Kinetik: Game Technology, Information System, Computer Network, Computing, Electronics, and Control
Not a member yet
424 research outputs found
Sort by
Improvement of AC Bus Voltage Stability with Current Control Inverter
Full text linkThis research focuses on the development and analysis of a current control method for inverters, which demonstrates superior performance compared to the more conventional voltage control method. Current control in inverters offers several significant advantages, including faster dynamic response, constant switching frequency, and the ability to effectively reduce harmonic distortion, which is often a challenge in modern power systems. Additionally, this method is capable of maintaining system stability even when it had complex load variations and fluctuating operating conditions. In this study, we implement a fuzzy logic approach to simulate current control in an inverter integrated with a photovoltaic (PV) renewable energy system. The simulation results indicate that the proposed current control method not only enhances overall energy efficiency, but also extends the operating range of the inverter, allowing the system to operate optimally under various load conditions
Integrating Adaptive Sampling with Ensembles Model for Software Defect Prediction
Full text linkHandling class imbalance is a challenge in software defect prediction. Imbalanced datasets can cause bias in machine learning models, hindering their ability to detect defects. This paper proposes an integration of Adaptive Synthetic Sampling (ADASYN) and ensemble learning methods to improve prediction accuracy. ADASYN enhances the handling of imbalanced data by generating synthetic samples for hard-to-classify instances. At the same time, the ensemble stacking technique leverages the strengths of multiple models to reduce bias and variance. The machine learning models used in this study are K-Nearest Neighbors (KNN), Decision Tree (DT), and Random Forest (RF). The results demonstrate that ADASYN, combined with ensemble stacking, outperforms the traditional SMOTE technique in most cases. For instance, in the Ant-1.7 dataset, ADASYN achieved a stacking accuracy of 90.60% compared to 89.32% with SMOTE. Similarly, in the Camel-1.6 dataset, ADASYN achieved 91.56%, slightly exceeding SMOTE’s 91.32%. However, SMOTE performed better in simpler models like Decision Tree for certain datasets, highlighting the importance of choosing the appropriate resampling method. Across all datasets, ensemble stacking consistently provided the highest accuracy, benefiting from ADASYN's adaptive resampling strategy. These results underscore the importance of combining advanced sampling methods with ensemble learning techniques to address class imbalance effectively. This approach improves prediction accuracy and provides a practical framework for reliable software defect prediction in real-world scenarios. Future work will explore hybrid techniques and broader evaluations across diverse datasets and classifiers
Optimizing Autonomous Navigation: Advances in LiDAR-based Object Recognition with Modified Voxel-RCNN
Full text linkThis study aimed to enhance the object recognition capabilities of autonomous vehicles in constrained and dynamic environments. By integrating Light Detection and Ranging (LiDAR) technology with a modified Voxel-RCNN framework, the system detected and classified six object classes: human, wall, car, cyclist, tree, and cart. This integration improved the safety and reliability of autonomous navigation. The methodology included the preparation of a point cloud dataset, conversion into the KITTI format for compatibility with the Voxel-RCNN pipeline, and comprehensive model training. The framework was evaluated using metrics such as precision, recall, F1-score, and mean average precision (mAP). Modifications to the Voxel-RCNN framework were introduced to improve classification accuracy, addressing challenges encountered in complex navigation scenarios. Experimental results demonstrated the robustness of the proposed modifications. Modification 2 consistently outperformed the baseline, with 3D detection scores for the car class in hard scenarios increasing from 4.39 to 10.31. Modification 3 achieved the lowest training loss of 1.68 after 600 epochs, indicating significant improvements in model optimization. However, variability in the real-world performance of Modification 3 highlighted the need for balancing optimized training with practical applicability. Overall, the study found that the training loss decreased up to 29.1% and achieved substantial improvements in detection accuracy under challenging conditions. These findings underscored the potential of the proposed system to advance the safety and intelligence of autonomous vehicles, providing a solid foundation for future research in autonomous navigation and object recognition
Transfer Learning Approaches for Non-Organic Waste Classification: Experiments Using MobileNet and VGG-16
No full textThis paper develops machine learning (ML) models for classifying non-organic waste automatically. The goal is to support more effective waste management by increasing recycling rates, reducing landfill use, and minimizing environmental impact. The ML models proposed in this paper classify 20 types of non-organic waste collected from the internet, which consists of 2,552 instances. Our experiments reveal several key findings. First, MobileNet, which achieved 86% accuracy, outperforms VGG-16, which reaches only 72% accuracy. Second, both models show good classification performances in classifying glass bottles, toothbrushes, and cigarette butts. Third, both models suffer from misclassification in visually similar categories, especially when it comes to paper-based waste like books, cardboard, foam packaging, and carton packaging. Fourth, MobileNet has difficulty detecting plastic packaging, carton packaging, and books, while VGG-16 exhibits higher misclassification rates for foam packaging, cardboard, and newspapers. These results pose a further critical development of the model to classify non-organic waste with similar textures and shapes. Moreover, it presents the urgency of improving the model to distinguish visually similar waste materials. Considering the number of labels used in this paper compared with existing studies, the findings demonstrate the competitiveness of our models for non-organic waste classification
