AAArtem Akulynichev

Artem Akulynichev

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Artem Akulynichev has a total of 28 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Sensor-Modality-Aware Human Activity Recognition with the Convolutional Tsetlin Machine: Interpretable and Resource-Efficient Neuro-Symbolic Learning

    by Olga Tarasyuk, Anatoliy Gorbenko, Oleksandr Gordieiev, Artem Akulynichev, Rishad Shafik, Alex Yakovlev

    Human activity recognition (HAR) based on smartphone and wearable sensor data is commonly addressed using statistical learning methods and deep neural networks that often provide strong predictive performance, but at the expense of limited interpretability and substantial computational and energy requirements. Such limitations reduce their suitability for deployment in practical sensing environments where model decisions must be transparent, verifiable and executable on resource-constrained devices. In this work, we investigate the Convolutional Tsetlin Machine (CTM) for multimodal HAR using only the raw inertial signals (9 × 128) of the UCI-HAR dataset, rather than its pre-computed 561-feature representation. The Tsetlin Machine is a novel neuro-symbolic machine learning approach that offers two important advantages over many conventional machine learning methods: (i) it learns logic-based decision rules that support human inspection and provide a transparent basis for analyzing model decisions, and (ii) it operates with comparatively low computational complexity, making it well suited to efficient and low-power on-device learning. The proposed study systematically analyses the contribution of different feature modalities by decomposing the inertial signals space into semantically defined subsets according to: (i) sensor source: accelerometer and gyroscope; (ii) signal group: gyroscope angular velocity, body and total acceleration (including gravity); (iii) coordinate axis: x, y and z. A separate CTM classifier was trained for each modality and its combinations in order to determine the relative discriminative value of each modality group for activity classification. In addition to predictive performance, the study emphasizes the interpretability of the CTM model ensured by expressing each decision in the form of propositional clauses, thereby enabling visualization and direct inspection of the modality-specific patterns supporting each activity class. Owing to its symbolic structure and modest computational demands, the CTM provides a principled framework for the design of explainable, resource-efficient and deployable HAR systems. The proposed work therefore contributes toward trustworthy multimodal sensing by jointly addressing predictive performance, interpretability and suitability for embedded and mobile platforms.

  • Open Access

    Sensor Modality-Aware Human Activity Recognition with the Convolutional Tsetlin Machine: Interpretable and Resource-Efficient Neuro-Symbolic Learning

    by Olga Tarasyuk, Anatoliy Gorbenko, Oleksandr Gordieiev, Artem Akulynichev, Rishad Shafik, Alex Yakovlev

    Human activity recognition (HAR) based on smartphone and wearable sensor data is commonly addressed using statistical learning methods and deep neural networks that often provide strong predictive performance, but at the expense of limited interpretability and substantial computational and energy requirements. Such limitations reduce their suitability for deployment in practical sensing environments where model decisions must be transparent, verifiable and executable on resource-constrained devices. In this work, we investigate the Convolutional Tsetlin Machine (CTM) for multimodal HAR using the UCI-HAR dataset. The Tsetlin Machine is a novel neuro-symbolic machine learning approach that offers two important advantages over many conventional machine learning methods: (i) it learns logic-based decision rules that are human-readable and formally verifiable, and (ii) it operates with comparatively low computational complexity, making it well suited to efficient and low-power on-device learning. The proposed study systematically analyses the contribution of different feature modalities by decomposing the inertial signals space into semantically defined subsets according to: (i) sensor source: accelerometer or gyroscope; (ii) physical component: body or gravity; (iii) coordinate: x, y or z. A separate CTM classifier was trained for each modality and their combination in order to determine the relative discriminative value of each modality group for activity classification. In addition to predictive performance the study emphasizes the interpretability of the CTM model ensured by expressing each decision in the form of propositional clauses, thereby enabling visualization and direct inspection of the modality-specific patterns supporting each activity class. Owing to its symbolic structure and modest computational demands, the CTM provides a principled framework for the design of explainable, resource-efficient and deployable HAR systems. The proposed work therefore contributes toward trustworthy multimodal sensing by jointly addressing predictive performance, interpretability and suitability for embedded and mobile platforms.

  • Open Access

    АНАЛІЗ СФІГМОГРАМ ЗА ДОПОМОГОЮ БІАМПЛІТУДИ

    by Вікторія Науменко, Олександр Тоцький, Сергій Підченко, Артем Акулинічев

    В роботі розглянуті результати аналізу восьми тестових сфігмограм, як за допомогою як добре відомих методів, таких як аналіз часових інтервалів, швидкості пульсової хвилі та стандартного спектрального аналізу, так і за допомогою запропонованих нових методів біспектрального аналізу , включно з аналізом кількості локальних максимумів біамплітуд. Встановлено, що оцінювання розподілу локальних максимумів біамплітуд може використовуватись як новий перспективний інтегральний показник відхилень стану серцево-судинної системи від норми.

