Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Machine learning framework and tools in precision farming
Get PDFFarming using machine learning (ML) techniques has a role to play in the current globalization scenario due to the advantages it offers for costeffective harvesting of the crop. The areas such as crop disease detection, soil nutrient detection, fertilizer analysis and optimization, weather and irrigation schedule prediction, are investigated utilizing a range of deep learning and ML techniques, such as K-nearest neighbors (KNNs), convolutional neural networks (CNNs), and support vector machines (SVMs). The article concentrates on preparing the recommendation system for the farmer to take a quick and timely decision for crop disease, use of optimal fertilizer for crop growth, and water requirement prediction to overcome water wastage. A massive amount of data, including image data from publicly accessible sources, such as PlantVillage, Kaggle is used to train the model. Sensor data is fed into the ML model for the nutrients analysis and water requirement analysis. An Android application is developed, which can be used from any handheld device by the farmers to take advantage of the proposed recommendation system. The result shows the promising future with better accuracy than previously available models in the same area. Parameters including recall, accuracy, precision, and F1-score are considered to gauge performance
A comparative analysis of GPUs, TPUs, DPUs, and QPUs for deep learning with python
Get PDFIn the rapidly evolving field of deep learning, the computational demands for training sophisticated models have escalated, prompting a shift towards specialized hardware accelerators such as graphics processing units (GPUs), tensor processing units (TPUs), data processing units (DPUs), and quantum processing units (QPUs). This article provides a comprehensive analysis of these heterogeneous computing architectures, highlighting their unique characteristics, performance metrics, and suitability for various deep learning tasks. By leveraging python, a predominant programming language in the data science domain, the integration and optimization techniques applicable to each hardware platform is explored, offering insights into their practical implications for deep learning research and application. the architectural differences that influence computational efficiency is examined, parallelism, and energy consumption, alongside discussing the evolving ecosystem of software tools and libraries that support deep learning on these platforms. Through a series of benchmarks and case studies, this study aims to equip researchers and practitioners with the knowledge to make informed decisions when selecting hardware for their deep learning projects, ultimately contributing to the acceleration of model development and innovation in the field
A hybrid machine learning approach for malicious website detection and accuracy enhancement
Get PDFMalicious URLs are web addresses purposely generated for a user’s detriment. Some examples include phishing scams in which the victim is fooled into logging into a fake site or portals for downloading malware where any click on a link invites a hostile program to the user’s device. The damage done to an individual’s finances, confidential information, and even reputation due to malicious URLs makes it crucial to devise means of countering these threats. This can be achieved by creating an intelligent model that identifies suspicious characteristics common to these websites. The objective of this research is to design a novel hybrid machine learning algorithm-based model for detecting malicious websites. A random forest, decision tree, and extreme gradient boosting (XGBoost) are the three hybrid classification algorithms proposed for the study. Accuracy in detection will help prevent and reduce the effects of such websites. The accuracy rate in this research is 98.7%, precision is at 98.9%, and recall at 98.5%. With these results, it follows that the hybrid model is more effective than training any individual algorithm with the given dataset
The variety of phosphor Ca2MgSi2O7:Eu2+ emission color affect white light LEDs
Get PDFThe conventional phosphor-converted white light emitting diode (WLED) suffers from several drawbacks relevant to heat generation and low rendered performance. Thus, using ultraviolet LEDs was introduced as a solution. It is essential to choose the phosphors with high stability that can activated under 350-410 nm to be compatible with the chips. Rare-earth-doped silicate phosphor is among the most reserched materials for solid-state light devices, thanks to its high stability and low-cost production. This work presents the Eu2+ -doped Ca2MgSi2O7 green phosphor to serve the pursuit of comprehensively enhancing the WLED performances. The f–d transitions and Eu2+ ions mixture take possession of two seperate cation spots in main grids with the help of two emission peaks, one at 465 nm and another at 520 nm. The composition of YAG:Ce3+ and Ca2MgSi2O7:Eu2+ phosphors, and a near-UV chip of 370 nm were utilized to compose WLEDs. Results show that by increasing the Ca2MgSi2O7:Eu2+ phosphor amount, the lumen output, correlated color homogeneity, and color rendering factors can be improved. The paper emphasizes the necessity for the optimal selection of the Ca2MgSi2O7:Eu2+ phosphor concentration, which would be about 10 wt%. The phosphors could be promising in making green-induced white luminous materials for white pc-LEDs with near UV-base
Energy efficient distributed intelligence on cognitive IoT gateway using MQTT protocols
Get PDFInternet of Things (IoT) facilitates communication between machines and devices which plays a crucial role in the conservation of energy. In largescale multidomain environments securing the data exchange among various IoT devices and key sharing creates a significant challenge. However, the message queuing telemetry transport (MQTT) lacks functional security mechanisms as well as mutual authentication between brokers and clients. To address these issues, a novel Cognitive IoT in Teroperability Recognition USing deep learning (CITRUS) framework is developed for real-time decision-making and sharing information among multiple IoT systems. Initially, the healthcare and weather data are collected remotely by using interoperable sensors which are then fed to the deep learning (DL) module for efficient decision-making. The MQTT module makes an energy-efficient IoT data communication over a resource-constrained network and the QoS1 introduces an acknowledgment and retransmission mechanism to ensure message delivery. The efficacy of the CITRUS model has been analyzed in terms of accuracy (AC), recall (RC), F1-score (F1S), sensitivity, packet delivery ratio (PDR), transmission speed, communication overhead, packet loss ratio (PLR) and delay. The experimental result shows that the CITRUS method achieves 89.89% of delay whereas, the IHPEC, SemBox, and DynoIoT methods achieve 161.63%, 128.99%, and 111.70% respectively for efficient data transmission
