JITA - APEIRON

Pattern Proposal for detecting object occupancy in AVs utilizing ML, Edge Cloud, AI, and CV for parking LoTs

Mon, 27 Jan 2025 00:00:00 +0100

The purpose of this article is to provide an overview of current parking system management issues as well as support for effective object identification in circumstances of autonomous parking. In particular, we will look at existing solutions and patterns that have the most immediate use in automobiles and other businesses that rely on these technical approaches. The question arises as to the effectiveness and expense of existing solutions during various times of day (rain, fog, snow) and weather conditions, which can significantly alter the visibility of any systems that use cameras for object detection. This is applicable to parking lots that use cameras and a computer system’s integrated capabilities to detect and classify items using a machine model that makes certain decisions or suggests certain objects for object occupancy tracking. Our parking system concept proposal can be summed up in a dozen successful implementations for autonomous vehicle (AV) control and other Internet of things (IoT) systems. Image processing could be carried out using a mix of one or more algorithms and patterns that require specific hardware and software in the background, such as CI/CD, HPC, edge, cloud computing, and cluster microservices. Key contributions and findings can be given in the form of a novel pattern method known as combined detection of parking occupancy on chip (ComDPOCh), which can be applied in a real-world setting using appropriate hardware and software.

Transforming Traffic Safety: Detection of Car-Pedestrian Contact Using Computer Vision Technologies

Vesna Radojčić, Miloš Dobrojević — Mon, 27 Jan 2025 00:00:00 +0100

This paper explores the integration of computer vision technologies to enhance traffic safety through the effective detection of car-pedestrian interactions. As urban environments become more congested, pedestrian safety remains a critical concern. The system’s performance was evaluated using real-life footage from vehicle-mounted cameras, as well as images and videos sourced from online platforms. These real-world scenarios enabled a detailed assessment of the system’s accuracy and efficiency in practical conditions. The study highlights the potential for significant improvements in traffic safety, particularly in Bosnia and Herzegovina, where over 38% of registered vehicles are older than 23 years, and nearly 62% exceed 14 years. The aging vehicle fleet heightens the risk of accidents, underscoring the need for advanced detection methods. The proposed system automates the identification of hazardous situations on roads, allowing timely responses from relevant authorities.

Development of an Open-Source Voice-Controlled Smart Home System

Olja Krčadinac, Željko Stanković, Dragana Dudić, Lazar Stošić — Mon, 27 Jan 2025 00:00:00 +0100

This paper explores the design and implementation of a voice-controlled smart home system utilizing the Raspberry Pi platform and Node.js framework. The system aims to provide users with an intuitive method for managing household devices through voice commands while enhancing functionality with integrated sensors. Leveraging the Hidden Markov Model Toolkit (HTK) for speech recognition, the system accurately interprets user commands, facilitating control over lighting, temperature, and various IoT devices. In addition to voice activation, the system incorporates multiple sensors: a temperature sensor for monitoring ambient conditions, a motion sensor (PIR) for detecting occupancy, and a light sensor for assessing natural light levels. These sensors work in harmony with the voice control features, enabling automated responses such as adjusting the heating or cooling based on temperature readings, activating lights upon detecting movement, and regulating artificial lighting according to available daylight. Comprehensive testing demonstrated the system’s high accuracy in command recognition and its responsiveness to user inputs, as well as its robust integration with additional smart devices. User feedback was instrumental in refining the system, leading to improvements in command clarity and operational efficiency. This research highlights the potential of combining voice control and sensor technology to create a more responsive and user-friendly smart home environment.

Exploring the Impact of Technology on Human Interaction and Engaging Business Needs Through Software Design Patterns

Mohammad Daud Haiderzai, Ihsan Ul Haq Safi, Pavle Dakić — Mon, 27 Jan 2025 00:00:00 +0100

This paper explores the impact of technology on human interactions by integrating software engineering patterns to address business needs. A specialized development environment is proposed to enhance collaboration, focusing on bridging the technical and business perspectives through adaptable patterns. By leveraging classic and emergent software patterns, the approach supports real-time feedback and iterative improvements, enabling agile development practices that align with evolving user and business requirements. Key contributions include a modular framework for improved team communication and a Conceptual Framework on Human Interaction (CFI), along with an Embedded Business and Safety Human-AI Interaction Pattern (EBSH-AI). These tools aim to enhance interaction across technical and business domains while addressing ethical considerations in AI, such as privacy and authenticity in lifelike avatars. The findings indicate that structured software patterns improve cross-functional communication, fostering productivity and trust in AI-driven environments. This research provides a pathway to more meaningful interactions and user-centered outcomes in software engineering, with particular emphasis on the emerging role of AI avatars in the Metaverse and their impact on business and social engagements.

