Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

31,291 Article Results

A chip level design of a multi-mode compressive sensing image sensor

10.11591/ijece.v16i5.pp2393-2404
Zahra Sepehri , Sayed Masoud Sayedi , Ehsan Yazdian
This paper presents the full chip-level design of a multi-mode CMOS vision sensor, emphasizing the detailed implementation of its circuit architecture. The proposed chip incorporates our previously developed photodiode sensing array together with the on-chip design of control circuitry. By embedding these building blocks, the chip enables pixel-level compressive sensing and supports dual operation modes, allowing the transmission of image data in both compressed and non-compressed formats. In either mode, the sensor is capable of capturing both scene images and difference images between consecutive video frames, operating at a frame rate of 40 fps. A 64*64 vision chip is implemented using TSMC 0.18um standard CMOS technology. In the normal scene image mode, with compression (N-C) and  without compression (N-nC), the structure consumes 36.99uW and 37.09uW, respectively. Meanwhile, in the difference scene image mode, with compression (D-C) and  without compression (D-nC), it consumes 38.67uW and 38.75uW, respectively.
Volume: 16
Issue: 5
Page: 2393-2404
Publish at: 2026-10-01

A comparative analysis of hybrid FFNN-LSTM and FFNN-RNN architectures for short term electricity load forecasting

10.11591/ijece.v16i5.pp2347-2356
Temitope Akinyede , Josephine Adenike Akinyede , Paul Kehinde Olulope , Emmanuel Taiwo Fasina , Temitope Adewale Olominu
Short-term accurate forecasting of electricity demand is crucial for power-system operation and energy scheduling and the equilibrium between electricity generation and consumption. The predicting of electric power consumption is however difficult because electricity-demand profiles are non-linear and time varying. In this paper, we consider two hybrid deep learning architectures feedforward neural network long short-term memory (FFNN-LSTM) and feedforward neural network recurrent neural network (FFNN-RNN) for multi-horizon electricity-load forecasting. The architectures we proposed combine the ability of FFNN to represent data non-linearly with the sequential modelling capability of LSTM and RNN. The authors assess the models based on historical electricity-load observations from Ado-Ekiti, at three different forecasting horizons of 24 hours, 72 hours, and 168 hours. Root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) assessed predictive performance. The findings reveal that FFNN-LSTM achieves a consistently lower RMSE across the evaluated horizons, resulting in greater effectiveness to confine relatively large forecasting errors. On the other hand, the separate LSTM and RNN models achieve lower MAE and MAPE in various scenarios, suggesting stronger short-term reaction to electricity demand. The results, therefore, show a trade-off between forecast stability and sensitivity to rapid load changes. In general, the performance of hybrid architecture is more stable compared to the standalone recurrent models which are more responsive to short-term changes when we look at most of the forecasting horizons. According to the research paper, the forecasting architectures can now be selected for smart grid applications.
Volume: 16
Issue: 5
Page: 2347-2356
Publish at: 2026-10-01

Modeling and performance analysis of batteries in PV systems

10.11591/ijece.v16i5.pp2294-2303
Abdelfettah Boussaid , Hadjer Bounechba , Seif El-Islam Chelli , Salah Hammoudi
This research paper comprehensively examines photovoltaic batteries, essential elements of sustainable solar energy systems. This research includes a pre- cise mathematical model development to represent behavior and performance of these batteries, in order to better understand the storage and use of electri- cal energy generated by solar panels. In addition, a complete simulation was carried out to evaluate the performance of these models under various operating conditions, thus contributing to improve our understanding on various factors that influence batteries efficiency in real situations. In the field of battery man- agement, their performances were examined, such as charging and discharging mode as well as continuous monitoring of the battery status. The objective is to strengthen the use efficiency of solar energy and improve its sustainability, thus contributing to achieving the global sustainability goals and reducing harmful emissions associated with traditional energies.
Volume: 16
Issue: 5
Page: 2294-2303
Publish at: 2026-10-01

