Andriy Fechan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Andriy Fechan has a total of 126 co-authors. Here are the top co-authors by number of shared publications:
Lviv Polytechnic National University
8 shared publications- Ihor Rabiichuk
Lviv Polytechnic National University
6 shared publications - Vladyslav Cherpak
Lviv Polytechnic National University, 79013 Lviv, Ukraine
6 shared publications O. Yasynovska
Lviv Polytechnic National University , Lviv , Ukraine
5 shared publicationsV. Kotsun
Lviv Affiliate of European University, Lviv, Ukraine
5 shared publications- Zinoviy Mykytyuk
Lviv Polytechnic National University , Lviv , Ukraine
5 shared publications
Latest Publications
- 25 June 2026
- Published by Ukrainian National Forestry University in Scientific Bulletin of UNFU
- Vol. 36 (3)
SEIR-LT hybrid asynchronous cellular automata model for information diffusion in social networks
A hybrid asynchronous cellular automata model is proposed for simulating information diffusion processes in social networks, combining the functionality of classical compartmental structures (Susceptible-Exposed-Infected-Recovered) with linear threshold theory. An individualized mechanism for updating user states is implemented to more accurately capture temporal heterogeneity: each network participant is assigned unique activity intervals instead of simultaneous state transitions across the entire population. It is demonstrated that such an asynchronous modeling approach provides a more detailed representation of real human behavior, as individuals respond to external stimuli with varying speeds and levels of attention. A generalized linear threshold (GLT) methodology is introduced to determine the probability of a user transitioning to an active state – that is, the moment when social influence from the surrounding network exceeds the individual resistance threshold, prompting information adoption. The development patterns of a heterogeneous population are characterized, with specific identification of influential and resistant classes of network participants. It is established that abandoning global synchronization of users completely eliminates unrealistic wave-like artifacts inherent in traditional computational approaches. The functionality of the proposed asynchronous cellular automata architecture is tested using Monte Carlo methodology on a two-dimensional lattice. The impact of temporal heterogeneity on the average cascade size is evaluated, revealing an increase from 5.93 % under full synchronization to 6.17 % in localized heterogeneous configurations. A significant increase in the overall probability of large-scale information diffusion is observed: the proposed mechanism achieves successful propagation in 96 out of 100 macroscopic iterations, compared to only 88 % under conventional constraints. A natural smoothing of process intensity is recorded, confirmed by a reduction in peak prevalence (the maximum proportion of the population in the infected state actively disseminating information) from 0.99 % to 0.92 %, along with a more efficient redistribution of load over a slightly extended sequence of simulation stages. The influence of model parameters is analyzed, and a phase transition of the baseline probability is identified, with a critical point near 0.25. A fundamental asymmetry in information spread within the network is revealed: initiating large-scale cascades using a small number of influential users is significantly easier than suppressing them by increasing the number of resistant participants. The developed asynchronous cellular automata model is proposed as a tool for forecasting and testing strategies aimed at mitigating the spread of undesirable content.
- 25 June 2026
- Published by Ukrainian National Forestry University in Scientific Bulletin of UNFU
- Vol. 36 (3)
SEIR-LT hybrid asynchronous cellular automata model for information diffusion in social networks
A hybrid asynchronous cellular automata model is proposed for simulating information diffusion processes in social networks, combining the functionality of classical compartmental structures (Susceptible-Exposed-Infected-Recovered) with linear threshold theory. An individualized mechanism for updating user states is implemented to more accurately capture temporal heterogeneity: each network participant is assigned unique activity intervals instead of simultaneous state transitions across the entire population. It is demonstrated that such an asynchronous modeling approach provides a more detailed representation of real human behavior, as individuals respond to external stimuli with varying speeds and levels of attention. A generalized linear threshold (GLT) methodology is introduced to determine the probability of a user transitioning to an active state – that is, the moment when social influence from the surrounding network exceeds the individual resistance threshold, prompting information adoption. The development patterns of a heterogeneous population are characterized, with specific identification of influential and resistant classes of network participants. It is established that abandoning global synchronization of users completely eliminates unrealistic wave-like artifacts inherent in traditional computational approaches. The functionality of the proposed asynchronous cellular automata architecture is tested using Monte Carlo methodology on a two-dimensional lattice. The impact of temporal heterogeneity on the average cascade size is evaluated, revealing an increase from 5.93 % under full synchronization to 6.17 % in localized heterogeneous configurations. A significant increase in the overall probability of large-scale information diffusion is observed: the proposed mechanism achieves successful propagation in 96 out of 100 macroscopic iterations, compared to only 88 % under conventional constraints. A natural smoothing of process intensity is recorded, confirmed by a reduction in peak prevalence (the maximum proportion of the population in the infected state actively disseminating information) from 0.99 % to 0.92 %, along with a more efficient redistribution of load over a slightly extended sequence of simulation stages. The influence of model parameters is analyzed, and a phase transition of the baseline probability is identified, with a critical point near 0.25. A fundamental asymmetry in information spread within the network is revealed: initiating large-scale cascades using a small number of influential users is significantly easier than suppressing them by increasing the number of resistant participants. The developed asynchronous cellular automata model is proposed as a tool for forecasting and testing strategies aimed at mitigating the spread of undesirable content.