Multi-objective MPPT Optimisation for PV System Using QHBM Algorithm in Madura Island
No full textThis study presents the application of the Queen Honey Bee Migration (QHBM) algorithm, for Maximum Power Point Tracking (MPPT) in an off-grid photovoltaic (PV) system on Madura Island. Implemented in Python, QHBM optimizes a 3.3 kW PV array (six polycrystalline silicon panels, 550 W each, configured in 2-series and 3-parallel) under tropical conditions (irradiation: 860–970 W/m², temperature: 26–30°C) using data from the East Java BMKG Trunojoyo Meteorological Station. QHBM’s multi-objective optimization balances power conversion efficiency (95.0–99.1%), power quality (THD < 4.5%), and component longevity (current ripple: 3.1–3.2 A), outperforming Perturb and Observe (P&O: 78% efficiency under low irradiation and 34% under partial shading) and Particle Swarm Optimization (PSO: 85% and 88%). Trade-offs are managed by minimizing ripple-induced thermal stress (10–15% lower than P&O) and achieving rapid convergence (0–3 ms vs. 300–500 ms for PSO), ensuring reliability in Madura’s dynamic climate. The system, integrated with a single-phase full-bridge inverter (96% efficiency), delivers a consistent daily energy output of 14,941.87 Wh (SD ±267.45 Wh) and reduces CO2 emissions by 118.49 kgCO2e annually. QHBM was chosen over P&O and PSO for its superior efficiency, faster response, and robustness under partial shading and noisy irradiation (±10% variations), offering a scalable solution for sustainable electrification in Indonesia’s archipelagic regions
Modified U-Net for Leaf Segmentation of Eucalyptus pellita Seedlings in Open Natural Environments
No full textThis study addressed leaf segmentation in open nursery environments for Eucalyptus pellita seedlings, where fluctuating illumination, cluttered backgrounds, and overlapping foliage had hindered reliable monitoring at operational scale. We proposed a Modified U-Net that integrated a ResNet-50 encoder for high-resolution feature extraction, L2 regularization in the decoder to improve generalization, and a composite binary cross-entropy plus Dice loss to balance pixel-level accuracy with shape conformity. We assembled 2,424 RGB images from an operational nursery and evaluated three architectures (Modified U-Net as the primary model, SegNet, and DeepLabv3+) under cloudy, sunny, and scorching illumination. We conducted inference at native resolution and summarized per-image metrics using medians with interquartile ranges, followed by nonparametric significance testing. The Modified U-Net consistently outperformed the baselines across all scenarios, achieving median Dice coefficients of 0.872 (cloudy), 0.841 (sunny), and 0.854 (scorching), with corresponding Intersection over Union values of 0.773, 0.725, and 0.745. A Kruskal-Wallis test on per-image Dice and Intersection over Union yielded no significant differences across lighting conditions (H = 4.012, p = 0.1345), indicating stable performance under natural illumination variability. Qualitative overlays revealed localized errors, including glare-induced false positives in sunny scenes and shadow-related artifacts under scorching light, which did not materially shift global overlap distributions. We concluded that the proposed architecture delivered robust, high-fidelity segmentation in realistic nursery conditions and provided a practical basis for field deployment, with further gains expected from glare- and shadow-aware augmentation and lightweight optimization for near real-time inference on edge devices
The Application of the Adaptive Neuro Fuzzy Inference System (ANFIS) Method in Estimating State of Charge (SOC) and State of Health (SOH) of Lithium-Ion Batteries
No full textThe increasing reliance on lithium-ion batteries (LIBs) for electric vehicles and portable electronics demands accurate monitoring of battery performance, particularly the State of Charge (SOC) and State of Health (SOH). Conventional estimation methods—such as Coulomb counting, Kalman filtering, and equivalent circuit modeling—face challenges under dynamic conditions due to drift and limited adaptability. Recent studies have explored machine learning and neuro-fuzzy approaches to enhance prediction accuracy, yet many lack integration of real-time hybrid learning or struggle with high estimation error in noisy data environments. This research aims to apply the Adaptive Neuro-Fuzzy Inference System (ANFIS) to estimate SOC and SOH using experimental data from a 48V lithium-ion battery. The novelty lies in combining voltage, current, and capacity data within a MATLAB-based ANFIS framework that employs a hybrid learning algorithm integrating backpropagation and Recursive Least Squares Estimation (RLSE). Training data for SOC estimation used charging voltage and current, while SOH estimation incorporated discharging data and capacity. Results show that ANFIS achieved high accuracy with RMSE of 0.1466 and MAE of 0.021 for SOC, and RMSE of 0.012 and MAE of 0.0017 for SOH. The estimated SOH was 33.61%, closely aligned with actual values. These findings confirm ANFIS as a robust and adaptive method for real-time battery diagnostics. Future work will explore multi-input hybrid models, the integration of IoT-based BMS telemetry, and testing across diverse battery chemistries to generalize the model's performance and extend its application in smart energy systems
Ambidextrous Blockchain Governance to Strengthen BankCo’s Digital Transformation through COBIT 2019 Traditional and DevOps
No full textThe adoption of blockchain in banking accelerates digital transformation by enhancing transparency and operational efficiency. However, it also brings with it governance issues pertaining to accountability, compliance, and system integrity within a highly regulated environment. This study addresses these challenges by developing a blockchain governance solution based on ambidextrous approach within COBIT 2019’s Traditional and DevOps Focus Areas. A governance model was built and evaluated through iterative steps. Until saturation was reached, information was gathered through key stakeholder interviews and checked with internal documentation such as yearly reports, risk frameworks, and policy records. The ambidextrous COBIT 2019 framework was used in the analysis for all seven governance components. Governance and Management Objectives (GMOs) were prioritized based on design factors, national regulations (POJK No.11/2022 and SOE Minister Regulation No.PER-2/MBU/03/2023), and insights from prior studies. APO12: Managed Risk was identified as the most prioritized GMO. A capability gap analysis revealed missing leadership roles, overlapping security responsibilities, and underdeveloped risk management practices. Recommendations include formalizing key governance roles and strengthening risk management process for blockchain and DevOps environments. These enhancements are expected to increase the maturity level of APO12 from 3.5 to 4.1, thereby improving BankCo’s risk management, compliance, and innovation capabilities. Ultimately, the findings contribute to continuous digital innovation by aligning risk management practices with strategic performance goals and adaptive control mechanisms rooted in emerging technology principles
Integrating Ensemble Learning and Information Gain for Malware Detection based on Static and Dynamic Features
Full text linkThe rapid advancement of malware poses a significant threat to devices, like personal computers and mobile phones. One of the most serious threats commonly faced is malicious software, including viruses, worms, trojan horses, and ransomware. Conventional antivirus software is becoming ineffective against the ever-evolving nature of malware, which can now take on various forms like polymorphic, metamorphic, and oligomorphic variants. These advanced malware types can not only replicate and distribute themselves, but also create unique fingerprints for each offspring. To address this challenge, a new generation of antivirus software based on machine learning is needed. This intelligent approach can detect malware based on its behavior, rather than relying on outdated fingerprint-based methods. This study explored the integration of machine learning models for malware detection using various ensemble algorithms and feature selection techniques. The study compared three ensemble algorithms: Gradient Boosting, Random Forest, and AdaBoost. It used Information Gain for feature selection, analyzing 21 features. Additionally, the study employed a public dataset called ‘Malware Static and Dynamic Features VxHeaven and VirusTotal Data Set’, which encompasses both static and dynamic malware features. The results demonstrate that the Gradient Boosting algorithm combined with Information Gain feature selection achieved the highest performance, reaching an accuracy and F1-Score of 99.2%
Analysis of Public Opinion on The Governor Candidate Debate Using LDA and IndoBERT
No full textThe gubernatorial candidate debate was broadcast live streaming through various YouTube channels, which attracted public attention. Many discussions and conversations appeared in the comments section of each YouTube channel that broadcasted the debate. Given the numerous public discussions, it is undoubtedly interesting to analyze the contents of the conversations, as well as the expectations and feedback from the public. However, analyzing conversations in the form of text data will be challenging using conventional methods. Therefore, in this study, public opinion will be analyzed using the topic identification and sentiment classification approaches. Topic identification is conducted to obtain accurate information about what the public is discussing, while sentiment classification is used to determine whether each comment contains positive or negative sentiments. This research is novel because it utilizes data collected from various major media YouTube channels and includes a qualitative analysis of the findings. This study uses public comment data taken from the KPU, NarasiTV, and KompasTV YouTube channels; the results obtained included 4,147 data points. Data preprocessing involves identifying topics using the LDA method, evaluating the LDA model, performing sentiment classification using IndoBERT, and visualizing the results of the public opinion analysis. The results revealed five topics with a perplexity value of -7.7909 and a coherence score of 0.5109. In addition, topic 4 is the most dominant compared to other topics, with 1,146 comments classified as positive sentiment and 504 classified as negative sentiment. Topic 4 reflects how religion, culture, and frequently mentioned figures are perceived and discussed by the public, especially in relation to the gubernatorial election (pilgub) or gubernatorial candidate debates