Navigation and steering control of a service robot using Kinectv2 and LiDAR

10.11591/ijece.v16i5.pp2454-2472
Suci Dwijayanti , Silfani Sandra Miranda , Bhakti Yudho Suprapto , Dlulil Amri , Robi Prasetio
Service robots operating in complex indoor environments with dynamic obstacles and narrow spaces require accurate mapping, localization, navigation, and motion control for safe and efficient operation. However, integrated systems that combine all these processes within a unified framework remain limited. This study proposes an integrated navigation and steering control system using RPLIDAR A3 and Kinect v2 sensors within a ROS2-based PLAN framework. The proposed system combines GMapping SLAM and RTAB-Map for hybrid mapping, extended Kalman filter (EKF) for localization enhancement, bidirectional RRT* for global path planning, artificial potential field (APF) for local obstacle avoidance, and a Type-2 fuzzy logic controller (FLC) for steering control. Experimental results show that the mapping system achieved a mean squared error (MSE), F1-score, recall, and precision of 0.26, 0.1714, 0.0312, and 0.0498, respectively. In steering control evaluation, the proposed Type-2 FLC with seven membership functions outperformed the proportional–integral–derivative (PID) controller under the tested conditions, achieving a rise time of 0.2901 s, settling time of 1.6819 s, peak time of 1.6997 s, overshoot of 0.4167%, and steady-state error of 10 (3.6). These results indicate that the proposed system improves navigation stability, steering performance, and environmental perception, demonstrating strong potential for autonomous service robot applications in complex indoor environments.
Volume: 16
Issue: 5
Page: 2454-2472
Publish at: 2026-10-01

Rainfall detection on a tropical island using a communication satellite downlink signal and artificial intelligence models

10.12928/telkomnika.v24i5.26012
Raden Yudha; Badan Meteorologi Klimatologi dan Geofisika Mardyansyah , Budhy; Universitas Indonesia Kurniawan , Santoso; Universitas Indonesia Soekirno , Danang; Badan Meteorologi Klimatologi dan Geofisika Eko Nuryanto , Ferry; Universitas Negeri Jakarta Budhi Susetyo
The development of rainfall estimates based on satellite links in several high latitude regions shows good results. However, its development for tropical regions is still very less promising, leaving the question of which method is worth using. In this work, we investigate artificial intelligence (AI) models by utilizing the Ku-band satellite downlink signal for rainfall detection on a tropical island. Long-term evaluation of all the models is performed using one year of rainfall intensity data. The findings of this study demonstrate that one-dimensional convolution neural network (1DCNN), long short-term memory (LSTM), and random forest (RF) models can effectively identify rainfall occurrences on tropical islands, with the area under the receiver operating characteristic curve (ROC) value of 0.84, 0.84, and 0.83, respectively. These three models consistently provide better accuracy in detecting the region’s diverse rainfall patterns throughout the year. The 1DCNN has better performance compared to the other models, achieving accuracy values of 0.76 and 0.82 during the periods of lowest and highest rainfall, respectively. Considering the presence of numerous potential mislabeled rain samples in the dataset, this result is excellent. In addition, the utilization of higher resolution rainfall intensity data can be a great potential for real-time rainfall monitoring in the tropical region.
Volume: 24
Issue: 5
Page: 1632-1643
Publish at: 2026-10-01

Performance and quality analysis of brain MRI image transmission over free-space optical communication systems under severe atmospheric conditions

