Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Modelling a neural network for analysing the results of segmentation of satellite images
Get PDFThe study's relevance lies in addressing inaccuracies within satellite image segmentation, necessitating the development and implementation of neural network models for automated segmentation. The purpose of study is to develop a model of a neural network for training with data obtained from the segmentation of satellite images. The basis of the methodological approach in study is a combination of methods of system analysis of neural networks, which have had a substantial impact on the development of the computer vision industry, with an empirical study of the general principles of neural network modelling for the training on satellite images segmentation. In this study, the results were obtained, indicating that there is a fundamental possibility of developing and practical implementation of a neural network model to determine the quality of the obtained segmentation of images of agricultural fields. Satellite images of agricultural fields of the Republic of Kazakhstan are obtained, and segmentation of field images is performed using the developed neural network model for learning segmentation results. The practical importance of the results obtained in study lies in the possibility of their use in the development of functional models of neural networks for training the results of the segmentation of satellite images
Leveraging transformer models for enhanced temperature forecasting: a comparative analysis in the Beni Mellal region
Get PDFThe remarkable impact of transformers in artificial intelligence, exemplified by applications like GPT-3 in language processing, has sparked interest in their potential for time series analysis. This study aims to explore whether transformers, specifically temporal fusion transformers (TFT), can outperform conventional methods in this domain. The research question is whether TFT exhibits superior performance compared to conventional recurrent neural network (RNN) methods, specifically gated recurrent unit (GRU), and traditional machine learning approaches, notably autoregressive integrated moving average (ARIMA), in the context of time series analysis and temperature prediction. A comparative analysis is conducted among three models: ARIMA, GRU, and TFT. The study utilizes time series data spanning from 1984 to the end of 2022. The models’ performances are evaluated using multiple metrics: mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and the coefficient of determination (R2). The TFT model achieves the lowest MAE, indicating high accuracy in its predictions. It outperforms both the RNN and traditional machine learning in temperature prediction tasks. Integrating the TFT model with the FAO penman-monteith method could improve irrigation scheduling due to more accurate temperature predictions, potentially enhancing water efficiency and crop yields
Artificial neural network-based intelligent sensor-based electronic nose for food applications
Get PDFFood commerce, especially for the general public, is greatly impacted by the capacity to identify and recognize chemical samples for food applications. Every chemical sample response has a unique, distinguishing smell. These advancements highlight the method of an artificial neural networks (ANN) to distinguish the distinctive fragrance from the reaction of substances. The categorization of various smell patterns has diminished confidence in ANN technology. Using an ANN technique and a sensor-based e-nose system for food applications, each chemical’s identification has been done commercially. The system comprises a 5-gas sensor selection that recognizes chemical talk while allowing for an improvement in permitting while falling gas is planned outside. To build a model of a different signal reaction, individual sensors are equally collected and merged into the innovation -favored sensor array. Demonstrates how it is related to the chemical test. The e-nose categorization has been tested with five different chemical samples and five different sensor classes. The e-nose approach, which comprises five sensors, can classify each chemical reaction model, starting with the results. With more sensors being employed, the classification accuracy of the precise chemical reaction improves. These data demonstrate that the ANN-based e-nose method promises a successful classification system for chemical sample responses for a characteristic odor sample
Stochastic geometry-based resource allocation scheme over cellular shotgun systems
Get PDFThis paper presents a resource allocation scheme that fulfills the maximum possible aggregate rate of the capacity region by targeting the corner points of the multiple-input multiple-output multiple access channel. This corner points of the channel’s capacity region are attainable whenever each user’s transmission has minimum possible interference among other users. This work aims to investigate the non-singularity of such situations by the exploitation of users’ geographic location seeking the opportunity of getting users’ transmission spatially multiplexed. The developed model demonstrates that similar results can be achieved with partial channel state information knowledge under certain conditions throughout the operational signal to noise ratio range. The proposed resource allocation scheme is designed for a shotgun cellular system with a random distribution of users over a circular coverage area. The proposed model uses stochastic geometry to prove that when number of users grows up within the coverage area, the probability of achieving the corner points sum rate increases rapidly. The developed model was evaluated, and the results show that for a circular coverage area with a radius of 10 km, the probability of having users whose transmissions can be spatially multiplexed with minimum interference increases as the number of users grows to 300 users
Enhancing EEG-based brain-computer interface systems through efficient machine learning classification techniques
Get PDFAdvances in the fields of neuroscience and computer science have greatly enhanced the human brain’s ability to communicate and interact with the surrounding environment. In addition, recent steps in machine learning (ML) have increased the use of electroencephalography (EEG)-based BCIs for artificial intelligence (AI) applications. The prevailing challenge in recording EEG sensor data is that the captured signals are mixed with noise, which makes their effective use difficult. Therefore, strengthening the classification stage becomes extremely important and plays a major role in addressing this problem. In this study, we chose five most widely used classification models that obtained the best results in this field and tested them on two open-source databases. We also focused on improving the hyperparameters of each algorithm to obtain best results. Our results indicate excellent results on the first dataset and acceptable for most models on the second, while RF showed superior performance on both with an accuracy of 100% on the first dataset and 86.47% on the second. This was achieved with the lowest training costs, and better performance compared to previous works we evaluated that used the same databases. These results provide valuable insights and advance the development of brain-computer interface (BCI) technology and design