- Open Access
- 17 May 2026
- Published by Lviv Polytechnic National University in Computer systems and network
- Vol. 19, 271-292
Processes of Modelling Political Polarization
Political polarization has become a defining structural feature of contemporary democratic societies, driven by complex interactions between individual bias, localized social influence, and broader informational environments. Understanding the mechanisms through which heterogeneous opinions evolve into structured ideological divisions requires formal computational modelling capable of integrating spatial, behavioural, and temporal components. This study proposes a modelling framework based on an asynchronous cellular automaton in which agents are embedded in a continuous two-dimensional ideological space representing economic and social attitudes. Each agent’s political state is defined as a vector, enabling simultaneous representation of multiple ideological dimensions rather than reducing political orientation to a single scalar indicator. Opinion evolution is governed by three principal mechanisms: persistence of prior beliefs, local social influence within a Moore neighbourhood, and directional tendencies toward ideological extremity. Agent heterogeneity is introduced through behavioural parameters such as stubbornness, charisma, and differential sensitivity to internal and external factors. To introduce gradualism into the dynamics, the model incorporates momentum-based state updates and inertia parameters inspired by swarm intelligence principles, without assuming explicit optimization or collective objective-driven behaviour. In this context, swarm intelligence is understood not in its classical sense, but as a source of structural principles for modelling behaviour and individual opinion trajectories. In addition to local peer interactions, global time-varying exogenous ideological signal acts uniformly across the population, representing diffuse environmental pressures such as media narratives or macro-political events. The emergence and evolution of polarization are assessed using both qualitative and quantitative approaches. A visualization procedure based on transformation from Cartesian to polar coordinates and mapping to the HSV colour space provides an interpretable representation of ideological direction and intensity on the grid. Quantitative assessment relies on a complementary set of polarization metrics, including Moran’s I for spatial autocorrelation, the Esteban-Ray polarization index, measures of inter-group separation derived from clustering, circular variance, entropy, extremity indicators, and inter-dimensional correlation coefficients. The combined application of these metrics enables differentiation between polarization, mere diversity, consensus formation, and radicalization. Simulation results demonstrate that, starting from an initially heterogeneous and spatially disordered distribution of opinions, structured ideological clusters emerge progressively through local interactions and reinforcement mechanisms. Over time, fragmented clusters coalesce into large-scale contiguous regions characterized by internal homogeneity and external separation. Importantly, polarization stabilizes without universal radicalization, indicating that durable ideological division may arise endogenously from decentralized interactions and heterogeneous behavioural traits. Across five independent runs of the primary configuration, key polarization metrics, including the Moran's I and the Esteban-Ray polarization index, confirm the reproducibility and structural robustness of this outcome. Simulations under alternative parameter configurations further demonstrate that the framework is capable of reproducing qualitatively and quantitatively distinct dynamics, including radicalization and ideological fragmentation. The proposed framework provides a flexible and interpretable computational platform for analysing polarization as a multiscale dynamic process shaped by both local and global influences.
- 17 May 2026
- Published by Lviv Polytechnic National University in Computer systems and network
- Vol. 19, 271-292
Open AccessProcesses of Modelling Political Polarization
Political polarization has become a defining structural feature of contemporary democratic societies, driven by complex interactions between individual bias, localized social influence, and broader informational environments. Understanding the mechanisms through which heterogeneous opinions evolve into structured ideological divisions requires formal computational modelling capable of integrating spatial, behavioural, and temporal components. This study proposes a modelling framework based on an asynchronous cellular automaton in which agents are embedded in a continuous two-dimensional ideological space representing economic and social attitudes. Each agent’s political state is defined as a vector, enabling simultaneous representation of multiple ideological dimensions rather than reducing political orientation to a single scalar indicator. Opinion evolution is governed by three principal mechanisms: persistence of prior beliefs, local social influence within a Moore neighbourhood, and directional tendencies toward ideological extremity. Agent heterogeneity is introduced through behavioural parameters such as stubbornness, charisma, and differential sensitivity to internal and external factors. To introduce gradualism into the dynamics, the model incorporates momentum-based state updates and inertia parameters inspired by swarm intelligence principles, without assuming explicit optimization or collective objective-driven behaviour. In this context, swarm intelligence is understood not in its classical sense, but as a source of structural principles for modelling behaviour and individual opinion trajectories. In addition to local peer interactions, global time-varying exogenous ideological signal acts uniformly across the population, representing diffuse environmental pressures such as media narratives or macro-political events. The emergence and evolution of polarization are assessed using both qualitative and quantitative approaches. A visualization procedure based on transformation from Cartesian to polar coordinates and mapping to the HSV colour space provides an interpretable representation of ideological direction and intensity on the grid. Quantitative assessment relies on a complementary set of polarization metrics, including Moran’s I for spatial autocorrelation, the Esteban-Ray polarization index, measures of inter-group separation derived from clustering, circular variance, entropy, extremity indicators, and inter-dimensional correlation coefficients. The combined application of these metrics enables differentiation between polarization, mere diversity, consensus formation, and radicalization. Simulation results demonstrate that, starting from an initially heterogeneous and spatially disordered distribution of opinions, structured ideological clusters emerge progressively through local interactions and reinforcement mechanisms. Over time, fragmented clusters coalesce into large-scale contiguous regions characterized by internal homogeneity and external separation. Importantly, polarization stabilizes without universal radicalization, indicating that durable ideological division may arise endogenously from decentralized interactions and heterogeneous behavioural traits. Across five independent runs of the primary configuration, key polarization metrics, including the Moran's I and the Esteban-Ray polarization index, confirm the reproducibility and structural robustness of this outcome. Simulations under alternative parameter configurations further demonstrate that the framework is capable of reproducing qualitatively and quantitatively distinct dynamics, including radicalization and ideological fragmentation. The proposed framework provides a flexible and interpretable computational platform for analysing polarization as a multiscale dynamic process shaped by both local and global influences.