10.11591/ijece.v16i5.pp2526-2536
Entidhar Mhawes Zghair , Seham Hashem , Ali Hammadi
Reliable transfer of brain magnetic resonance imaging (MRI) data over atmospheric free-space optical (FSO) links is a key enabler of telemedicine, yet conventional FSO studies judge link quality by communication metrics such as the bit error rate (BER) alone, which cannot guarantee the structural and contrast fidelity that diagnosis demands. This study proposes a quality-aware FSO transmission framework for brain MRI in which link performance is assessed jointly through BER, peak signal-to-noise ratio (PSNR), and the structural similarity index (SSIM). A physical-layer FSO channel is modelled in OptiSystem 20 and co-simulated with MATLAB R2023b, which performs image serialization, reconstruction, and quality analysis. Thirty axial T2-weighted slices (256×256, 8-bit) from the public IXI dataset are transmitted at 1550 nm over clear-air, rain, and fog channels at 500, 1000, and 2000 m. Adopting conservative diagnostic thresholds of PSNR ≥ 30 dB and SSIM ≥ 0.85, the link is diagnostically usable in clear air at all tested distances (PSNR = 42.1 dB, SSIM = 0.98 at 500 m) and in rain up to 2000 m (PSNR ≥ 31.2 dB), whereas fog degrades quality below the thresholds at every distance, reaching PSNR = 22.7 dB and SSIM = 0.68 at 2000 m. A concatenated forward-error-correction (FEC) scheme is then shown to restore diagnostic quality under fog up to 1000 m, extending the usable fog range, while 2000 m remains infeasible and motivates hybrid FSO/RF operation. The framework provides quantitative deployment limits for FSO-based medical image transport.
Volume: 16
Issue: 5
Page: 2526-2536
Publish at: 2026-10-01

Numerical investigation on the effect of blade count on the performance of a small-scale Savonius wind turbine

10.11591/ijece.v16i5.pp2335-2346
Dezetty Monika , Muchlishah Muchlishah , Nuha Nadhiroh , Ahmad Fikri , Muhammad Rafi Anwarrahman , Willy Nurhidayat , Mutiar Mutiar
The growing demand for clean energy has increased interest in small-scale wind turbines, particularly vertical axis wind turbines (VAWTs), which perform well in turbulent wind conditions and low wind speeds. This study numerically analyzes Savonius rotors with two, three, and four blades to evaluate their aerodynamic performance in uniform inflow. Simulations were performed with a steady-state multiple reference frame (MRF) model at an inflow velocity of 10 m/s. Key parameters used included the torque coefficient (𝐶𝑇), power coefficient (𝐶𝑝), and tip speed ratio (𝜆). The results show a clear trade-off between configurations. The four-blade rotor produces the highest torque (0.1297 N·m at 950 rpm) and strong self-starting capability, although its peak efficiency (𝐶𝑝≈0.60) is likely too high due to model simplification. The two-bladed rotor achieves higher efficiency (𝐶𝑝≈0.24) at 𝜆≈0.52 but exhibits significant torque fluctuations. Conversely, the three-bladed rotor produces the lowest efficiency (𝐶𝑝≈0.086) but offers smoother torque, beneficial for stable small-scale power generation. Flow analysis confirms this pattern: two blades cause a wide, asymmetrical wake; three blades reduce wake asymmetry with better stability; and four blades produce a compact wake but with higher drag. Overall, the optimal number of blades depends on design priorities—efficiency, torque, or operational smoothness.
Volume: 16
Issue: 5
Page: 2335-2346
Publish at: 2026-10-01

A hierarchical federated multi-task transfer learning framework for paroxysmal arrhythmia detection, vital sign monitoring, and activity recognition

