Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Not a member yet
776 research outputs found
Sort by
A New Fault Tolerant Scheme for Switch Failures in LLC Resonant Converter
Full text linkThe LLC (Inductance Inductance capacitance) resonant converter offers advantages such as high power density, high efficiency, and compact size, making it widely used in photovoltaic power generation systems. Its operational reliability is crucial for the continuous performance of these systems. However, complex operating conditions and variable climates can adversely affect power equipment. Switch fault diagnosis and remedial measures are essential aspects of designing isolated full-bridge DC-DC converters, significantly enhancing overall system reliability. When a switching component fails, the resonant converter cannot operate near its resonant point, leading to substantial reductions in efficiency and output power. To improve system fault tolerance and reduce maintenance costs, this paper proposes an improved LLC topology and a rapid switch short-circuit fault diagnosis method for phase-shift full-bridge converters. By real-time monitoring of the average voltage of the resonant capacitor, the method quickly identifies switch short-circuit faults within a single switching cycle, enabling topological control of faulty and redundant components. The modified topology ensures stable output voltage and power while allowing the converter to operate near the resonant frequency. The paper discusses the working principle, design considerations, and implementation of this approach. Simulation results verify the effectiveness of the proposed method
CyberShieldDL: A Hybrid Deep Learning Architecture for Robust Intrusion Detection and Cyber Threat Classification
Full text linkIn modern network environments, securing systems from newly emerging attacks is essential, and a constructive approach is the use of an IDS (Intrusion Detection System). When faced with attacks that are not in the list of predefined patterns, traditional IDS methods such as signature-based detection or standalone machine learning models may not function properly to detect such attacks because they are not adaptable and not designed to deal with this type of attack. The current IDS systems that employ deep-learning architectures have enhanced detection capabilities; however, most prior art systems are limited by partial feature learning, which only learns features of either spatial or temporal traffic structures. Meanwhile, the lack of contextaware mechanisms, such as attention, limits their ability to attend more to the most informative network components, leading to suboptimal detection performance and generalization. To counter this issue, in this work, we introduce CyberShieldDL, which is the first deep learning-based IDS framework with a novel hybrid architecture: IntruNet-Hybrid, combining Convolutional Neural Networks (CNN) for spatial pattern extraction, Bidirectional Long Short-Term Memory (Bi-LSTM) networks for sequential feature extraction, and an attention mechanism to learn the salient features for intrusion detection dynamically. To create the framework, an optimized preprocessing and feature selection pipeline is presented to effectively and costeffectively prepare the model input. Extensive experiments on the CICIDS2017 dataset demonstrate that CyberShieldDL consistently outperforms the state-of-the-art, achieving an overall accuracy of 98.35% and high precision, recall, and F1-score in various attack scenarios. Cross-dataset validations on NSL-KDD and UNSW-NB15 also verify the system's generalization. The design provides a scalable and flexible solution for realworld network security, offering the flexibility and adaptability necessary to enhance classification accuracy and robustness against evolving attack patterns. Its modular construction enables us to extend it for real-time deployment and future adversarial robustness easily
Improving Channel Gain of 6G Communications Systems Supported by Intelligent Reflective Surface
Full text linkThe 6G wireless communication networks may use intelligent reflecting surfaces (IRS). It can enhance energy efficiency (EE). The IRS can enhance wireless communication by selectively reflecting incident signals in favorable directions. A potential method to improve the efficacy of wireless channels is to use a software-controlled metasurface that reflects signals when the direct transmission line from the source to the destination is insufficient. The IRS may redesign the environment to facilitate radio signal transmission. The decrease in channel gain in 6G communications networks using multiple reflective elements of the IRS is one of the challenges. This study seeks to propose a solution to enhance the channel gain and performance of the IRS in 6G communication systems. The research aimed to improve channel gain in assisted-IRS 6G communication systems by artificial intelligence algorithm (DS-PSO: dynamic and static particle swarm optimization). This study's technique enhances the effectiveness of aided-IRS communication methods. The simulation results of the optimized IRS model proposed in this paper show a significant improvement in channel gain compared to the results of previous studies
