Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Not a member yet
776 research outputs found
Sort by
Design of Robust Centralized PID Optimized LQR Controller for Temperature Control in Single-Stage Refrigeration System
Full text linkRefrigeration systems are used for many purposes such as food preservation, cooling and others. They require controller to ensure that the refrigerating cycle can go ON and OFF to maintain a setpoint temperature. For instance, in preservation of food or other perishables, deterioration can occur without efficient system to ensure that temperature within refrigerating space is kept at a setpoint value. This paper presents robust centralized proportional integral and derivative (PID) optimized linear quadratic regulator (LQR) temperature control system for single-stage refrigeration system. A composite technique in which PID algorithm was used to adjust the gains of LQR is proposed. The model of single-stage vapour compressor refrigeration (VCR) system was established in terms of the evaporator, compressor, condenser and the expansion valve’s temperatures. An LQR was initially designed. Then a PID optimized LQR was design. The results indicated that the PID optimized LQR controller outperformed the LQR by providing 73.4% and 62.7% improvement for the evaporating temperature, 45.6% and 71.4% improvement for the compression temperature, 30% and 84.6% improvement for the condensing temperature, and lastly 72% and 70.2% improvement for the expansion temperature in terms of response time and settling time. Simulation with test data proved its robustness and effectiveness in tracking setpoint temperature. Generally, the proposed system has shown capacity to offer robust and centralized tracking in the presence of changing setpoint values.
Cyber Security Threat Prediction using Time-Series Data With LSTM Algorithms
Full text linkCyber security remains a paramount concern in the digital era, with organizations and individuals increasingly vulnerable to sophisticated cyber-attacks. This study aims to develop and evaluate Long Short-Term Memory (LSTM) regression models to predict three types of cyber attacks: flood, spyware, and vulnerability. The LSTM algorithm is used to construct regression models for spyware, flood, and vulnerabilities within a firewall log dataset. The experiments demonstrate that preprocessing techniques such as normalization and standardization can positively impact model performance by reducing prediction errors and enhancing accuracy. The results of the experiments show that the model developed in this research exhibits potential in predicting cyber attacks. For the flood attack model, the best performance was achieved with an RMSE of 59.8810 and an R-Squared of 0.9214 after data standardization. The spyware attack model's best results were an RMSE of 133.9567 and an R-Squared of 0.7685 after standardization. In contrast, the vulnerability attack model showed limited improvement, with the best RMSE of 503.5521 and an R-Squared of 0.2358 after standardization. Moreover, real-time implementation and testing of these models in live network environments could validate their practical applicability and effectiveness
Improving Bi-LSTM for High Accuracy Protein Sequence Family Classifier
Full text linkThe primary nutrient that is crucial for identifying biochemical processes and biological norms in living cells is protein. Proteins are usually centered around one or a few functions which are defined by their family type. Hence, identification and classification are needed to separate the proteins according to their structure and families. In this work, we built a model to classify families of protein sequences. We used the protein sequences dataset consists of various macromolecules of biological significance. The classifier is built up using deep learning of Bi-LSTM. We began the research by collecting the dataset from the Protein Data Bank of the Research Collaboratory for Structural Bioinformatics, pre-processing the data using tokenizing, and modeling the classifier based on deep learning network of Bi-LSTM. As we get the best accuracy rate of the trained model, we figure out the model performance using the evaluation metrics of learning curve, accuracy rate, and loss. The results show that Deep Bi-LSTM provides excellent performance with fit learning curve, 99% accuracy rate, and 0.042 loss
The Effectiveness of Using the Pearson’s Correlation Coefficient for Compression Quality Assessment: Case of ECG Signals Compression with Discrete Cosine Transform
Full text linkThis work presents a comparative study which aims to validate the importance of using the Pearson’s Correlation Coefficient (PCC), for the first time in this field of research, as an effective parameter for quantitative measurement of Electrocardiogram (ECG) signal compression quality. The comparison with the Percent of Root mean squared Difference coefficient (PRD) was carried out using the Discrete Cosine Transform (DCT). The ECG signals of the three derivations DI, DII and DIII, used for the test, belong to five categories of patients with various pathologies, each category of which includes four patients. The obtained results, based on the morphology comparison of P waves, T waves and QRS complexes before compression and after reconstruction, showed that the range of values between 99.90% and 100% for the PCC, ensures a very good signal reconstruction quality with a Compression Ratio (CR) that could reach 16
Regulation of Active and Reactive Powers in Doubly-Fed Induction Generators Utilizing Proportional-Integral and Artificial Neural Network Controllers
Full text linkIn this paper, vector orientation and neural networks are used to simulate and regulate a Doubly Fed Induction Generator (DFIG) wind turbine. The aerodynamic turbine and DFIG dq models are developed. PI current regulation is used in vector control to separate active and reactive power control. To reproduce the PI response, training networks create a different neural vector control scheme. Comparative simulations confirm the effectiveness of both control methods in following set points and counteracting disturbances. The neural vector control scheme outperforms the PI scheme in managing short-term changes. In contrast to the PI control, it has quicker response times for both rising and settling. Neural vector control enables precise and rapid tracking of electromagnetic torque. Neural vector control could improve the performance of DFIG wind turbines because it has an adaptive architecture that lets it respond well to changes in parameters and maintain its accuracy over time. Additional investigation is needed to improve neural network training techniques and incorporate them with conventional control systems
