and right channels have different amplitude intensities and onset times. Thus, this study addressed these issues by proposing the IRawNet method with fused feature layers to merge different amplitude from multichannel raw audio. IRawNet aims to transcribe Indonesian classical music notes. It was validated with the Gamelan music dataset. The Synthetic Minority Oversampling Technique (SMOTE) overcame the class imbalance of the Gamelan music dataset. Under various experimental scenarios, the performance effects of oversampled data, hyperparameters tuning, and fused feature layers are analyzed. Furthermore, the performance of the proposed method was compared with Temporal Convolutional Network (TCN), Deep WaveNet, and the monochannel IRawNet. The results proved that proposed method almost achieves superior results in entire metric performances with 0.871 of accuracy, 0.988 of AUC, 0.927 of precision, 0.896 of recall, and 0.896 of F1 score
Modified Deep Pattern Classifier on Indonesian Traditional Dance Spatio-Temporal Data
Full text linkTraditional dances, like those of Indonesia, have complex and unique patterns requiring accurate cultural preservation and documentation classification. However, traditional dance classification methods often rely on manual analysis and subjective judgment, which leads to inconsistencies and limitations. This research explores a modified deep pattern classifier of traditional dance movements in videos, including Gambyong, Remo, and Topeng, using a Convolutional Neural Network (CNN). Evaluation model's performance using a testing spatio-temporal dataset in Indonesian traditional dance videos is performed. The videos are processed through frame-level segmentation, enabling the CNN to capture nuances in posture, footwork, and facial expressions exhibited by dancers. Then, the obtained confusion matrix enables the calculation of performance metrics such as accuracy, precision, sensitivity, and F1-score. The results showcase a high accuracy of 97.5%, indicating the reliable classification of the dataset. Furthermore, future research directions are suggested, including investigating advanced CNN architectures, incorporating temporal information through recurrent neural networks, exploring transfer learning techniques, and integrating user feedback for iterative refinement of the model. The proposed method has the potential to advance dance analysis and find applications in dance education, choreography, and cultural preservation
A Combination of Lexicon-based and Distributional Representations for Classification of Indonesian Vaccine Acceptance Rates
Full text linkWhen the COVID-19 pandemic hit, the use of vaccines was advertised as the end of the pandemic by the entire world. However, the chances of vaccination depended on the sentiments of society and individuals about the vaccine. People's acceptance of vaccines can change depending on conditions and events. Social media platforms such as Twitter can be used as a source of information to find out the conditions and attitudes of the community toward the program. By implementing a machine learning technique on the COVID-19 vaccine dataset, we hope to impact the classification result with text. This study suggests three distinct machine learning models for classifying texts of the COVID-19 vaccination, namely a model based on the first lexicon using the feature extraction method; second, using the word insertion technique to utilize distribution representation; and third, a combination model of distribution representation and feature extraction based on the lexicon. From the evaluation that has been carried out, we found that a combination of lexicon-based and distributional representation methods succeeded in giving the best results for classifying the level of acceptance of the COVID-19 vaccine in Indonesia with an accuracy score of 71.44% and an F1-score of 71.43%
Analytical Analysis of Flexible Microfluidic Based Pressure Sensor Based on Triple-Channel Design
Full text linkIn designing a flexible microfluidic-based pressure sensor, the microchannel plays an important role in maximizing the sensor's performance. Similarly, the material used for the sensor's membrane is crucial in achieving optimal performance. This study presents an analytical analysis and FEA simulation of the membrane and microchannel of the flexible pressure sensor, aimed at optimizing it design and material selection. Different types of materials, including two commonly used polymers, Polyimide (PI) and Polydimethylsiloxane (PDMS) were evaluated. Moreover, different designs of the microchannel, including single-channel, double-channel, and triple-channel, were analyzed. The applied pressure, width of the microchannel, and length of the microchannel were varied to study the normalized resistance of the microchannel and maximize the performance of the pressure sensor. The results showed that the triple-channel design produced the highest normalized resistance. To achieve maximum performance, it is found that using a membrane with a large area facing the applied pressure was optimal in terms of dimensions. In conclusion, optimizing the microchannel and membrane design and material selection is crucial in improving the overall performance of flexible microfluidic-based pressure sensors
Development of a Mobile Application for Plant Disease Detection using Parameter Optimization Method in Convolutional Neural Networks Algorithm
Full text linkPlant diseases are a serious problem in agriculture that affects both the quantity and quality of the harvest. To address this issue, authors developed a mobile software capable of detecting diseases in plants by analyzing their leaves using a smartphone camera. This research used the Convolutional Neural Networks (CNN) method for this purpose. In the initial experiments, authors compared the performance of four deep learning architectures: VGG-19, Xception, ResNet-50, and InceptionV3. Based on the results of the experiments, authors decided to use the CNN Xception as it yielded good performance. However, the CNN algorithm does not attain its maximum potential when using default parameters. Hence, authors goal is to enhance its performance by implementing parameter optimization using the grid search algorithm to determine the optimal combination of learning rate and epoch values. The experimental results demonstrated that the implementation of parameter optimization in CNN significantly improved accuracy in potato plants from 96.3% to 97.9% and in maize plants from 87.6% to 93.4%
Integrated Multi-view 3D Image Capture and Motion Parallax 3D Display System
Full text linkWe propose an integrated 3D image capture and display system using a transversely moving camera, regular 2D display screen and user tracking that can facilitate the multi-view capture of a real scene or object and display the captured perspective views in 3D. The motion parallax 3D technique is used to capture the depth information of the object and display the corresponding views to the user using head tracking. The system is composed of two parts, the first part consists of a horizontally moving camera interfaced with a customized camera control and capture application. The second part consist of a regular LCD screen combined with web camera and user tracking application. The 3D multi-view images captured through the imaging setup are relayed to the display based on the user location and corresponding view is dynamically displayed on the screen based on the viewing angle of the user with respect to the screen. The developed prototype system provides the multi-view capture of 60 views with the step size of 1 cm and greater than 40˚ field-of-view overlap. The display system relays 60 views providing the viewing angle coverage of ±35˚ where the angular difference between two views is 1.2˚
An Improvement of Computer Based Test System Based on TCExam for Usage with A Large Number of Concurrent Users
Full text linkComputer-based test or assessment has been used widely, especially in the current COVID-19 pandemic, where many schools are conducting distance learning as well as distance examination. The need for a computer or software system to support education is inevitable. A range of solutions, from the free/open source software systems to the paid/proprietary ones have been publicly available. Still, an organization with limited resources prefers to find free or low-budget, while yet demanding reliable solutions. We have reported the use of the computer-based test in a new student recruitment test which is held country-wide. We developed the system based on TCExam, a free and open source computer-based test software, and successfully fulfilled the requirements, but with some tweaks. We found that the TCExam has a performance degradation when used by a large number of examinees concurrently, especially during specific phases during the test. This paper reports the result of our investigation to address the problem and suggests some modifications to the base codes as well as a recommendation of the hardware configuration. We evaluated the modified system in a simulated environment. We successfully achieved up to 56% performance gain using the modified system
An Image Processing Framework for Breast Cancer Detection Using Multi-View Mammographic Images
Full text linkBreast cancer is the leading cause of cancer death in women. The early phase of breast cancer is asymptomatic, without any signs or symptoms. The earlier breast cancer can be detected, the greater chance of cure. Early detection using screening mammography is a common step for detecting the presence of breast cancer. Many studies of computer-based using breast cancer detection have been done previously. However, the detection process for craniocaudal (CC) view and mediolateral oblique (MLO) view angles were done separately. This study aims to improve the detection performance for breast cancer diagnosis with CC and MLO view analysis. An image processing framework for multi-view screening was used to improve the diagnostic results rather than single-view. Image enhancement, segmentation, and feature extraction are all part of the framework provided in this study. The stages of image quality improvement are very important because the contrast of mammographic images is relatively low, so it often overlaps between cancer tissue and normal tissue. Texture-based segmentation utilizing the first-order local entropy approach was used to segment the images. The value of the radius and the region of probable cancer were calculated using the findings of feature extraction. The results of this study show the accuracy of breast cancer detection using CC and MLO views were 88.0% and 80.5% respectively. The proposed framework was useful in the diagnosis of breast cancer, that the detection results and features help clinicians in making treatment