Asian Journal of Convergence in Technology
Not a member yet
868 research outputs found
Sort by
Simulation of a bi-directional buck-boost converter and application for the vehicle to grid, grid to vehicle
Get PDFThis paper presents a bidirectional buck-boost converter for bidirectional power flow, electric vehicle applications, and for power flow from source to load and load to source. Which is a vehicle to grid and grid to vehicle in terms of electric vehicle applications. The simulation has been done on a bidirectional buck-boost converter in MATLAB software using a hybrid energy storage system. Simulation proves the suitability for usage in electric vehicle applications and a hardware prototype has been created of the bidirectional buck-boost converter to verify the results of the converters working. In this converter soft switching can be used to reduce switching losses at many parts and to increase the overall efficiency of the bidirectional buck-boost converter
Application of Heuristic Search Algorithm to Design Automatic Generation Control in Power System
Get PDFThis paper presents a Heuristic Search Algorithm, Conventional Particle Swarm Optimization (CPSO) which is employed to design Automatic Generation Control (AGC) for an isolated power system to minimize the variation in frequency due to changes in load demand. Further, Advanced Particle Swarm Optimization (APSO) with a velocity update strategy is introduced to ensure that the particles will orbit toward their optimum point quickly which guarantees faster convergence
An Image-based Intelligent System for Data Extraction
Get PDFAutomation is the process of providing goods and services with fewer to no human interventions. The major advantage of automation is reduction in human error. The system proposes to extract data from images that are tilted at different angles and noisy. The system reduces human error by storing the data directly in the database. The proposed system will take image input from the user through a user interface. This interface is a web application. The input image is pre-processed and forwarded to a machine learning model. The machine learning model is trained and tested using a character data set and convolutional neural network. The model will detect the characters and will give the output as recognized text. This output will be automatically stored in the database and shared with the user through the same interface
Novel Design of Highly Directive Mictrostrip Patch Antenna with Air Substrate
Get PDFThis paper presents a novel design and analysis of inset fed microstrip patch antenna with an air substrate operating at 2.4 Ghz. The proposed antenna has a high directivity and high radiation efficiency which are 2 important parameters in antennas for communications, radars, wireless power transmission and other applications. The S parameters, VSWR, Directivity, gain, beamwidth, radiation efficiency, electric field intensity and size of the design are presented. Ansys HFSS was used for the study. Small amounts of FR4 material was also used
Steganography Information Retrieval Mechanism Using Deep Neural Network
Get PDFSteganography has captivated the interest of a rising number of academics in recent years, as its applications have grown beyond information security. The most common approach is digital signal processing (DSP), which includes least significant bit (LSB) encoding. Deep learning has lately been used in various innovative approaches to the steganography problem. The bulk of existing methods, on the other hand, are designed for image in picture steganography. This study proposes using deep learning algorithms to disguise clandestine audio in digital photos. The first network conceals the concealed audio in a picture, while the second network decodes the image to retrieve the actual audio. In-depth research make use of a set of 24K images and the VIVOS Corpus audio dataset1. Experimental data shows that our strategy is more effective than earlier methods. Both the visual and audio integrity are preserved, and the maximum length of the concealed audio is substantially extended
Machine Learning Algorithms for Heart Disease Prediction
Get PDFCardiovascular disease, otherwise known as heart disease, encompasses many diseases that affect the heart. Heart disease prediction is among the most complicated tasks in medical field. In the modern age, about one person dies every minute as a result of heart disease. In addition to many factors that contribute to heart disease, it's necessary at this point in time to acquire accurate, reliable, and sensible approaches to make an early diagnosis so that the disease may be managed appropriately. Due to the complexity of finding out the heart condition, the prediction process must be automated to avoid risks related to it and to alert the patient at an early stage. In the healthcare domain, data mining is commonly used to analyze huge, complex medical data and predict heart disease. Researchers apply a variety of data mining and machine learning approaches to analyse huge complex medical data and predict heart disease. In this study, various heart disease attribute are presented, and model is developed on the basis of supervised learning algorithm as K-nearest neighbor, Decision Tree, Random Forest, Logistic Regression, SVM, Light GBM and Naïve bayes. This Paper makes use of heart condition dataset available in Kaggle repository. The purpose of this study is to anticipate heart disease risk in patients. The results show that K-nearest neighbor provides the most accurate result
Artificial Neural Network: A brief study
Get PDF An Artificial Neural Network (ANN) is a data processing paradigm inspired by the way biological nervous systems, such as the brain, process data. The unique structure of the information processing system is a crucial component of this paradigm. It is made up of a huge number of highly interconnected processing elements (neurons) that work together to solve issues. ANNs, like humans, learn by example, and a huge dataset results in more accuracy. Through a learning process, an ANN is trained for a specific application, such as pattern recognition or data classification. This is also true of ANNs. This paper provides an overview of Artificial Neural Networks (ANN), their working, and training. It also describes the application and benefits of ANN
EXERCISE POSE PREDICTION USING CONVOLUTIONAL NEURAL NETWORK (CNN) AND RESIDUAL NETWORKS (ResNet)
Get PDFImage classification is broadly used in almost all the fields. It can be used in medical, military, surveillance and many other fields. In this paper, we carried out image classification for four classes of exercise poses. Out of all available methods for image classification, we chose neural networks for classification. The CNN and ResNet50 algorithms were implemented and results were included in this paper. This can be applied in various exercise related fields like exercise monitoring, exercise posture correctness, virtual exercise training, etc
Review of Various Applications of Machine Learning
Get PDFThe usage of machine learning proposes an intelligent diagnostic study program that supports a web-based learning model aiming to develop students' ability to integrate information by allowing them to select study topics of interest, find information on those topics by searching online for related reading courseware and discussing what they have learned with their peers. The suggested learning program can effectively help students improve their knowledge while browsing online using the "webbased learning" approach, based on our test results. This study, on the other hand, proposes to use a machine-learning algorithm to anticipate future stock prices by combining open source libraries with pre-existing algorithms to help make this uncertain business model predictable. The result is entirely dependent on numbers and is predicted by many assumptions that may or may not occur in the real world, such as the forecast period. At the same time, the study also aims to provide a tool to anticipate accurate and timely traffic data. This fact has prompted us to pursue a solution to the problem of predicting traffic flow based on traffic data and models. Due to a large amount of available data for the transport system, it is difficult to accurately predict traffic flow
Detection of SAR and Penetration Depth of EM waves on Human body with respect to Cellular 4G/LTE Base Stations
Get PDFThe advancement of cell phone correspondence design on the planet has further developed, leading to public worry over conceivable medical problems and openness to radio recurrence of electromagnetic energy produced from the cell base stations. The miniature strip fixes receiving wire display a huge job in electromagnetic energy communicating and getting conditions in a cell phone. In day-to-day existence, people are impacted by the electromagnetic radiations produced from cell base stations. Because of the intricacy of the human body structure, the estimation of the impact of the radiation is so troublesome. Two central points are predominantly considered among different elements while considering the impact of radiation from cell base station, they are specific absorption rate (SAR) and skin depth (SD). The specific absorption rate (SAR) and skin depth (SD) are straightforwardly registered as more complicated. In this paper, we assess the mathematical investigation of the SAR and SD concerning the genuine utilization of working recurrence of the cell phone for 4G/LTE correspondence. To start the model, MATLAB recreation devices were utilized and the result from that model was dissected results were contrasted hypothetically and with different scientists