Linear Programming Routing for Wireless Body Area Networks – Loban Algorithm

Goran Đukanović, Goran Popović — Mon, 27 Jan 2025 00:00:00 +0100

IoMT systems are one of the most important segments of the future global IoT concept, which includes complete networking and data exchange related to all aspects of human life and activity. As a physical layer of these systems, WBAN is used as a network of wireless sensor nodes placed on certain parts of the human body with the aim of collecting data that is relevant for monitoring the health status of the monitored patient. Data transmission in networks of this type is specific in many respects compared to classic WSN networks, so even when solving problems typical for wireless sensor networks, such as data routing, it is necessary to apply specific solutions. Energy consumption of sensor nodes is one of the basic goals that are set by designers of wireless sensor networks. Short-range technologies implemented on networks that cover smaller areas of the terrain have additional specificities that require special attention. In this paper, the influence of different criteria on problem analysis is observed, as well as the application of several methods characteristic for optimization through linear programming. We propose an original algorithm for routing data through the network called LOBAN, which takes into account two criteria when choosing a relay node, the importance of which is emphasized through different combinations of coefficients in linear programming. The goal of the algorithm is to optimize energy consumption in the network and extend its lifetime.

Full Linearization of Ranking and Unranking of Catalan Cipher Vectors Using Catalan Triangle Abstraction

Adrijan Božinovski, Biljana Stojčevska — Mon, 27 Jan 2025 00:00:00 +0100

This paper demonstrates how to abstract a version of the Catalan Triangle necessary to compute the rank value from a given Catalan Cipher Vector, which is a process called ranking, and the process of obtaining a Catalan Cipher Vector from a given rank value, which is a process called unranking. That version of the Catalan Triangle is called the Bottom Ones Catalan Triangle and is not required to be computed in its entirety for the purpose of ranking and unranking, but only elements of it that are required for the current computation. A formula is derived to compute each element of this triangle and it is demonstrated how the processes of both ranking and unranking are fully linear.

On Underwater Data Centers: Surveillance, Monitoring, and Environmental Management in the Baltic Sea

Mark Abner, Sanja Bauk — Mon, 27 Jan 2025 00:00:00 +0100

This article analyzes underwater surveillance and monitoring technologies aimed at enhancing security and environmental management, using a hypothetical underwater data center in the Baltic Sea as a case study. It explores cutting-edge solutions such as remotely operated vehicles (ROVs), autonomous underwater vehicles (AUVs), and smart buoys, focusing on their integration for monitoring underwater infrastructure and safeguarding against infrastructural threats. With rising concerns over maritime security due to recent events like the Nord Stream outage, this research highlights the need for advanced technological systems to address such a kind of vulnerabilities. The analysis also considers multi-layered potential of these systems for security, safety, and environmental resilience. Consequently, this study provides insights into the feasibility, challenges, and future directions for deploying underwater data centers as a sustainable alternative to traditional land-based facilities of this type, contributing to the broader discourse on securing critical underwater infrastructure and promoting eco-friendly data storage solutions.