Holographic-based design, building, and testing of an RRP spherical robot for olive fruits harvesting
Get PDFA revolute-revolute-prismatic (RRP) spherical robot has been designed, simulated, built, and tested. The robot is intended to perform olive fruit harvesting tasks. The design simulation is done using hologram tools. The design factors considered include reach, dexterity, accuracy, and productivity. Based on the results of the holographic simulation, a prototype was built and tested on real olive fruits. The end effector is equipped with a rake tool so that the robot can harvest multiple fruits in each stroke. The robot is controlled by Raspberry Pi while a stereovision camera enables 3-D vision. Once the camera detects the fruits, an inverse kinematics algorithm is initiated to find the location of the fruits. The fruit coordinates are commanded to the manipulator to perform the harvesting. The field tests showed that the manipulator is successful in performing the harvesting operations. To increase the harvesting efficiency, it is recommended to build a larger prototype
Empowering microgrids: harnessing electric vehicle potential through vehicle-to-grid integration
Get PDFElectric vehicles (EVs) can potentially be integrated into microgrids via vehicle-to-grid (V2G) technology, which enhances the energy system's stability and durability. This paper provides an in-depth examination and evaluation of V2G integration in microgrid systems. It analyses the present state of research as well as possible uses, challenges, and directions for V2G technology in the future. This paper addresses the technological, economic, and regulatory aspects of implementing V2G and provides case studies and pilot projects to shed light on potential benefits and barriers associated with its adoption. The research highlights how V2G contributes to more efficient integration of renewable energy sources, grid stabilization, and cost savings for EV owners. It also addresses the latest developments in technology and proposed laws aimed at encouraging growing applications of V2G
Energy-efficient knapsack algorithm for intelligent cluster head selection in IoT enabled wireless sensor networks
Get PDFThe demand for wireless sensor networks (WSN) has grown rapidly with the development of the internet of things (IoT), which requires sensors that are both energy-efficient and scalable to support continuous data collection and real-time monitoring applications. The main challenge is limited battery life in network nodes, which necessitates effective energy management strategies to prolong network lifespan. This paper introduces an energyefficient knapsack algorithm (EEKA) for smart cluster head (CH) selection in IoT WSNs, aiming to optimize energy use while enhancing network stability and data transmission efficiency. The approach features a CH selection strategy based on residual energy, ensuring an even distribution of energy among sensor nodes. The incorporation of the knapsack optimization technique enhances resource allocation, thereby minimizing energy consumption and maximizing transmission reliability. Simulation results using NS2.34/2.35 show remarkable improvement in performance metrics compared to existing techniques: EEKA extends the network lifetime by 16% whereas throughput is enhanced by 17% with reduced latency by 14% under efficient data distribution. Moreover, adaptive CH selection strategy extends coverage by another 20% for wider and effective monitoring. All these results therefore confirm that EEKA has successfully focused on improving energy efficiency, stability, and scalability regarding IoT-driven WSNs to make it a practical solution for real-world applications like smart cities, environmental observation, and industrial automation
FPGA-based implementation of an S-Box cryptographic co-processor for high-performance applications
Get PDFThe increasing demand for reliable cryptographic operations for securing current systems has given birth to well-advanced and developed hardware solutions, in this paper we consider issues within the traditional symmetric advanced encryption standard (AES) cryptographic system as major challenges. Additionally, problems such as throughput limitations, reliability, and unified key management are also discussed and tackled through appropriate hierarchical transformation techniques. To overcome these challenges, this paper presents the design and field programmable gate array (FPGA)-based implementation of a cryptographic coprocessor optimized for substitution box (S-Box) operation which is considered as a key component in many cryptographic algorithms such as AES. The architecture of the co-processor proposed in this article is based on the advanced characteristics of FPGAs to accelerate the S-Box transformation, improve throughput and reduce latency compared to software implementations. We discussed carefully the design considerations along with resource utilization, speed optimization, and energy efficiency. The obtained experimental results present significant performance improvements, the FPGA-based implementation ensured higher throughput and lower execution time compared to traditional CPU-based methods. We presented in this work the effectiveness of using FPGAs for the acceleration of cryptographic operations in secure applications which will therefore be a robust solution for the next generation of secure systems
Performance evaluation of path planning algorithms for blind people
Get PDFBlind people face difficulties in identifying objects of interest and moving to them safely and efficiently in unfamiliar environments. Thanks to highperformance computers, high-quality sensors and artificial intelligence algorithms, it is possible to perform real-time tasks such as locating the user, generating occupancy grids that represent the environment and identifying objects of interest. From this information, paths can be generated that allow the user to reach a point of interest in an optimal way. This paper presents the performance evaluation of four path planning algorithms that were implemented in MATLAB and tested with synthetically generated occupancy grids, varying their size and occupancy percentage. The evaluation criteria include time to reach the goal, number of expanded cells and number of cells in the path. In addition, a single indicator that integrates all performance criteria is proposed to identify the best algorithm. The results show that the A* algorithm presents the best performance in static environments, under certain hardware requirements for data processing and restrictions on grid size for real-time applications. These findings expand the fields of application of path planning algorithms, quantify their performance under different conditions of the environment, and make them attractive for implementation in embedded systems