Systematic Literature Review: Security Challenges of IoT-based Smart Home Systems
Full text linkInternet of things has a wide range of applications such as healthcare, agriculture, transportation, and industrial manufacturing. Smart homes automation occupies a large segment of applications. Due to the proliferation of IoT-based smart homes systems, the attack vector on these devices expanded and became a target for attackers. Although these devices are improving constantly, security is still a challenge for them. The lack of security standardizations and hardware limitations resulted in a slow or lack of security practices in these devices. In this study, we conduct a systematic literature review to identify the exposed threats in the last five years in these devices and introduce novel countermeasures to mitigate the security issues. IEEE, ACM, Scopus, Science Direct, Springer, and MDPI databases were selected for the systematic review. The result of the systematic review were 731 articles collected. Based on reading the abstract, 605 articles were excluded, and 41 articles were excluded based on reading the full text. The result 70 articles were filtered using quality assessment criteria which resulted in 35 articles related to our search domain and answering the research questions. Additionally, a survey is conducted to elicit experts in the field knowledge to enhance our findings. We were able to identify 22 unique threats that endanger smart homes systems and the proposed countermeasures classified into 5 classes
Detection and Estimation of Schizophrenia Severity from Acoustic Features with Inclusion of K-means as Voice Activity Detection Function
Full text linkSchizophrenia symptom severity estimation provides quantitative information that is useful at both the detection and treatment stages of the mental disorder, as the information helps in decision-making and improves the management of the illness. Very limited studies have been recorded for estimating the symptom severity as a regression task with machine learning, especially from speech recordings, which is the aim of this study coupled with detection. Acoustic features, which comprise frequency-domain and time-domain features, were extracted from 60 schizophrenia subjects and 59 healthy controls enrolled in this research. The acoustic features were used to train GridSearchCV-optimized XGBoost as a classifier. Three Multi-Layer Perceptron (MLP) networks, hyper-parameter-tuned by Bayesian Optimizer, were trained to predict the sub-type symptom severity from acoustic extracted features from the schizophrenia groups. The XGBoost classification model that discriminates between schizophrenia and healthy groups achieved a classification accuracy of 98.6%. The three MLP regression models yielded Mean Absolute Errors of 1.975, 2.856, and 1.555, as well as correlation coefficients of 0.888, 0.806, and 0.786 for predicting positive, negative, and cognitive symptom scores, respectively. Solution architecture for the deployment of the models for practical use was suggeste
Enhanced Field-Oriented Control for Synchronous Reluctance Motor Using Fuzzy Logic
Full text linkThis paper presents a fuzzy logic-based Field Oriented Control (FOC) strategy for synchronous reluctance motors (SynRMs). The proposed algorithm addresses the inherent nonlinearities and parameter sensitivities of SynRMs by integrating fuzzy logic control (FLC) into the FOC framework, enhancing system robustness and adaptability. The SynRM model is derived in the rotor reference frame, with two control loops implemented: one for speed control and the other for flux control. Two FLCs are utilized in the speed control loop, while one FLC is adopted in the flux control loop. Fuzzy sets, membership functions, and rule bases enable dynamic parameter tuning. The entire system is simulated in MATLAB/Simulink. The system's dynamic performance is rigorously evaluated in two scenarios: with decoupling control components between the speed and flux control loops, and without these components under various loading conditions. Comprehensive simulations demonstrate that the proposed control algorithm, without decoupling control components, exhibits superior dynamic performance in terms of rise time, overshoot, and settling time. Furthermore, eliminating the decoupling components reduces the system's dependency on machine parameters while having a minor effect on undershoot
Deep Learning-Driven Intrusion Detection System for Distributed Denial of Service Mitigation
Full text linkDDoS attacks continue to pose a serious risk to digital infrastructures, as they can render online services inaccessible without altering system files or gaining direct control over the target. Traditional security mechanisms often fall short in identifying these attacks promptly due to their massive scale and the subtlety with which they blend into regular traffic. With the advancement of artificial intelligence, especially in the realm of deep learning, new solutions are emerging to enhance the detection and classification of such threats. In this work, we focus on strengthening Intrusion Detection Systems (IDS) by leveraging deep learning methods to improve accuracy and responsiveness in detecting DDoS attacks. Using the comprehensive CIC-DDoS-2019 dataset, we experimented with several deep learning architectures including Feedforward Neural Networks (MLP), Convolutional Neural Networks (CNN), and Recurrent models incorporating Long Short-Term Memory (LSTM). These models were evaluated for their ability to analyze complex traffic behaviors and identify malicious activity within diverse network environments. his study contributes to the ongoing research on intelligent cybersecurity solutions by proposing deep learning-based IDS frameworks that not only detect threats with higher accuracy but also adapt to dynamic attack patterns. Our findings suggest that such models can serve as a critical component in modern security infrastructures, offering scalable and resilient defense mechanisms against increasingly sophisticated cyberattacks like DDoS. Our empirical results demonstrate that the MLP model yielded the most reliable performance, achieving an outstanding classification precision of 99.62% across various traffic categories. This highlights its effectiveness in isolating harmful flows from legitimate ones, thereby reducing the risk of false alarms and improving detection reliability
Blind Image Quality Metric for Color Images Based on Human Vision System and Deep CNN
Full text linkThis article introduces a novel blind image quality metric (BIQM) for color images which is designed taking into account human visual system characteristics. The BIQM has a four-stage framework: RGB to YUV transformation, denoising with convolutional neural network , quality evaluation, and weighting to make it compatible with the human visual system. Experimental results, including Spearman's rank-order correlation coefficient, confirm BIQM's effectiveness, particularly in scenarios involving white noise and its compatibility with the human visual system. Furthermore, a survey involving 100 participants ranks images based on three distinct qualities, validating the method's alignment with the human visual system. The comparative analysis reveals that the proposed BIQM can compete with commonly used non-referenced quality measures and is more accurate than some of them. The MATLAB codes for the development of the BIQM are made available through the provided link: are available in the link: https://bit.ly/49MrbF
Optimized Dual-Band Reconfigurable Power Amplifier for 5G
Full text linkThis article presents the design of a dual-band power amplifier capable of operating in three different classes: A, AB, and B. Simulation results reveal that a single power amplifier can efficiently operate at two specific frequencies of the 5G core band, namely 3.5 GHz and 3.8 GHz. The amplifier demonstrates exceptional stability and matching across both frequency bands. It achieves a maximum gain of 17.7 dB, a maximum output power of 41.2 dBm, and a maximum power-added efficiency (PAE) of 70%. These performance characteristics are achieved through an innovative design that allows for frequency band reconfiguration via a PIN diode switch, as well as the selection of the operating mode among classes A, AB, and B. This flexibility makes the amplifier ideal for applications in 5G communication systems, offering an optimal balance between linearity, energy efficiency, and overall performanc
Robust Stereo Matching for Driver Assistance Systems Under Adverse Driving Conditions
Full text linkDeep stereo networks perform effectively when both training and testing data come from the same domain. However, their accuracy tends to drop significantly in efficiency-focused target scenarios due to domain shifts between training and testing datasets. These shifts often arise from differences in factors such as color, lighting, contrast, and texture. Additionally, the architecture of deep networks generally results in processing times that are unsuitable for real-time applications. To address these issues, this paper proposes a lightweight and robust stereo matching approach tailored for diverse driving environments. It leverages attention mechanisms for feature extraction and uses evolutionary algorithms for optimizing parameters. The method outperforms existing deep learning and traditional stereo matching techniques in terms of both processing speed and the percentage of bad pixels, as demonstrated on three challenging outdoor datasets: KITTI, HCI, and Driving Stereo. These results indicate that the proposed solution is highly effective for real-world applications where both precision and flexibility are essential