New neural PLL Architecture

Vladimir V. Đokić, Dragana Đokić — Mon, 27 Jan 2025 00:00:00 +0100

A PLL or phase-locked loop is a control system that creates an output signal whose phase is related to the phase-locked loop and represents controlled input signal. The goal of this research is to first investigate the functioning of new PLL neural networks and then, in the research section, explore an approach involving the extraction of neural symmetrical voltage components.
The architectural characteristics of phase-locked loops (PLLs) typically include capture and lock ranges, bandwidth, and transient response. The new neural PLL architecture offers several advantages, such as low noise performance, reduced silicon area, and compatibility with low supply voltages. However, it may also present disadvantages, including hardware dependency and potential design complexity compared to traditional PLL architectures. Evaluating these factors is crucial, depending on the specific needs of the application.
In this paper, we present the scientific research included in the experimental part where we investigate the performance of the proposed neural PLL, for which experimental comparisons with the conventional PLL in a distorted reference frame are necessary. Structural columns or structural circles will be used for graphic display.
The following research methods and techniques will be applied: instruments, basic methods and data processing procedures - if they are foreseen. What makes this work a scientific research work is a descriptive method that will be used.
To better understand how PLLs work, we propose an original three-phase neural approach for components of the system’s phase and symmetry. The quality of the electricity can be assessed and managed with this framework. Our study shows that the full neural architecture may be applied to three-phase power systems because it is based on DSP supplies. Additionally, we present the performance of the PLL system in a three-phase power supply context. Different regulators, such as PI and RST based on phase logic, are incorporated into the PLL scheme. The results suggest that the neural PLL could make a significant contribution in applications where the quality and efficiency of three-phase power systems are essential.

Improving Trust in Information Sources: Serbian Wikipedia Use Case

Dragana Dudić, Nebojša Ratković, Ivana Firaunović, Olja Krčadinac — Mon, 27 Jan 2025 00:00:00 +0100

For almost two decades, Wikimedia Serbia has active collaboration with educational institutions in Serbia, promoting Wikipedia to young people, training and encouraging them to create and edit Wikipedia with special focus on Wikipedia trust concepts as the important factor of every information source. This paper presents our findings on the current level of students’ understanding of trust mechanisms in information sources and examines the impact of conducted targeted educational intervention on this domain. This study was conducted as part of the collaboration between Wikimedia Serbia, the Faculty of Computer Science and Informatics at Union Nikola Tesla University, and Šabac Grammar School. We used structured questionnaires to determine the initial state of trust mechanisms awareness in students. After analyzing the obtained results, we organized instructional session which included comprehensive lectures on Wikipedia, its structure, rules and its editing process. Afterwords, students selected Wikipedia-related assignments aligned with their course interests. The follow-up questionnaire revealed great improvement in correct referencing and licensing usage, in both educational institutions emphasizing the importance of proper referencing and licensing as trust mechanism and its influence in digital literacy in general.

Revitalisation of Old Hydraulic Press Brake Machines Through Novel Software Solutions

Daniel Menićanin, Jelena Radanović, Nikola Račić, Branimir Trenkić — Mon, 27 Jan 2025 00:00:00 +0100

In the era of Industry 4.0, the modernization of traditional machines has become crucial for maintaining competitiveness and enhancing production capacities. Hydraulic CNC press machines manufactured in the Balkans, particularly in the former Yugoslavia, renowned for their robustness and reliability, are increasingly falling behind in terms of the technological innovations required to meet modern production demands. Rather than being replaced by expensive new machines, this paper presents a transformative approach to the digital upgrade of existing systems. By employing advanced software solutions and novel control methods, older presses are converted into high-precision, intelligently managed machines capable of automatically adjusting operating parameters according to material and task specifications. This system not only significantly reduces operational costs but also extends the lifespan of these machines, increases production speed and minimizes errors. The core innovation of the system is its adaptability, enabling continuous evolution and seamless integration of new technological advancements without requiring substantial investments in entirely new equipment. This paper offers a vision of the future where digital transformation allows traditional machines to become part of smart, interconnected production lines.

Young Learners’ Use of Google Translate and Glosbe in a Writing Task: A Case Study

Mirela Simić — Mon, 27 Jan 2025 00:00:00 +0100

The aim of this study was to examine the use of digital translation tools by near-beginner adolescent foreign language learners and their beliefs about the potential of these tools for language learning. A group of 15 participants, aged 13 and 14, enrolled in a German language course at a language studio, completed a writing task using the Google Translate and Glosbe tools. Semi-structured interviews provided insight into their usage of these tools and revealed contrasting beliefs. Some learners, in favor of using digital translation tools in language classes, argued that the tools facilitated communication and helped them acquire new vocabulary, conjugation patterns, and syntax. In contrast, others expressed that they had learned little, were critical of translating entire texts, and were concerned about becoming dependent on the tools and developing a false sense of competence. The lack of reflexivity and the challenges faced by some students underscore the importance of providing guidance and support to learners at this stage when using digital translation tools.