10.11591/ijece.v16i5.pp2494-2515
Poomari Durga K. , M. S. Abirami
An internet of medical things (IoMT) enabled remote patient monitoring system needs to accurately detect heterogeneous physiological and activity signals while maintaining patient privacy and data flow constraints on sensitive information. Most previous federated or multi-task learning methods focus on one of the following aspects: one-modality per site, one-task per site, or a distributed model optimization, but they do not consider both the signal processing specific to each modality and the knowledge transfer specific to each task. This study introduces a hierarchical federated multi-task transfer learning approach to solve this problem, which is used to detect paroxysmal arrhythmia, monitor heart rate from electrocardiogram (ECG) signals, and recognize human activities. The framework consists of three computational layers: Edge layer for signal preprocessing and learning local features, Federated layer for privacy preserving model aggregation, Cloud-level multi-task transfer learning layer for sharing transferable representations across related monitoring tasks. Wavelet-based denoising, R-peak detection, QR-interval delineation and convolutional neural network (CNN) dan bidirectional long short-term memory (BiLSTM)-attention based classification are performed on the processed ECG signals. The accelerometer and gyroscope data are not temporally synchronized with the ECG dataset, and all processing and activity recognition is done separately from the other data. With the use of the MIT-BIH Arrhythmia database and human activity recognition dataset, experimental results showed that the arrhythmia classification accuracy was 96.8% and the overall activity-recognition accuracy was 91.0%. The findings show that the proposed architecture is able to enable heterogeneous health-monitoring tasks under a shared decentralized learning architecture without moving raw data away from their source. While the framework offers a basis for scalable and privacy-compliant remote monitoring, there are key points that warrant further exploration such as communication overhead and the inclusion of client data diversity, computational complexity, and clinical validation.
Volume: 16
Issue: 5
Page: 2494-2515
Publish at: 2026-10-01

BDLock: A blockchain-enabled two tier privacy-aware federated idam service platform using RBAC

10.11591/ijece.v16i5.pp2537-2548
Muhammad Shakil Pervez , Md. Nasim Adnan , Sarker Tanveer Ahmed Rumee , Moinul Islam Zaber
Centralized identity and access management (IDAM) systems suffer from sin-gle points of failure, lack of authorization transparency, and susceptibility to in-sider threats and privilege abuse. While role-based access control (RBAC) of-fers structured permission management, its enforcement through centralized pol-icy engines introduces auditability gaps unacceptable in modern distributed service delivery environments. This paper presents BDLock, a blockchain-enabled two-tier privacy-aware federated IDAM platform integrating OAuth 2.0, OpenID Connect (OIDC), and Hyperledger Fabric 2.4. The first tier validates JSON Web Tokens (JWT) issued by Keycloak against a Spring Boot resource server, the second tier enforces immutable scope-based RBAC rights on the Hyperledger Fabric ledger, ensuring every access decision is tamper-proof and auditable. Unlike prior approaches, BDLock uniquely bridges OAuth-authenticated off-chain identities to cryptographic on-chain Fabric wallet identities, satisfying all six STRIDE-modelled threats categories across both Web2 and Web3 identity models. Validated with up to 1,800 concurrent users, BDLock achieves a peak throughput of approximately 200 transactions per second using round-robin load balancing. At high concurrency, it outperforms single-peer fallback by up to 25%. Furthermore, it maintains uninterrupted access control during peer failures, eliminating the single point of failure found in all nine compared state-of-the-art systems.
Volume: 16
Issue: 5
Page: 2537-2548
Publish at: 2026-10-01

Design science research in developing a religious chatbot based on Bulugh al-Maram

10.11591/ijece.v16i5.pp2782-2794
Aris Tjahyanto , Irmasari Hafidz , Faizal Johan Atletiko
Chatbots have recently gained significant popularity. For instance, ChatGPT has become a preferred tool for many individuals seeking instant answers without relying on human responses. This immediacy sets chatbots apart from books, which require users to search for information manually. This time-consuming process does not align with millennials' preference for convenience and efficiency. Studying hadith independently using the Bulugh al-Maram book demands considerable time and effort. The limited use of natural language processing technologies in religious chatbots restricts their ability to handle complex inquiries effectively. A chatbot capable of answering hadith-related questions could greatly assist the public in studying hadith texts by providing direct responses without extensive searching. This chatbot was designed for web browsers, utilizing deep learning as its core technology. This research led to the development of a chatbot prototype for learning hadith from Bulugh al-Maram. Built using the design science research (DSR) methodology, the prototype achieves an intent recognition rate (IRR) of 86.82%. However, its capabilities are below the BERT model, demonstrating a strong ability to accurately interpret user questions and statements.
Volume: 16
Issue: 5
Page: 2782-2794
Publish at: 2026-10-01