- Open Access
- 1 January 2026
- Published by Lviv Polytechnic National University in Ukrainian Journal of Information Technology
- Vol. 8 (1) , 105-114
ADAPTIVE MULTI-TARGET RECONNAISSANCE BY UAV SWARMS IN EW ENVIRONMENTS: A HYBRID AUCTION-PSO APPROACH
The rapid proliferation of Unmanned Aerial Vehicle (UAV) swarms in modern combat theaters has necessitated the development of robust autonomous control systems capable of operating effectively in environments heavily saturated with Electronic Warfare (EW) assets. Traditional navigation paradigms frequently fail under such conditions due to global navigation satellite system spoofing and the systematic disruption of remote operator control channels. This paper proposes a novel hybrid control architecture for multi-target reconnaissance that synergistically integrates a market-based Auction Algorithm for strategic target allocation and a modified Particle Swarm Optimization (PSO) algorithm for tactical path planning. The integration of these distinct logical levels utilizes a Finite State Machine (FSM) mechanism, which ensures strict deterministic switching between cruising, pursuit, and evasive flight modes during the mission. The strategic allocation phase employs a complex non-linear cost function designed to evaluate multiple parameters simultaneously. This function accounts not only for the geometric distance to the target but also for the current energy states of individual agents and the predicted environmental threat levels calculated through discrete spatial sampling. Furthermore, the tactical navigational logic incorporates advanced bio-inspired mechanisms to enhance survivability. The modified algorithm introduces adaptive inertia weights and tangential repulsive forces, which collectively empower the drones to autonomously execute sophisticated "sliding mode" maneuvers directly along the boundaries of active jamming zones. The research methodology included a rigorous comparative analysis using Monte Carlo simulations within a comprehensive 60×60 kilometer operational area. This simulated environment featured highly rugged terrain and modeled a swarm of 20 UAVs tasked with engaging 100 distant targets while navigating through 15 dynamically active EW suppression zones. The simulation results conclusively demonstrate that the proposed hybrid approach significantly outperforms classical greedy algorithms. Specifically, the hybrid system reduced the integral damage sustained from EW irradiation by 18.4% and decreased the total time spent within high-risk danger zones by 12.2%. Moreover, the analysis revealed a particularly significant "energy paradox" during the autonomous operations. Despite a 3.4% increase in the average flight distance caused by the necessary evasive maneuvers, the swarm achieved superior overall energy efficiency by consuming 10.79% of battery resources compared to the 10.88% consumed by straight-line trajectories. Ultimately, these findings prove the high effectiveness of the proposed hybrid method for executing autonomous reconnaissance missions under conditions of extreme uncertainty. The developed control architecture offers a highly viable and practical pathway for the future deployment of resilient unmanned aerial systems in actively contested airspace.
