2024. 2^nd Issue

Volume XVI, Number 2

Table of contents 

Full issue  

 

MESSAGE FROM THE EDITOR-IN-CHIEF

Pal Varga
Infocommunications Journal welcomes HTE 75  
Our publisher, HTE – the Scientific Association for Infocom munications – reaches its platinum age, turning 75. This ever evolving, independent, professional organization was officially reg istered on January 29, 1949, and for 75 years, it has been providing objective representation for the entire infocommunications sector (ICT; telecommunications, information technology, media), facili tating the acceptance and development of technological advance ments. The secret to its long-standing success lies in its ability to continuously renew itself: by creating modern scientific and profes sional platforms, it expands the knowledge base of those interested, and by utilizing the collective expertise of its members, it provides opinions on Hungarian and EU projects, compiles comprehensive technical reports, and offers customized training sessions.

Download

 

PAPERS FROM OPEN CALL

Simon Dahdal, , Lorenzo Colombi, Matteo Brina, Alessandro Gilli, Mauro Tortonesi, Massimiliano Vignoli, and Cesare Stefanelli
An MLOps Framework for GAN-based Fault Detection in Bonfiglioli’s EVO Plant 
In Industry 5.0, the scarcity of data on defective components in smart manufacturing leads to imbalanced data sets. This imbalance poses a significant challenge to the develop ment of robust Machine Learning (ML) models, which typically require a rich variety of data for effective training. The imbal ance not only restricts the models’ accuracy but also their ap plicability in diverse industrial scenarios. To tackle this issue, our research delves into the capabilities of Deep Generative Models, with a special focus on Generative Adversarial Networks, for the generation of synthetic data. This approach is aimed at rectify ing dataset imbalances, thereby enhancing the training process of ML models. We demonstrate how synthetic data can substan tially bolster the performance and reliability of ML models in industrial settings. Furthermore, the paper presents an innova tive MLOps pipeline and architecture, meticulously designed to incorporate Deep Generative Models (DGMs) into the entire ML development cycle. This solution is automated and goes beyond mere automation; it is self-optimizing and capable of making necessary corrections, specifically engineered to address the dual challenges of data imbalance and scarcity, thus enabling more precise and dependable ML applications in smart manufacturing. 

Reference
DOI: 10.36244/ICJ.2024.2.1
Download 

 

Gábor Árpád Németh and Máté István Lugosi
MTR Model-Based Testing Framework 
In this article we propose a novel, free and open- source model-based testing framework for finite state machine specifications. The various model handling and test generation options make the framework suitable for testing complex systems and provide a solid background for investigating different automated test design methodologies. The complexity and fault detection capabilities of the available algorithms are investigated through analytical analyses and simulations applying randomly injected faults. 

Reference
DOI: 10.36244/ICJ.2024.2.2
Download 

 

M. Z. I. Nazir, M. Alqaradaghi, and T. Kozsik 
Automated checker for detecting method-hiding in Java  programs 
Method overriding is a valuable mechanism that happens when an instance method is defined in a subclass and has the same signature and return type as an instance method in the superclass. However, in Java, if those methods are static, then instead method hiding happens, which is a programming weakness and may produce unexpected results. Static analysis is an approach in software testing that examines code to identify various programming weaknesses throughout the software de velopment process without running it. This paper addresses the detection of method-hiding problem in Java programs. We implemented a new automated checker under the SpotBugs static analysis tool that can detect the mentioned problem. According to our results, the checker precisely detected zhe related issues in both custom test cases and realword programs.

Reference
DOI: 10.36244/ICJ.2024.2.3
Download 

 

