TELKOMNIKA (Telecommunication Computing Electronics and Control)
Not a member yet
3120 research outputs found
Sort by
Detecting fake news through deep learning: a current systematic review
Get PDFThis systematic review explores the domain of deep learning-based fake new detection employing advanced search practices on Scopus and Web of Science (WoS) databases with keywords “fake news,” “deep learning,” and “method.” The study encompasses 33 articles categorized into three main themes: i) dataset and benchmarking for fake news detection, ii) multimodal approaches for fake news detection, and iii) deep learning applications and techniques for fake news detection. The analysis reveals the significance of curated datasets and robust benchmarking in improving the efficacy of fake news detection models. Additionally, the review highlights the emergence of multimodal approaches that integrate textual and visual information for improved detection accuracy. The findings clarify the essential role of deep learning applications, emphasizing the development of sophisticated models for automated identification of fake news. This systematic study adds to a thorough grasp of current research trends and offers insightful information for future developments in the field of deep learning-based false news identification
Imposing neural networks and PSO optimization in the quest for optimal ankle-foot orthosis dynamic modelling
Get PDFIndividuals with abnormal walking patterns due to various conditions face significant challenges in daily activities, especially walking. Ankle-foot orthosis (AFO) devices are crucial in providing essential support to their lower limbs. Accurately modeling the dynamic behavior of AFO systems, particularly in predicting ground reaction forces, is a complex yet vital task to ensure their effectiveness. This research develops dynamic models for AFO systems using advanced modeling techniques, employing both parametric and non-parametric approaches. Parametric methods, such as particle swarm optimization (PSO), and non-parametric methods, like multi-layer perceptron (MLP) neural networks, are utilized through system identification methods. According to the findings, the MLP neural network continuously generates objective results and performs exceptionally well in correctly detecting the AFO system, attaining a noticeably lower mean squared prediction error of 0.000011. This research highlights the potential of advanced modeling techniques, particularly MLP neural networks, in enhancing AFO system modeling accuracy. Although parametric techniques like PSO are useful, the MLP approach performs better, offering insightful information about modelling AFO systems and indicating that non-parametric techniques like MLP neural networks have potential to further AFO creation and control
Comparing global system for mobile and G-NetTrack signal strength in drive test study
Get PDFThe quality of wireless signals is determined by the received signal strength indicator (RSSI), which can be measured using various tools, including Android apps and global system for mobile (GSM) modules. To enhance the measurement of signal quality and communication reliability, this study aimed to develop a complementary instrument consisting of a microcontroller, global positioning system (GPS) module, GSM module, and microSD card. The firmware was developed using C++ code and compiled using the Arduino integrated development environment (IDE). In addition, this study utilized the Kolmogorov-Smirnov test (K-S test) normality and the Mann-Whitney U test to investigate any differences in RSSI quality between the GSM module and global positioning system network track (G-NetTrack). The results showed that the GSM module consistently produced higher RSSI values than G-NetTrack in most locations along the route. Furthermore, the K-S test normality suggested that the RSSI values obtained from both tools were not normally distributed, and the Mann-Whitney U test revealed a significant difference between the two samples, with G-NetTrack having lower values than the GSM module (U=14730, p<0.05). This study demonstrated that the GSM module provides stronger signal strength measurements than G-NetTrack during the drive test, and highlighted the importance of using appropriate statistical tests to analyze RSSI data
High-bandwidth millimetre wave multiple-input multiple-output antenna for 38 GHz 5G mobile applications
Get PDFThis study assesses the efficacy of an industrial and innovation antenna by scrutinizing its performance using simulations and an equivalent resistor, inductor, and capacitor (RLC) circuit model. By utilizing computer simulation technology (CST) modeling techniques, the antenna’s small dimensions of 37.75×31.75 mm2 are considered concerning the minimum frequency. The antenna functions at a frequency of 38 GHz, with a bandwidth of 11 GHz. It has a maximum gain of 8.875 dB and demonstrates excellent isolation (-27.627 dB) and efficiency (98.859%), respectively. By designing and simulating a comparable RLC circuit in advanced design system (ADS), we have confirmed the accuracy and reliability of the data acquired via CST. Both CST and ADS simulators yielded similar reflection coefficients. This antenna is a superb option for the 38 GHz frequency range in 5G wireless communication
Optimized tri-band MIMO antenna design for 6G terahertz applications and future connectivity
Get PDFThis paper presents an industrial and innovation rectangular-shaped multiple input multiple output (MIMO) antenna designed for terahertz (THz) frequency applications, specifically targeting 6G communication. The proposed antenna achieves triple-band operation at 3.62 THz, 6.248 THz, and 7.613 THz by incorporating four T-shaped slots. It is designed on a polyimide substrate with a dielectric constant of 3.5 and a tangent loss of 0.0027, with dimensions of 80 μm by 180 μm and a thickness of 11 μm. The patch and the ground plane are constructed from copper, ensuring robust performance. The antenna provides bandwidths of 0.7 THz, 2.2 THz, and 1.1 THz, with isolation levels exceeding -31.3 dB. It achieves a peak gain of 14.3 dB and a high efficiency of 94%, demonstrating its potential for high-performance THz applications. MIMO performance parameters, such as the envelope correlation coefficient (ECC), diversity gain (DG), mean effective gain (MEG), and total active reflection coefficient (TARC), exhibit excellent agreement with theoretical values. The design is further validated through simulations using computer simulation technology (CST) and a circuit model in advanced design system (ADS). The results of these tests mirrored those of the CST simulations, confirming the reliability of future 6G THz communication systems
Challenges in the technological adoption of document management systems