- 1 January 2026
- Published by Lviv Polytechnic National University in Ukrainian Journal of Information Technology
- Vol. 8 (1) , 105-114
Open AccessADAPTIVE MULTI-TARGET RECONNAISSANCE BY UAV SWARMS IN EW ENVIRONMENTS: A HYBRID AUCTION-PSO APPROACH
The rapid proliferation of Unmanned Aerial Vehicle (UAV) swarms in modern combat theaters has necessitated the development of robust autonomous control systems capable of operating effectively in environments heavily saturated with Electronic Warfare (EW) assets. Traditional navigation paradigms frequently fail under such conditions due to global navigation satellite system spoofing and the systematic disruption of remote operator control channels. This paper proposes a novel hybrid control architecture for multi-target reconnaissance that synergistically integrates a market-based Auction Algorithm for strategic target allocation and a modified Particle Swarm Optimization (PSO) algorithm for tactical path planning. The integration of these distinct logical levels utilizes a Finite State Machine (FSM) mechanism, which ensures strict deterministic switching between cruising, pursuit, and evasive flight modes during the mission. The strategic allocation phase employs a complex non-linear cost function designed to evaluate multiple parameters simultaneously. This function accounts not only for the geometric distance to the target but also for the current energy states of individual agents and the predicted environmental threat levels calculated through discrete spatial sampling. Furthermore, the tactical navigational logic incorporates advanced bio-inspired mechanisms to enhance survivability. The modified algorithm introduces adaptive inertia weights and tangential repulsive forces, which collectively empower the drones to autonomously execute sophisticated "sliding mode" maneuvers directly along the boundaries of active jamming zones. The research methodology included a rigorous comparative analysis using Monte Carlo simulations within a comprehensive 60×60 kilometer operational area. This simulated environment featured highly rugged terrain and modeled a swarm of 20 UAVs tasked with engaging 100 distant targets while navigating through 15 dynamically active EW suppression zones. The simulation results conclusively demonstrate that the proposed hybrid approach significantly outperforms classical greedy algorithms. Specifically, the hybrid system reduced the integral damage sustained from EW irradiation by 18.4% and decreased the total time spent within high-risk danger zones by 12.2%. Moreover, the analysis revealed a particularly significant "energy paradox" during the autonomous operations. Despite a 3.4% increase in the average flight distance caused by the necessary evasive maneuvers, the swarm achieved superior overall energy efficiency by consuming 10.79% of battery resources compared to the 10.88% consumed by straight-line trajectories. Ultimately, these findings prove the high effectiveness of the proposed hybrid method for executing autonomous reconnaissance missions under conditions of extreme uncertainty. The developed control architecture offers a highly viable and practical pathway for the future deployment of resilient unmanned aerial systems in actively contested airspace.
- Open Access
- 29 December 2025
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 6 (2(86)) , 15-22
Development of adaptive reconfiguration method for stream data processing systems using system metrics
The object of research is the process of adaptive configuration changes for stream processing applications which is focused on improving specific performance properties. The absence of the generalized automated approach for dynamic reconfiguration of state-store in limited hardware environment is the research problem addressed in this paper. The proposed solution helps to avoid a need for manual application reconfiguration from engineers. The implementation is based on Kafka Streams but designed to be portable across other frameworks that use RocksDB as a state store. Static configuration of modern stream processing systems limits efficiency under variable workloads. In this study, an adaptive module is proposed that monitors system metrics in real-time and automatically updates state-store configurations. The module performs deterministic check to derive new configuration based on predefined thresholds or utilizes a fine-tuned Large Language Model (LLM) to select new configuration values when decisions are vague. The method dynamically applies updates to the affected instance. High-load experimental results reveal the fact that adaptive executions eliminated write stalls, increased memtable hit ratio from 2% to 40% and block-cache hit ratio from 15% to 80%, reduced disk I/O by approximately 50%, and improved throughput by around 5%, at the cost of higher memory consumption. To avoid redundant adaptive updates and outlier-based bias a 10-minute observation frequency was selected. The approach is suitable for systems with fixed resources, state-intensive workloads with high key cardinality. Additionally, if covers the need for safe configuration change under operational constraints. The architecture is framework agnostic for the RocksDB-based based stream processing with state stores.
- 29 December 2025
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 6 (2(86)) , 15-22
Open AccessDevelopment of adaptive reconfiguration method for stream data processing systems using system metrics
The object of research is the process of adaptive configuration changes for stream processing applications which is focused on improving specific performance properties. The absence of the generalized automated approach for dynamic reconfiguration of state-store in limited hardware environment is the research problem addressed in this paper. The proposed solution helps to avoid a need for manual application reconfiguration from engineers. The implementation is based on Kafka Streams but designed to be portable across other frameworks that use RocksDB as a state store. Static configuration of modern stream processing systems limits efficiency under variable workloads. In this study, an adaptive module is proposed that monitors system metrics in real-time and automatically updates state-store configurations. The module performs deterministic check to derive new configuration based on predefined thresholds or utilizes a fine-tuned Large Language Model (LLM) to select new configuration values when decisions are vague. The method dynamically applies updates to the affected instance. High-load experimental results reveal the fact that adaptive executions eliminated write stalls, increased memtable hit ratio from 2% to 40% and block-cache hit ratio from 15% to 80%, reduced disk I/O by approximately 50%, and improved throughput by around 5%, at the cost of higher memory consumption. To avoid redundant adaptive updates and outlier-based bias a 10-minute observation frequency was selected. The approach is suitable for systems with fixed resources, state-intensive workloads with high key cardinality. Additionally, if covers the need for safe configuration change under operational constraints. The architecture is framework agnostic for the RocksDB-based based stream processing with state stores.