Calibration-guided score fusion for robust multimodal traffic anomaly detection

10.11591/ijece.v16i5.pp2516-2525
Quang Hiep Do , Thien Tan Nguyen
Multimodal traffic anomaly detection is affected by differences in visual and audio score ranges, temporal fluctuations, and unstable decision thresholds. This paper proposes a calibration-guided score fusion (CGSF) framework that processes video frames and audio spectrograms through separate reconstruction-based models. The resulting anomaly scores are temporally smoothed, normalized using validation data, and combined at the score level. A percentile estimated from normal validation samples is then used as the decision threshold. The framework was evaluated on the MAVD and DADA2000 datasets. On MAVD, CGSF achiev,,,,,,ed a ROC-AUC of 0.553, a PR-AUC of 0.082, and an F1-score of 0.129. It outperformed direct fusion in precision, recall, and F1-score, although the gain in ROC-AUC was small. Analysis on DADA2000 showed smoother temporal score behaviour after calibration and smoothing. The results indicate that CGSF mainly improves score comparability and threshold consistency rather than producing a large increase in detection accuracy. Its modular design also allows the visual and audio branches to be trained and updated independently.
Volume: 16
Issue: 5
Page: 2516-2525
Publish at: 2026-10-01

Constrained finite-control-set predictive control of an H-bridge direct current motor drive

10.12928/telkomnika.v24i5.27909
Bouiri; University of Tindouf Abdesselam , Kendzi; University Center Salhi Ahmed Naama Mohammed , Boughazi; University of Bechar Othmane , Drauoi; University of Tindouf Abdelghani
This paper presents a constrained finite-control-set model predictive control (FCS-MPC) scheme for an H-bridge converter-fed direct current (DC) motor drive, aimed at fast dynamic response and low steady-state armature-current ripple. By selecting the converter switching state directly at each sampling instant, the method removes the separate pulse-width modulation (PWM) stage used in conventional cascaded proportional-integral (PI) control and embeds the armature-current limit directly in the cost function. When the predicted armature current exceeds the predefined limit, a penalty term drives the optimizer to choose the sw itching state that returns the current within safe bounds while preserving speed tracking. The controller is evaluated in simulation under fixed-speed reference tracking, variable-speed reference tracking, and a step load-torque disturbance. The results show fast settling, low steady-state speed error, and bounded current during transients; under the idealized simulation conditions the steady-state armature-current total harmonic distortion (THD) is 0.21 %, indicating low current ripple for the studied operating points. The present study is limited to simulation; experimental validation and a quantitative comparison against conventional PI-PWM control are identified as future work.
Volume: 24
Issue: 5
Page: 1816-1826
Publish at: 2026-10-01

A comparative assessment of machine learning, statistical, and hybrid approaches in amending ECMWF rainfall bias in the Batanghari river basin

10.12928/telkomnika.v24i5.27848
Arif; IPB University Ma'rufi , Tania; IPB University June , I Putu; IPB University Santikayasa , Apip; National Research and Innovation Agency (BRIN) Apip , Siswanto; Agency for Meteorology, Climatology, and Geophysics (BMKG) Siswanto
In tropical basins, rainfall predictions using international numerical weather prediction (NWP) models, like the European Centre for Medium-Range Weather Forecast (ECMWF), exhibit significant persistent biases and insufficient ensemble dispersion. This study evaluates statistical, machine learning (ML)-based, and hybrid bias correction methods to improve daily rainfall estimates for Indonesia’s Batanghari River Basin. Findings indicate that the ECMWF model significantly overestimates wet-day frequency (WDF) (81.5% versus 43.4% observed) and underestimates extreme maximum magnitudes, exhibiting an annual maximum rainfall bias of -15.6 mm/day. Standalone ML-based methods, such as random forest (RF) and extreme gradient boosting (XGBoost), experience significant variance deflation, inadequately addressing these substantial underestimations. In contrast, quantile mapping (QM) significantly reduces the intrinsic drizzle bias and achieves the highest monthly temporal stability among the tested frameworks (Kling-Gupta efficiency (KGE) = 0.341) with low processing requirements. Moreover, hybrid methodologies, including RF-QM and XGBoost-QM, successfully restore the natural ensemble spread and robustly reconstruct the upper tails of extreme rainfall, albeit at the expense of long term volume conservation (resulting in negative monthly KGE). Ultimately, QM is favored as an effective baseline for continuous hydrological modeling, whereas hybrid methodologies are highly suited for evaluating localized flash flood hazards and short-term severe intensities.
Volume: 24
Issue: 5
Page: 1611-1631
Publish at: 2026-10-01