Energy baseline model enhanced based on artificial neural network in industrial buildings
Get PDFIn this article, a new energy-efficient reference model has been established for a plastic injection molding plant. However, the proposed model handled difficulties due to the lack of robust and complete data, such as production mix and cooling degree-days. In addition, the proposed model applies three distinct enhanced modeling methodologies, including regression modeling, artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). Furthermore, these performance parameters were established to assess the accuracy of each model in this work. Moreover, the numerical results show that among the methodologies used in this work, the ANN demonstrated effective performance despite uncertainties in the measured input variables. The ANN numerical results in this paper highlight the ability to accurately assess baseline consumption in the industrial sector, providing a practical tool for decision-makers to improve energy efficiency
Sampled-data observer design for sensorless control of wind energy conversion system with PMSG
Get PDFThis paper presents a nonlinear observer for a variable-speed wind energy conversion system (WECS) utilizing a permanent magnet synchronous generator (PMSG). The study addresses the design of high-gain sampled-data observers based on the nonlinear WECS model, supported by formal convergence analysis. An essential aspect of this observer design is the incorporation of a time-varying gain, significantly enhancing system performance. Convergence of estimation errors is demonstrated using the input-to-state stability method. Simulation of the proposed observer is conducted using the MATLAB-Simulink tool. The obtained results are presented and analyzed to showcase the overall effectiveness of the proposed system
Spatial domain noise removal filtering for low-resolution digital images
Get PDFIn this research work, six different filters are applied on a low resolution 8 b/pixel gray-scale images, which operate on small sub-images (windows of 3×3 to 11×11 pixels). The enhanced images are used to compare the efficiency of the different six filters using the peak signal to noise ratio (PSNR) image quality measure. Noise peak elimination filter (PSNR)=36.63) outperforms others, such as median filter (PSNR=36.61), while corruption estimation (PSNR=36.03) significantly cuts processing time by only processing the corrupted pixels while maintaining image details. Mean filter (PSNR=34.05) is sensitive to outliers, which cause the image's sharpness and fine features to be lost. By avoiding averaging across edges, bimodal-averaging filter (PSNR=35.30), which improves on the mean filter, chooses the mean of the biggest population. The median-mean filtering (PSNR=36.32), which combines median and mean filters and determines the output pixel by averaging the median and some nearby pixels, is another improvement above averaging
Notice of Retraction An NFMF-DBiLSTM model for human anomaly detection system in surveillance videos
No full textNotice of Retraction-----------------------------------------------------------------------After careful and considered review of the content of this paper by a duly constituted expert committee, this paper has been found to be in violation of IAES's Publication Principles. We hereby retract the content of this paper. Reasonable effort should be made to remove all past references to this paper.-----------------------------------------------------------------------In response to the increasing demand for an intelligent system to avoid abnormal events, many models for detecting and locating anomalous behaviors in surveillance videos have been proposed. Nevertheless, significant flaws of inadequate discriminating ability are present in the majority of these models. A novel newton form and monotonic function based deep bidirectional long short-term memory (NFMF-DBiLSTM) human anomaly recognition system was discussed in this paper to tackle those issues. Initially, videos are transformed into frames; after that, the duplicate frames are removed, and by utilizing the shannon entropy centered contrast limited adaptive histogram equalization (SE-CLAHE) algorithm, the contrast has been elevated. By using the probabilistic matrix factorization kernel density estimation (PMF-KDE) technique, the background is subtracted after estimating only the motion of the object. After this, the silhouette function is performed utilizing the dirac depth silhouette function (DDSF). In addition, clustering is done by sorting and average-based K-means (SA-KM). The features are extracted from the suspected human and are then chosen by utilizing Poisson Eurasian oystercatcher optimization (PEOO). For classifying normal or anomaly, the selected features are subjected directly into the NFMF-DBiLSTM. When contrasted with the prevailing methodologies, the proposed model is found to be more efficient
Design of energy efficient and reconfigurable sample rate converter using FPGA devices
Get PDFThe technique of sampling rate conversion is frequently employed in various fields. A discrete time-varying filter, as well as a sample skip or sample duplicate operation, are required for the most general instance of an irrational and time-variable conversion factor. A wide band of signals is employed in a communication system, especially in specific situations where data must be transferred directly. A broadband sample rate converter with changeable filter parameters is necessary in such cases. Sample rate conversion is a communication system technology that accepts a band-limited high sample rate modulated signal and uses filtering to retrieve the original message signal. In this work, an energy-efficient implementation of a reconfigurable field programmable gate arrays (FPGA) architecture for a sample rate converter is proposed. In applications such as multi-rate signal processing and the construction of channelized receivers, sample rate conversion is used. In this work, a new FPGA based design is proposed to perform multiple sample rate conversion for various data transmission protocols such as Wi-Fi, ZigBee and Bluetooth. A lowpass filter with a 2.45 GHz filter with the minimum number of taps is used to avoid the aliasing effect. Xilinx synthesis tools are used to estimate hardware resource utilization and speed analyses. XC6VCX240t-2FF484 FPGA achieves 15% hardware resource occupancy at a maximum clock speed of 133 MHz