- 20 October 2025
- Published by Lviv Polytechnic National University in Computational Problems of Electrical Engineering
- Vol. 15 (1) , 7-17
Use of swarm intelligence in unmanned vehicles
This article explores the use of swarm intelligence algorithms in unmanned vehicles (UVs), focuses on their main advantages for improving the efficiency and productivity of systems. Unmanned vehicles, which can operate autonomously or under remote control, play a significant role in such areas as surveillance, search and rescue, agriculture and military operations. The main focus of the article is on algorithms such as ant colony optimisation (ACO), artificial bee colony (ABC), particle swarm optimization (PSO), glow-worm swarm optimization (GSO), firefly algorithm (FA), bat algorithm (BA), grey wolf optimizer (GWO), and whale optimization algorithm (WOA). Each of these algorithms is discussed in detail, particularly their core principles, specific applications in UVs, and their levels of effectiveness in different environments. Each algorithm has been examined to highlight its operational strengths and its limitations, such as computational demands and environmental suitability. This paper discusses the algorithms in terms of managing critical functions of UVs, such as resource allocation and multi-agent coordination, which are essential for complex mission scenarios. Particular attention is paid to the adaptability of each algorithm, especially in unpredictable or hostile environments, where rapid recalibration of UV behaviour is necessary for mission success. By analysing each algorithm capacity to adjust the UV to new data in real-time, the article highlights their potential to optimize UV performance and reliability in challenging contexts. Special attention is given to collaborative task management in swarm intelligence, emphasizing its ability to enhance unmanned aerial vehicle (UAV) group coordination and decision-making for efficient operation in complex and dynamic scenarios. In general, the article provides deep analysis of swarm intelligence algorithms, and the information that will help choose the most effective algorithm to help solve specific tasks using different types of UVs. Future research will focus on improving the scalability, adaptability, and integration of these algorithms with latest technologies in order to enhance their effectiveness in solving complex UV missions. In addition, a comparative table of the main characteristics of the algorithms was created and a review of similar studies comparing swarm algorithms was made.
- 20 October 2025
- Published by Lviv Polytechnic National University in Computational Problems of Electrical Engineering
- Vol. 15 (1) , 7-17
Use of swarm intelligence in unmanned vehicles
This article explores the use of swarm intelligence algorithms in unmanned vehicles (UVs), focuses on their main advantages for improving the efficiency and productivity of systems. Unmanned vehicles, which can operate autonomously or under remote control, play a significant role in such areas as surveillance, search and rescue, agriculture and military operations. The main focus of the article is on algorithms such as ant colony optimisation (ACO), artificial bee colony (ABC), particle swarm optimization (PSO), glow-worm swarm optimization (GSO), firefly algorithm (FA), bat algorithm (BA), grey wolf optimizer (GWO), and whale optimization algorithm (WOA). Each of these algorithms is discussed in detail, particularly their core principles, specific applications in UVs, and their levels of effectiveness in different environments. Each algorithm has been examined to highlight its operational strengths and its limitations, such as computational demands and environmental suitability. This paper discusses the algorithms in terms of managing critical functions of UVs, such as resource allocation and multi-agent coordination, which are essential for complex mission scenarios. Particular attention is paid to the adaptability of each algorithm, especially in unpredictable or hostile environments, where rapid recalibration of UV behaviour is necessary for mission success. By analysing each algorithm capacity to adjust the UV to new data in real-time, the article highlights their potential to optimize UV performance and reliability in challenging contexts. Special attention is given to collaborative task management in swarm intelligence, emphasizing its ability to enhance unmanned aerial vehicle (UAV) group coordination and decision-making for efficient operation in complex and dynamic scenarios. In general, the article provides deep analysis of swarm intelligence algorithms, and the information that will help choose the most effective algorithm to help solve specific tasks using different types of UVs. Future research will focus on improving the scalability, adaptability, and integration of these algorithms with latest technologies in order to enhance their effectiveness in solving complex UV missions. In addition, a comparative table of the main characteristics of the algorithms was created and a review of similar studies comparing swarm algorithms was made.
- 20 October 2025
- Published by Lviv Polytechnic National University in Computational Problems of Electrical Engineering
- Vol. 15 (1) , 18-24
Gesture recognition system for controlling iot systems
The development of the Internet of Things (IoT) opens up new opportunities for creating intelligent services that enhance user interaction with surrounding devices. Modern IoT systems primarily use touchscreens and mobile applications for control; however, gesture-based methods can significantly expand their functionality. This work proposes a gesture recognition system applied to the control of IoT devices. The core of the system is the classification of finger movement trajectories using a Hidden Markov Model (HMM). The system consists of three main stages: initial hand segmentation using colour and depth information, fingertip detection based on hand contours, and the use of clustering in polar coordinates to extract dynamic features. The Baum-Welch and Viterbi algorithms are applied for training and gesture recognition, respectively. Experimental results show that the developed system is capable of classifying gestures with consideration of spatiotemporal variability with high accuracy. In particular, the average recognition rate reached 98.61% for the training set and 93.06% for the test data. The proposed approach demonstrates effectiveness under challenging conditions, including changes in lighting and partial occlusion of objects in the scene.
- 20 October 2025
- Published by Lviv Polytechnic National University in Computational Problems of Electrical Engineering
- Vol. 15 (1) , 18-24
Gesture recognition system for controlling iot systems
The development of the Internet of Things (IoT) opens up new opportunities for creating intelligent services that enhance user interaction with surrounding devices. Modern IoT systems primarily use touchscreens and mobile applications for control; however, gesture-based methods can significantly expand their functionality. This work proposes a gesture recognition system applied to the control of IoT devices. The core of the system is the classification of finger movement trajectories using a Hidden Markov Model (HMM). The system consists of three main stages: initial hand segmentation using colour and depth information, fingertip detection based on hand contours, and the use of clustering in polar coordinates to extract dynamic features. The Baum-Welch and Viterbi algorithms are applied for training and gesture recognition, respectively. Experimental results show that the developed system is capable of classifying gestures with consideration of spatiotemporal variability with high accuracy. In particular, the average recognition rate reached 98.61% for the training set and 93.06% for the test data. The proposed approach demonstrates effectiveness under challenging conditions, including changes in lighting and partial occlusion of objects in the scene.