Real-time multimodal fatigue detection using facial vision and alert integration via ESP32 for occupational health applications

10.11591/ijece.v16i5.pp2750-2768
Andrés Enrique Rojas Primo , Alfredo Lazaro Gutierrez , Felix Pucuhuayla-Revatta
Early detection of work fatigue is a major challenge in industrial settings due to the lack of non-invasive, accessible, and low-cost systems capable of operating in real time. In this context, this research proposes a multimodal real-time fatigue detection system using facial vision and artificial intelligence, aimed at risk prevention and promoting occupational health. The system integrates geometric and behavioral parameters, such as eye aspect ratio (EAR), head tilt, and mouth opening, processed on a Raspberry Pi 5 using MediaPipe and a hybrid convolutional neural network (CNN) MobileViT model. Visual and audible alerts are managed by an ESP32 microcontroller using the message queuing telemetry transport (MQTT) protocol, while a graphical interface developed in Tkinter allows real-time monitoring of operator status. Experimental results, evaluated in a simulated work environment using AI-generated synthetic videos, show an accuracy greater than 97% and a latency of less than 250 ms, confirming the system's effectiveness in the early detection of signs of drowsiness and attention deficit. In conclusion, the proposal represents a non-invasive, scalable, and efficient solution that combines computer vision, deep learning, and the Internet of Things (IoT) to strengthen workplace safety and well-being.
Volume: 16
Issue: 5
Page: 2750-2768
Publish at: 2026-10-01

Robust resource allocation in multi-cell UE-specific RIS-assisted D2D relay networks under imperfect CSI

10.11591/ijece.v16i5.pp2575-2594
Kayode Popoola , Ayodeji Ajani , Stuart Nicholson , Muheeb Ahmed , Srilatha Narayangari Pamuri , Ibrahim Bala Alhassan
Device-to-device (D2D) communication enhances spectral efficiency but remains constrained by limited transmission range, underlay interference, and the half-duplex overhead of conventional relays. User equipment-specific reconfigurable intelligent surfaces (UE-RIS) offer a promising alternative by enabling passive beamforming to strengthen D2D links without additional spectrum consumption. However, existing studies typically assume perfect channel state information (CSI) and single-cell operation, limiting their applicability to practical deployments. This paper proposes a robust multi-cell resource allocation (RMRA) framework for UE-RIS-assisted D2D relay networks under imperfect CSI. A hybrid uncertainty model is adopted, combining statistical Gauss-Markov CSI errors for intra-cell links with bounded norm-ball uncertainty for inter-cell links. The joint optimisation of resource reuse, transmit power allocation, and RIS phase configuration is formulated as a stochastic mixed-integer nonlinear program that maximises network spectral efficiency while satisfying outage and quality-of-service constraints. To efficiently solve the problem, a three-stage algorithm is proposed comprising distance-pruned Hungarian assignment, robust power control using Bernstein-type inequality and S-procedure based semidefinite programming, and soft actor-critic (SAC) based passive beamforming. Simulation results show that RMRA achieves a 94% D2D access rate at light load and over 75% at full load, improves sum spectral efficiency by 34.7% and 70.2% over AF relaying and direct D2D, respectively, attains 118.5 bits/s/Hz/W energy efficiency, and maintains 30.2 bits/s/Hz under severe CSI uncertainty.
Volume: 16
Issue: 5
Page: 2575-2594
Publish at: 2026-10-01
1 2 3
...
2086 2087
Show 6 of 2087

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration