Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
ABSA of Indonesian customer reviews using IndoBERT: single- sentence and sentence-pair classification approaches
Get PDFAspect-based sentiment analysis (ABSA) task is important to identify user satisfaction from customer reviews by recognizing the sentiments of all aspects discussed in the reviews. This work investigates a novel study on the effectiveness and efficiency of three IndoBERT-based models for solving the ABSA task in Indonesian language. IndoBERT is a state-of-the-art transformer-based model, i.e., bidirectional encoder representations from transformers (BERT), that was pre-trained on Indonesian language. Our first model utilizes IndoBERT in a feature-based mode, paired with the convolutional neural network (CNN) and machine learning models, for single-sentence classification. Next, our second model is obtained by fine- tuning the IndoBERT model for a typical single-sentence classification to build an end-to-end model. At last, our third model also adopts a fine-tuning approach to use IndoBERT, but for sentence-pair classification by utilizing auxiliary sentences. Our results demonstrate that the third model, the fine- tuned IndoBERT for sentence-pair classification, gains the highest effectiveness. It demonstrates significant improvement over deep learning baselines (Word2Vec-CNN-XGBoost) by 23.6% and transformer-based baselines (mBERT-aux-NLIB) by 2.2% in terms of F-1 score. When considering both effectiveness and efficiency, the results show that the best- performing model is our second model, the fine-tuned IndoBERT for single- sentence classification
A novel framework of building operation algorithm for the block of technical diagnostics of aircraft’s automatic control system
Get PDFThis article presents the problem of designing an automatic control system that is stable against errors and failures of sensors on aircraft. The sensor system has a technical diagnostic block that ensures diagnosis and eliminates typical errors and failures. Based on the determination of the error vector, damage can occur by adding measurement elements corresponding to the measurement parameters to the control system. When there are errors or failures of the sensor elements, the state vector of the system changes and is determined by measurements. The difference between the measured vector components when there are errors, failures and when working normally is the basis of the working algorithm of the failure diagnosis block. The results demonstrate encouraging prospects for practical implementations
IoT-based fertigation system for agriculture
Get PDFFertigation system has been widely used by farmers to automate some processes of crops productions. A conventional system requires workers to prepare a fertilizer mixture, before transferring it into a main storage tank to be mixed with water. Then, electrical conductivity (EC) of the mixture will be measured. The existing fertigation system still relies heavily on workers and is manually operated and prone to human error. Therefore, internet of things (IoT) based fertigation system has been developed to deliver the fertilizer mixture with consistent EC value automatically to the plants. The main system controller is designed using ESP32 development module. The operation of the system can be monitored using an IoT dashboard and farmers can also control the system remotely. Alert will be given to the farmers if the condition of the system or plant does not meet the predefined settings. The values of EC together with temperature and humidity sensors are recorded for further analysis. A testbed is set up to provide fertigation to 120 polybags eggplants. Using the proposed fertigation system, the eggplants have been harvested earlier, therefore reducing the fertilizer usage. The cost of this IoT based fertigation system is lower compared to existing commercial products
The deep convolutional networks for the classification of multi-class arrhythmia
Get PDFAn arrhythmia is an irregular heartbeat. Many researchers in the AI field have carried out the automatic classification of arrhythmias, and the issue that has been widely discussed is imbalanced data. A popular technique for overcoming this problem is the synthetic minority oversampling technique (SMOTE) technique. In this paper, the author adds some sampling of data obtained from other datasets into the primary dataset. In this case, the main dataset is the Massachusetts Institute of Technology–Beth Israel Hospital (MIT-BIH) arrhythmia database and an additional dataset from the MIT-BIH supraventricular arrhythmia database. The classification process is carried out with one-dimensional convolutional neural network model (1D-CNN) to perform multiclass and subject-class advancement of medical instrumentation (AAMII) classifications. The results obtained from this study are an accuracy of 99.10% for multiclass and 99.25% for subject-class
Power loss estimation utilizing the flexibility of peak power loss regression equations based on 11 kV base case feeder
Get PDFDistribution network feeder characteristics can typically be divided into groups based on factors including length, load distribution along the feeder, peak demand, installed capacity, and load profile. By comparing the parameters to those of similar feeders with known losses, it is usually possible to predict the power losses and technical losses (TL) of the respective feeders pretty accurately. However, it is exceedingly difficult and time-consuming to estimate the losses with various variables and characteristics over such a large area. This paper proposed that through base case feeder modeling and simulation utilizing typical network and load data, feeders’ peak power loss (PPL) functions can be established as a simple and effective power loss estimation method. Hence, the least time-consuming way of using a PPL regression equation based on a base case feeder is established in this paper to estimate the losses. The flexibility of PPL is proven through the case study. In the end, the results obtained between PPL and peak power demand (PPD) are demonstrated to be precisely proportional and the method is proven as a simple power loss estimation method due to the flexibility of the PPL regression equation
Speed control for traction motor of urban electrified train in field weakening region based on backstepping method
Get PDFTractor motors always operate in the speed region higher than rated speed, but is limited to the module of the stator current, stator voltage vectors. Additionally, mathematical model of traction motor has shown nonlinearity through the product of the state variables ð‘–ð‘ ð‘‘, ð‘–ð‘ ð‘žÂ with the input variable ðœ”ð‘ :ðœ”ð‘ ð‘–ð‘ ð‘ž, ðœ”ð‘ ð‘–ð‘ ð‘‘. Therefore, this paper focuses on the study of speed control of traction motors in weakening field region while optimizing torque control, and choosing the backstepping method in designing speed–flux controller in order to solve the nonlinear structure. The simulation results of the responses: speed, torque, power, and flux performed on MATLAB/Simulink software with parameters collected from metro Nhon-Hanoi Station, Vietnam have proven the correctness in theoretical research
Robust sliding mode observer based-simultaneous state and actuator fault estimation for a class of switching systems
Get PDFThis paper investigates the state and actuator fault reconstruction problem in a class of switched linear systems subjected to unknown external disturbances according to average dwell time (ADT) technique. First, a robust switched sliding mode observer (SMO) is developed to simultaneously reconstruct the states of the switched system and the actuator faults. A novel and less conservative sufficient stability conditions are then established using the multiple quadratic Lyapunov function technique and the ADT approach. These conditions are formulated as linear matrix inequalities (LMI) to facilitate the design of the SMO. The observer gains matrices are obtained throughout the resolution of LMI using convex optimization techniques. Next, actuator faults are estimated by utilizing the concept of equivalent output injection, achieved through an analysis of the state error dynamics during the sliding motion. Finally, simulation results are considered to illustrate the applicability and efficiency of the developed method. It showcases the rapid and accurate convergence of the estimated system states and actuator fault to the real variables
Elitism-crossover barnacle mating optimization and its application to PID controller design for a buck converter
Get PDFThis paper presents an elitism-crossover barnacle mating optimization (ECBMO). It is an improvement of barnacle mating optimization (BMO). The original BMO suffers from local optima problem leading to a low accurate solution. A new method of offspring generation is adopted into the original BMO structure. Some features of the best-so-far solution are incorporated into the generated offspring. The accuracy performance of the proposed algorithm is tested on several IEEE functions. A statistical analysis is conducted to compare its performance over the original BMO. It is also applied to optimize proportional integral derivative (PID) parameters for controlling output voltage of a buck converter. Result of benchmark functions test shows the proposed algorithm has attained higher accuracy for all functions compared to BMO algorithm. Application on the real problem shows both algorithms control the converter voltage satisfactorily. However, the ECBMO has achieved more optimal PID parameters and leading to a better output voltage response
Unmanned aerial vehicle path planning in a 3D environment using a hybrid algorithm
Get PDFThe optimal unmanned aerial vehicle (UAV) path planning using bio-inspired algorithms requires high computation and low convergence in a complex 3D environment. To solve this problem, a hybrid A*-FPA algorithm was proposed that combines the A* algorithm with a flower pollination algorithm (FPA). The main idea of this algorithm is to balance the high speed of the A* exploration ability with the FPA exploitation ability to find an optimal 3D UAV path. At first, the algorithm starts by finding the locally optimal path based on a grid map, and the result is a set of path nodes. The algorithm will select three discovered nodes and set the FPA's initial population. Finally, the FPA is applied to obtain the optimal path. The proposed algorithm's performance was compared with the A*, FPA, genetic algorithm (GA), and partical swarm optimization (PSO) algorithms, where the comparison is done based on four factors: the best path, mean path, standard deviation, and worst path length. The simulation results showed that the proposed algorithm outperformed all previously mentioned algorithms in finding the optimal path in all scenarios, significantly improving the best path length and mean path length of 79.3% and 147.8%, respectively
Dynamics and kinematics of complex mechanical systems harnessing multibody dynamic program
Get PDFUnderstanding the behavior and performance of engineering applications like machines, transport machines, manipulators, and mechanisms like gears relies heavily on the study of the dynamics and kinematics of complex mechanical systems. This article provides a comprehensive overview of recent findings and advancements in this field. The purpose of this work is to provide an in-depth introduction to the theoretical and practical considerations involved in assessing the dynamic and kinematic properties of such complex systems. Understanding forces, torques, displacements, and velocities is highlighted as crucial to the design and study of complex mechanical systems, and the underlying mathematical models and concepts that control their motion are investigated. This paper also evaluates and critiques the most current developments in modeling and simulation approaches such as finite element analysis (FEA), computational dynamics, and optimization strategies. The multidisciplinary aspect of the topic and its potential to progress numerous engineering, robotics, and industrial applications constitute the topic's scientific uniqueness. The results include various advanced modeling and simulation techniques like FEA, computational dynamics, and multibody dynamics simulation. In conclusion, this article compiles a lot of information on the dynamics and kinematics of sophisticated mechanical systems, such as machines, transport machines, manipulators, and mechanisms