- Open Access
- 29 August 2025
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 4 (2(84)) , 6-12
Development of a method for using color in machine-readable optical codes to increase the information capacity
The possibility of increasing the capacity of QR codes by using color modules without adding new metadata is studied. A method for automatically determining the number of colors and their palette using image processing is proposed, which ensures compatibility with classical QR codes. A system is proposed that allows creating a QR code that uses 4, 8 or 16 colors in addition to the standard black and white version. The main problem is the optimal use of the available color space to minimize errors when reading an informative image with an optical camera, compensate for the effects of uneven lighting and poor image quality, and ensure backward compatibility with the black and white version. In the course of analyzing the use of different color spaces, the most promising perceptually uniform OKLCH space was determined. Algorithms for image preprocessing for correct decoding of information and an algorithm for encoding and decoding information using color have been developed. The results obtained are explained by the distribution of the color gamut after the test reading of informative images, the number of errors and successful readings. Using the OKLCH color space, it was possible to read 60% of 16-color test images, while in HSL it was not possible to read any image due to color overlap. However, both spaces have a fairly high rate of successful reads in 4 and 8 color codes. The use of color will allow the introduction of new standards for high-capacity color machine-readable codes without requiring changes to existing ones for additional metadata, while maintaining full backward compatibility and reliability of black and white codes. Increased information capacity in some cases allows to eliminate the need to be connected to the Internet, reduce the size of the code, and make it more visually appealing by using colors.
- 29 August 2025
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 4 (2(84)) , 6-12
Open AccessDevelopment of a method for using color in machine-readable optical codes to increase the information capacity
The possibility of increasing the capacity of QR codes by using color modules without adding new metadata is studied. A method for automatically determining the number of colors and their palette using image processing is proposed, which ensures compatibility with classical QR codes. A system is proposed that allows creating a QR code that uses 4, 8 or 16 colors in addition to the standard black and white version. The main problem is the optimal use of the available color space to minimize errors when reading an informative image with an optical camera, compensate for the effects of uneven lighting and poor image quality, and ensure backward compatibility with the black and white version. In the course of analyzing the use of different color spaces, the most promising perceptually uniform OKLCH space was determined. Algorithms for image preprocessing for correct decoding of information and an algorithm for encoding and decoding information using color have been developed. The results obtained are explained by the distribution of the color gamut after the test reading of informative images, the number of errors and successful readings. Using the OKLCH color space, it was possible to read 60% of 16-color test images, while in HSL it was not possible to read any image due to color overlap. However, both spaces have a fairly high rate of successful reads in 4 and 8 color codes. The use of color will allow the introduction of new standards for high-capacity color machine-readable codes without requiring changes to existing ones for additional metadata, while maintaining full backward compatibility and reliability of black and white codes. Increased information capacity in some cases allows to eliminate the need to be connected to the Internet, reduce the size of the code, and make it more visually appealing by using colors.
- Open Access
- 1 June 2025
- Published by Lviv Polytechnic National University in Computer systems and network
- Vol. 17, 17-33
Software Code Protection on the .NET Platform
In the work, a study of the functioning of complex information systems is carried out, which is aimed at the analysis of existing protection methods, the development of the concept of the code transformation mechanism, which will ensure a high level of protection of .NET applications. The architecture of the .NET Framework with modern universal improvements is considered. One of the important elements is the Base Class Library (BCL), which is a set of base classes and libraries that provide standard functionality, such as working with files, networks, databases, text processing, and basic data structures. It was determined that obfuscation is one of the most common methods of software protection, which consists in modifying the source or intermediate code in such a way as to preserve its functionality, but make it difficult to understand. The main features of code improvement are established, which consist in obfuscating execution flows, encrypting string literals, and using packers and crypters. A characteristic feature of packers and crypters is direct injection into memory, which reflects its functionality in unpacking or decrypting program code directly into random access memory (RAM), which makes it difficult to remove a dump of the process. In addition, the entanglement of execution threads allows you to modify the logic of the software operation by inserting unnecessary conditions, loops or transitions, which creates complex execution structures. The anti-debugging process is proposed, which is one of the key components of software protection against reverse engineering. A conceptual approach to the implementation of software anti-debugging mechanisms is formed, which consists in multi-level protection of software and changes the instructions and behavior of the virtual machine. One of the important elements of such algorithms is checking parent processes and detecting debugging using hardware breakpoints. This, in its turn, opens up the possibility during the design of a virtual machine in combination with other techniques to create significant obstacles for the analysis of the software code.