T. Padmavathil, Kusma Kumari Cheepurupalli, and R. Madhu
Evaluation of FBMC Channel Estimation using multiple Auxiliary symbols for high throughput and low BER 5G and beyond communications 
Filter Bank Multi-Carrier (FBMC) modulation has been recognized as a consistent contender and a possible successor for Orthogonal Frequency Division Multiplexing (OFDM) in 5G and beyond because of its outstanding spectral properties. The channel is assessed in FBMC using pilot-symbol aided channel estimation that provides robust estimates even for severe channel conditions. In the present work, neutraliz ing the imaginary interference at the pilot positions is focused while estimating the channel. To neutralize the imaginary in terference, multiple auxiliary symbols have been proposed to enhance the throughput and channel capacity. The Iterative Minimum Mean Squared Error (IMMSE) cancellation scheme has been proposed to reduce the interference at the pilot and data positions. Transmission power, Bit Error Rate (BER) and throughput are computed for Filter Bank Multi Carrier (FBMC), OFDM and proven that better system performance is obtained for FBMC. The performance of channel estimation is evaluated through 5G standards and indicates that the usage of multiple auxiliary symbols per pilot leads to better throughput and low Bit Error Rate at low power transmission.

Reference
DOI: 10.36244/ICJ.2024.2.4
Download 

 

Praveen Kumar R., M. P. Prabakaran, Durai Arumugam, and J. Selvakumar 
A Novel Hybridization of ML Algorithms for Cluster Head Selection in WSN 
Generally, Wireless Sensor Networks (WSNs) are infrastructure-less networks with thousands of sensor nodes that sense or monitor the physical and environmental changes and forward the collected data to a central node. Besides, WSN has become the most efficient technology for handling Internet of Things (IoT) devices. Still, challenges such as node failures, high traffic among the nodes, link failures, etc., limit the performance of WSNs. To solve the challenges in WSN, this paper aims to develop a novel non-uniform clustering model, where the Cluster Heads (CHs) are selected based on the candidate CH selection strategy that transfers the data. Moreover, unbalanced energy utilization and data redundancy are eliminated via multi-hop communication. For attaining the non-uniform clustering model, the routing among the data packets is done by the efficiency of the hybridization of the Machin Learning (ML) algorithms viz Genetic Algorithm (GA) and Lion Algorithm (LA) with the consideration of energy, cost, time, network lifetime, and data accuracy. Finally, the performance of the proposed model is verified and validated through a comparative study with the existing models.

Reference
DOI: 10.36244/ICJ.2024.2.5
Download 

 

 Bence Csóka, Péter Fiala, and Péter Rucz 
 Direction and Distance Estimation of Sound  Sources with Microphone Arrays   
This paper is concerned with the estimation of the direction and distance of sound sources with the MUSIC beamforming algorithm, and their tracking with the help of Kalman filter. Direction-of-arrival (DOA) estimations can be performed using a combination of acoustical focusing and beamforming. Distance estimation is usually not part of the process, but it is possible through an extension of the beamforming algorithm. MUSIC (Multiple Signal Classification) is a relatively fast and simple method to locate sound sources. It is based on the separation of the received signals’ cross-spectral matrix to signal and noise subspaces. We also use the Kalman filter and its extended non-linear version to track moving sound sources. We evaluate the performance of these methods through simulations in the MATLAB environment and measurements with unmanned aerial vehicles (UAV). DOA estimations and tracking are possible in both cases, but distance estimation is significantly more problematic in the latter. We aim to find the cause of the errors in the estimation during measurements, to develop a more robust method in the future.

Reference
DOI: 10.36244/ICJ.2024.2.6
Download 

 

György Wersényi, and Ádám Csapó
 Comparison of Auditory and Visual Short-Term Memory Capabilities using a Serious Game Application  
A comprehensive serious game application has been designed and implemented to examine the capacity and effective- ness of short-term auditory and visual memory, otherwise known as working memory in human subjects. Participants engaged in an adaptation of the well-known paired association game that entails turning over cards and recalling their placement within a 2D matrix structure of various resolutions. Each trial introduced either visual icons (vision only condition) or auditory objects (audio-only condition). User performance was evaluated through a detailed statistical analysis focusing only on the highest 6x8 resolution condition in the application. Findings suggest that visual memory did not conclusively outperform auditory memory in the context of this game. However, within the scope of auditory stimuli, familiar iconic sounds, such as excerpts of speech and commonplace sounds, were recalled more effectively than unfamiliar, synthetic sounds like parametric waveforms. Furthermore, performance appeared to be influenced by demographic factors, with male and younger subjects yielding superior results.

Reference
DOI: 10.36244/ICJ.2024.2.7
Download