Get PDFIncorporating information and communication technologies (ICT) in public organisations in keeping with digital government policies and access to ICT by citizens motivates public institutions to implement systems to provide better services to citizens. One of the most essential public services is the documentary process, which includes document management systems (DMSs) storing, organising, and managing the documentary flow. The acceptance and use of a DMS enabling digital signature in public institutions depends to a certain extent on a set of factors influencing user behaviour towards it. This paper reports the findings of a quantitative, cross-sectional, and correlational study examining the behavioural intention (BI) to use a DMS, employing three constructs of the unified theory of acceptance and use of technology (UTAUT). The research involved 998 workers from public institutions who participated in a survey, with quantitative data analysed using Spearman correlation. The results show that performance expectancy (PE), effort expectancy (EE), and social influence (SI) positively correlate with the BI to use a DMS and thus must be considered as relevant factors in DMS implementation in public institutions. The results provide relevant information to policymakers and DMS managers to promote the adoption of DMS in the digital transformation process that organisations are experiencing
XGBoost optimization using hybrid Bayesian optimization and nested cross validation for calorie prediction
Get PDFAccurately predicting calorie expenditure is crucial for wearable device applications, enabling personalized fitness and health recommendations. However, traditional models struggle with high data variability and nonlinear relationships in activity data, leading to suboptimal predictions. This study addresses these challenges by integrating extreme gradient boosting (XGBoost) with Bayesian optimization and nested cross validation to enhance predictive accuracy. Unlike previous approaches, our method systematically tunes hyperparameters using Bayesian optimization while employing nested cross validation to prevent overfitting, ensuring robust model evaluation. We utilize a dataset of daily activity records, including steps, distance, and active minutes, extracted from wearable devices. Our experimental findings indicate a substantial enhancement in prediction performance, achieving a mean squared error (MSE) of 4294.27, an Rsquared (R2) score of 0.9917, and a root mean squared error (RMSE) of 65.53. The proposed model outperforms baseline approaches such as random forest and support vector machines in terms of predictive accuracy. These findings underscore the advantage of our approach in predictive modeling. Beyond calorie estimation, the proposed methodology is adaptable to other domains requiring high-precision predictions, such as healthcare analytics and personalized recommendation systems
Convolutional neural network-based real-time drowsy driver detection for accident prevention
Get PDFDrowsy driving significantly threatens road safety, contributing to many accidents globally. This paper presents a convolutional neural network (CNN)-based real-time drowsy driver detection system aimed at preventing such accidents, particularly for deployment in Android applications. We propose a lightweight CNN architecture that effectively identifies drowsiness and microsleep episodes by categorizing driver facial expressions into four distinct categories: close-eye expressions, open-eye expressions, yawns, and no yawns. Our model, which employs facial landmark detection and various pre-processing techniques to enhance accuracy, achieves an impressive 96.6% accuracy. This performance surpasses several popular CNN architectures, including VGG16, VGG19, MobileNetV2, ResNet50, and DenseNet121. Notably, our proposed model is highly efficient, with only 0.4 million parameters and a memory requirement of 1.51 MB, making it ideal for real-time applications. The comparative analysis highlights the superior balance between accuracy and resource efficiency of our model, demonstrating its potential for practical deployment in reducing accidents caused by driver fatigue
Clustering of swamp land types against soil resistivity and grounding resistance
Get PDFIn theory, the resistivity value of the soil is one of the factors that must be taken into account when planning a grounding installation. The resistivity value of swamp soil is 30 Ωm, as per the general requirements for electrical installation of 2011 (PUIL 2011). This value is identical to the resistivity of the soil type in The Institute of Electrical and Electronics Engineers (IEEE standard 80 in 2000), where the wet soil type has a resistivity value of 100 Ωm. It is difficult for electrical engineers to design construction on swamp land because the standard's representation of the features of swamp land does not accurately reflect the types of swamps or wetlands that exist in reality. The focus of this investigation is the resistivity value of swamp soil types. The results of this investigation will make a scientific contribution to the clustering of land at each soil resistivity value in freshwater, brackish water, saltwater, and acidic water swamp land. These soils have pH values that range from 3.5 to above 6. The research on swamp land clustering has revealed that each swamp has a distinctive resistivity value for the different types of swamp soil
Identification of working memory status in children from EEG signal features using discrete wavelet transform
Get PDFThe conventional method for assessing the working memory performance of children is time-consuming and potentially inaccurate, especially when dealing with many samples. Therefore, an automated system that can produce swift and accurate results is required. Electroencephalograms (EEG) can be used to analyse the working memory status of children by extracting specific features from the EEG signal, which can be incorporated into an automatic system to reduce manpower and processing time for analysis. This project used EEG recording to identify children’s working memory status while they were performing working memory tasks. EEG signals were acquired from both children and adults using an automated computer-based working memory assessment tool, processed, and analyzed. The discrete wavelet transform (DWT) was then employed to identify five distinct working memory statuses: distracted, confused, daydreaming, losing focus, and active. DWT was also used to extract features that demonstrate these various statuses. The results showed that DWT could accurately identify the working memory status of both children and adults from their EEGs. This work has thus provided a more efficient method for extracting features from EEG signals to identify working memory statuses in both children and adults