- 1 June 2025
- Published by Lviv Polytechnic National University in Computer systems and network
- Vol. 17, 17-33
Open AccessSoftware Code Protection on the .NET Platform
In the work, a study of the functioning of complex information systems is carried out, which is aimed at the analysis of existing protection methods, the development of the concept of the code transformation mechanism, which will ensure a high level of protection of .NET applications. The architecture of the .NET Framework with modern universal improvements is considered. One of the important elements is the Base Class Library (BCL), which is a set of base classes and libraries that provide standard functionality, such as working with files, networks, databases, text processing, and basic data structures. It was determined that obfuscation is one of the most common methods of software protection, which consists in modifying the source or intermediate code in such a way as to preserve its functionality, but make it difficult to understand. The main features of code improvement are established, which consist in obfuscating execution flows, encrypting string literals, and using packers and crypters. A characteristic feature of packers and crypters is direct injection into memory, which reflects its functionality in unpacking or decrypting program code directly into random access memory (RAM), which makes it difficult to remove a dump of the process. In addition, the entanglement of execution threads allows you to modify the logic of the software operation by inserting unnecessary conditions, loops or transitions, which creates complex execution structures. The anti-debugging process is proposed, which is one of the key components of software protection against reverse engineering. A conceptual approach to the implementation of software anti-debugging mechanisms is formed, which consists in multi-level protection of software and changes the instructions and behavior of the virtual machine. One of the important elements of such algorithms is checking parent processes and detecting debugging using hardware breakpoints. This, in its turn, opens up the possibility during the design of a virtual machine in combination with other techniques to create significant obstacles for the analysis of the software code.
- 12 May 2025
- Published by Lviv Polytechnic National University in Information and communication technologies, electronic engineering
- Vol. 5 (1) , 154-161
ENVIRONMENT MONITORING WITH A PASSIVE PRESSURE-TEMPERATURE SENSOR BASED ON OPTICALLY ACTIVE MEDIA
The article is devoted to the study of the features of continuous monitoring of changes in the color of information channels of passive pressure-temperature sensors, tracking the spectral distribution of the light intensity of the liquid crystal color depending on the operating conditions. The main direction of application of such sensors can be “screening” of chemically active environments, deployment of temporary objects of short-term use in conditions of a pandemic, or military operations, etc. The functional component of a passive, contactless sensor is a combination of an elastic silicon membrane and cholesteric liquid crystals as an optically active environment, which, due to the action of an external physical factor, such as pressure or temperature, undergo a rearrangement of the geometry of the spiral of the supramolecular structure of the film. In this case, it is necessary to take into account the ratio between useful and parasitic radiation passing through the primary transducer, which is determined mainly by the ratio of the areas of a homogeneous polymer area and an area with liquid crystal inclusions in the cross section of the primary transducer. It was determined that the use of optical diagnostic methods in spectroscopy, including colorimetry, opens up the possibility of monitoring environments, substances and their electrophysical properties without using electrically induced methods that cause a high level of electromagnetic interference. Low cost and lack of power consumption ensure high competitiveness of such devices. It is proposed to use a hardware and software complex for monitoring, which contains a spectrometer and allows real-time display, as well as recording the received information in file arrays, various optical characteristics such as spectral intensity distribution, color rendering index, etc. These characteristics, which are pre-calibrated to the influence of external physical factors, provide information about the measured value. It was established that in the wavelength range from 370 to 650 nm, the proposed method for evaluating the color characteristics of samples using optically active media of liquid polymer cholesteric crystals allows evaluating the parameters of the spectral intensity distribution with sufficient sensitivity. Thus, for wavelengths from 450 to 550 nm, the maximum intensity is I=1.4 arb.u., and the resolution of the method lies in the wavelength range up to 100 nm, i.e., when the wavelength changes by 10 nm, the intensity increases by 0.14 arb.u.
- 12 May 2025
- Published by Lviv Polytechnic National University in Information and communication technologies, electronic engineering
- Vol. 5 (1) , 154-161
ENVIRONMENT MONITORING WITH A PASSIVE PRESSURE-TEMPERATURE SENSOR BASED ON OPTICALLY ACTIVE MEDIA
The article is devoted to the study of the features of continuous monitoring of changes in the color of information channels of passive pressure-temperature sensors, tracking the spectral distribution of the light intensity of the liquid crystal color depending on the operating conditions. The main direction of application of such sensors can be “screening” of chemically active environments, deployment of temporary objects of short-term use in conditions of a pandemic, or military operations, etc. The functional component of a passive, contactless sensor is a combination of an elastic silicon membrane and cholesteric liquid crystals as an optically active environment, which, due to the action of an external physical factor, such as pressure or temperature, undergo a rearrangement of the geometry of the spiral of the supramolecular structure of the film. In this case, it is necessary to take into account the ratio between useful and parasitic radiation passing through the primary transducer, which is determined mainly by the ratio of the areas of a homogeneous polymer area and an area with liquid crystal inclusions in the cross section of the primary transducer. It was determined that the use of optical diagnostic methods in spectroscopy, including colorimetry, opens up the possibility of monitoring environments, substances and their electrophysical properties without using electrically induced methods that cause a high level of electromagnetic interference. Low cost and lack of power consumption ensure high competitiveness of such devices. It is proposed to use a hardware and software complex for monitoring, which contains a spectrometer and allows real-time display, as well as recording the received information in file arrays, various optical characteristics such as spectral intensity distribution, color rendering index, etc. These characteristics, which are pre-calibrated to the influence of external physical factors, provide information about the measured value. It was established that in the wavelength range from 370 to 650 nm, the proposed method for evaluating the color characteristics of samples using optically active media of liquid polymer cholesteric crystals allows evaluating the parameters of the spectral intensity distribution with sufficient sensitivity. Thus, for wavelengths from 450 to 550 nm, the maximum intensity is I=1.4 arb.u., and the resolution of the method lies in the wavelength range up to 100 nm, i.e., when the wavelength changes by 10 nm, the intensity increases by 0.14 arb.u.