A Development of Supporting System for Historical Heritage Based Tourism
Full text linkTourism is a major economic contributor in Thailand. With the richness of historical heritage recognized as world heritages, Phra Nakhon Si Ayutthaya province is a famous destination for tourists who enjoy historical and cultural tourism. This work presents a development of a supporting system for tourism in Phra Nakhon Si Ayutthaya province in regarding of historical and cultural aspects of heritages. This work designs an ontology to represent a relation network of properties from tourist attractions based on historical and cultural relationship among them. Instances which are the heritages hence are related and can be visualized in a form of a graph. The suggestion module is designed to provide related tourist destination following the relations from the generated knowledge graph based on the initial query of a user. The experiment results signify that the system revealed hidden historical relations of destinations to users and made them learn the values of history lied within heritages. Furthermore, 87.5% of participants decided to make a tour plan following the suggested destinations since they found the linking in historical values to be more meaningful and interesting
Classification of Darknet Traffic Using the AdaBoost Classifier Method
Full text linkDarknet is famous for its ability to provide anonymity which is often used for illegal activities. A security monitor report from BSSN highlights that 290.556 credential data from institution in Indonesia have been exposed on the darknet. Classification techniques are important for studying and identifying darknet traffic. This study proposes the utilization of the AdaBoost Classifier in darknet classification. The use of variable estimator values significantly impact classification results. The best performance was obtained with an estimator value of 500 with an accuracy of 99.70%. The contribution of this research lies in the development of classification models and the evaluation of AdaBoost models in the context of darknet traffic classification
Efficient Invisible Color Image Watermarking Based on Chaos
Full text linkSeveral difficulties are faced in developing a robust and transparent color image watermarking system, which requires the blending of the human visual system (HVS) during its design. Therefore, employing masks that take into account the features of HVSs has become a very effective tool for boosting robustness requirements without significant alterations in image imperceptibility. The present article offers watermarking strategy for colored images employing a reverse self-reference image in conjunction with the HVS constraint. A color image first undergoes conversion through the Red, Green, and Blue (RGB) format to the National Television Systems Committee (NTSC) space. The reference image is derived from the luminance channel through the discrete wavelet transform (DWT) domain. However, the chaotic map serves to generate the watermark, and a 2D torus automorphism is subsequently used to scramble the watermark. Therefore, the watermark is scrambled and placed in the reference image. Moreover, the detecting phase involves the host image, where the reference image is extracted from both the host and the image with a watermark, and the correlation is subsequently used to assess the similarity between the retrieved and the introduced watermark. The proposed watermarking scheme can retain the watermarked image's perceptibility justified by the PSNR. In addition, it achieves high robustness to withstand a wide array of attacks.
A Diet Recommendation System using TF-IDF and Extra Trees Algorithm
Full text linkAcross the globe, there is a growing emphasis on health and lifestyle choices. However, refraining from unhealthy foods and staying active are not enough; maintaining a well-balanced diet is also essential. Recently, recommendation systems have focused on promoting healthy eating habits, tailoring suggestions for balanced diets based on some parameters like age, gender, height, weight, age, BMI and BMR. Pairing a nutritious diet with regular physical activity can aid in reaching and sustaining a healthy weight, reducing the likelihood of chronic ailments such as heart disease, and enhancing overall well-being. The present paper introduces a novel approach for constructing dietary recommendations with optimized calorie intake, using content-based filtering with the TF-IDF statistical method and machine learning with the Extra Trees algorithm. This approach can generate a dynamic diet based on the calories a person burns and other parameters including the current Body Mass Index (BMI) and BMR (Basal Metabolic Rate). The proposed approach has been tested on a new realworld diet dataset, showcasing its effectiveness in providing diverse and accurate diet recommendations compared to another content-based filtering method and other machine learning algorithms
A Translation Framework for Cross Language Information Retrieval in Tamil and Malayalam
Full text linkCross Language Information Retrieval (CLIR) stands as an essential element in multilingual information accessibility, enabling users to obtain relevant information even when the query language and the language of the documents diverge. This paper proposes a translation framework for CLIR in Tamil and Malayalam, two Dravidian languages widely spoken in South India. Different challenges prevail in CLIR of these languages due to their linguistic differences, translation equivalence, mapping source to target languages, semantic equivalence, limited dataset and tools for ongoing research in this domain. The proposed methodology resolves some of the issues around training of a corpus utilizing a Long Short-Term Memory (LSTM) based encoder-decoder translation model. The study incorporates two bilingual parallel corpora comprising 373 sentences pairs each. Evaluation of the model's accuracy is conducted by equivalency its translations against reference translations using the Bilingual Evaluation Understudy (BLEU Score). Furthermore, BLEU scores obtained from proposed LSTM-based encoder-decoder model is compared with those from Google Translate. The findings reveal that the LSTM model attains an average BLEU score of 0.933, where, performance of Google Translate, achieved a score of 0.813. Finally, the study conducts a comparative analysis with selected CLIR models in different languages, to evaluate the overall performance of the proposed approach