- Open Access
- 2 May 2025
- Published by Ukrainian National Forestry University in Scientific Bulletin of UNFU
- Vol. 35 (2) , 110-115
Features of using swarm intelligence algorithms for drone route optimization
This paper investigates the use of swarm algorithms for optimising delivery routes using Unmanned Vehicles (UVs) as an effective alternative to traditional algorithms. Classical path planning algorithms such as Dijkstra's algorithm, A* or APF (Artificial Potential Field) can struggle in real-world environments due to the ability to fall into local minima, or due to increased computational complexity in cases with multiple obstacles. To avoid these limitations, the research focuses on path-planning methods through swarm algorithms, many of which are inspired by animal biological behaviour. The article considers the features of operation and suitability for use in dynamic conditions of such algorithms as Particle Swarm Optimisation (PSO), Artificial Bee Colony (ABC), Cuckoo Search Algorithm (CSA) and Sparrow Search Algorithm (SSA). The study showed that there is no one universal algorithm for all types of environments. All algorithms demonstrate high performance when used under the best conditions. An algorithm's effectiveness depends on considering weaknesses and strengths when planning a task. It was found that PSO shows fast convergence even when using a large number of agents, which makes it suitable for route planning in swarm systems. Instead, the extensibility of the ABC algorithm and its ease of computational complexity allows for the efficient distribution of tasks among agents in large swarm systems. The CSA allows for the best local planning, which, in combination with other algorithms, allows for the best results in the balance between global and local route planning. SSA achieves fast convergence at the cost of computational resources, while ensuring execution in a dynamic environment. The results of the study show that the best algorithms for dynamic environments are those that combine local and global search capabilities, such as PSO modifications. The study emphasises the role of swarm intelligence algorithms in improving UAVs' performance by providing robust, efficient and scalable delivery routes. Future research could concentrate on exploring new hybrid algorithm models that combine swarm intelligence algorithms with machine learning technologies to improve algorithm outcomes in dynamic environments. In order to enhance the algorithms and analyse their limitations, further testing of the algorithms in complex environments with a large number of interferences is important to make further improvements to the algorithms.
- 2 May 2025
- Published by Ukrainian National Forestry University in Scientific Bulletin of UNFU
- Vol. 35 (2) , 110-115
Open AccessFeatures of using swarm intelligence algorithms for drone route optimization
This paper investigates the use of swarm algorithms for optimising delivery routes using Unmanned Vehicles (UVs) as an effective alternative to traditional algorithms. Classical path planning algorithms such as Dijkstra's algorithm, A* or APF (Artificial Potential Field) can struggle in real-world environments due to the ability to fall into local minima, or due to increased computational complexity in cases with multiple obstacles. To avoid these limitations, the research focuses on path-planning methods through swarm algorithms, many of which are inspired by animal biological behaviour. The article considers the features of operation and suitability for use in dynamic conditions of such algorithms as Particle Swarm Optimisation (PSO), Artificial Bee Colony (ABC), Cuckoo Search Algorithm (CSA) and Sparrow Search Algorithm (SSA). The study showed that there is no one universal algorithm for all types of environments. All algorithms demonstrate high performance when used under the best conditions. An algorithm's effectiveness depends on considering weaknesses and strengths when planning a task. It was found that PSO shows fast convergence even when using a large number of agents, which makes it suitable for route planning in swarm systems. Instead, the extensibility of the ABC algorithm and its ease of computational complexity allows for the efficient distribution of tasks among agents in large swarm systems. The CSA allows for the best local planning, which, in combination with other algorithms, allows for the best results in the balance between global and local route planning. SSA achieves fast convergence at the cost of computational resources, while ensuring execution in a dynamic environment. The results of the study show that the best algorithms for dynamic environments are those that combine local and global search capabilities, such as PSO modifications. The study emphasises the role of swarm intelligence algorithms in improving UAVs' performance by providing robust, efficient and scalable delivery routes. Future research could concentrate on exploring new hybrid algorithm models that combine swarm intelligence algorithms with machine learning technologies to improve algorithm outcomes in dynamic environments. In order to enhance the algorithms and analyse their limitations, further testing of the algorithms in complex environments with a large number of interferences is important to make further improvements to the algorithms.