  • Open Access

    ОПТИМІЗАЦІЯ КОМПАКТНИХ ЛОГПЕРІОДИЧНИХ ДИПОЛЬНИХ АНТЕН ІЗ РОЗШИРЕНОЮ НИЗЬКОЧАСТОТНОЮ ХАРАКТЕРИСТИКОЮ

    by Віктор Мішан, Артем Акулинічев, Євгеній Коломієць

    With the rapid technological advancements over the last decade, there is an increased demand for new generation wireless devices and communication systems. This work provides an overview of various methods of miniaturization of printed log-periodic dipole antennas in order to reduce their size. First, the work involves the design of a standard printed log-periodic dipole antenna that operates over a wide frequency range from 0.7 GHz to 8 GHz and has an acceptable gain over most of its bandwidth. This antenna serves as a baseline model for further research and optimization. Furthermore, this paper introduces a novel approach to extending the low-frequency response of these antennas. The approach entails replacing the longest direct dipole with a triangular dipole structure and optimizing the dimensions of the four longest dipoles. The outcome is an enhanced antenna capable of operating within the 0.4 GHz to 8 GHz range, albeit with a minor reduction in gain at lower frequencies. The research on extending the low-frequency response of printed log-periodic dipole antennas holds promise for enhancing antenna systems' performance across various applications, including wireless communications, radar, and satellite communications. The findings of this study can serve as a foundation for further innovations and developments in the field of antenna technology, making a significant contribution to the advancement of modern wireless technologies. This work is particularly valuable for engineers and scientists working in the realm of antenna technology and wireless communications. It introduces a fresh perspective on enhancing the low-frequency performance of printed log-periodic dipole antennas, which may find applications in diverse radio communication systems, thus addressing the growing demand for compact yet efficient antenna solutions in the ever-evolving world of wireless technology.

  • The Effects of Queueing Algorithms on QoS for Real-Time Traffic in Process of Load Balancing

    by Artem Proskochylo, Mikhail Zriakhov, Artem Akulynichev

    The paper considers both the basic mechanisms of queue management on the router and the principle of their operation. A comparative analysis of the following queuing mechanisms: FIFO (First-In First-Out), CQ (Custom Queuing), PQ (Priority Queuing), WFQ (Weighted Fair Queuing), CBWFQ (Class Based Weighted Fair Queuing) and LLQ (Low Latency Queuing) based on a number of parameters such as traffic volume, traffic type, service of mixed traffic, enable on router and number of queues was made. The results of analysis showed that each of the queuing mechanisms has its advantages when dealing with certain types of traffic. Whereas in the process of load balancing several interfaces of the router are involved simultaneously the choice of queuing mechanism on each of them will depend on QoS for real-time traffic. Thus in this work the effect of various queue management mechanisms (such as FIFO, PQ, CQ, WFQ, LLQ) to QoS for real-time traffic in the process of load balancing on the router in an IP network was also investigated. The simulation results showed that most of queuing mechanisms (except FIFO and CQ) in the process of load balancing allowed to provide high QoS for real time traffic by reducing the values of delays and jitter and a significant reduction of lost packets for such traffic flows. At the same time the results showed that in the load balancing process, the most efficient queues servicing mechanism is LLQ, which provides the lowest delay for VoIP and the end to-end delay commensurable with the PQ mechanism for video traffic.


  • Providing of invariance property for piezo resonance devices on the basis of adaptive systems contained predictive standard

    by Sergey Pidchenko, Alla Taranchuk, Alexander Totsky, Artem Akulynichev

    Conceptual regulations related to the development of invariant and multi-frequency piezo resonance devices with controlled dynamics (IPRD/CD) are considered. The control law is formulated for IPRD/CD in real time on the basis of predictive numerical analysis of the system core dynamics. The criterion of optimal control is designed analytically for minimization of energy expenses.

  • Open Access

    Researching the Impact of Parameters of the Developed Routing Models on Network Performance

    by Mohammad Alhihi, Hani Attar, Mohammad Samour, Artem Akulynichev

    Nowadays, it is hard to imagine work without applying the principle of computer networks, and every day the requirement for high-quality network maintenance is increasing. In order to have a high-quality network; always optimize QoS through the optimization of routing protocols is required. In this paper, the scientific task of optimizing routing processes in hybrid telecommunication networks with guaranteed quality of service is solved by developing models and methods of adaptive routing. To develop methods, a system of Telecommunications network (TN) mathematical models was constructed at the levels of its morphological and functional descriptions. We used a weighted oriented graph as a structural model. Formalization of the main indicators of the network operation efficiency is carried out, which are the network performance (or its derivatives – relative and normalized performances) and indicators of the degree of use of network resources – buffer memory capacities of nodes and bandwidth of the transmission paths. In this paper, an experimental study of the developed models and routing methods was carried out in order to verify their adequacy, evaluate the effectiveness, and develop practical recommendations. The scheme of experiment, focusing on the investigation of processes occurring in the network while solving routing tasks (data gathering, RT processing, distribution, and implementation), is proposed.


  • The technical invariance of piezoresonance devices of the infocommunication systems

    by Sergey Pidchenko, Alla Taranchuk, Anna Spivak, Artem Akulynichev

    The paper discusses the issues of technical invariance piezoresonance devices of infocommunication systems based on the use of multifrequency functional-algorithmic methods.


  • Overview of wireless technologies for organizing sensor networks

    by Artem Proskochylo, Andrey Vorobyov, Michail Zriakhov, Alexandr Kravchuk, Artem Akulynichev, Vladimir Lukin

    The paper deals with review of the main technologies of wireless access that are used in modern wireless sensor networks (WSN). Comparative analysis of access technologies for the most important parameters is provided. Main advantages and disadvantages for applying these technologies in WSN are reviewed.


  • Method of data compression for traffic monitoring

    by Nadejda Kozhemiakina, Vladimir Lukin, Nikolay Ponomarenko, Artem Akulynichev, Jaakko Astola, Karen Egiazarian

    In this paper a problem of compressing data containing information on basic parameters of network traffic is considered. Six test sets with different types of network traffic for known monitoring tool Wireshark are formed. Analysis of compression efficiency for these datasets by widely used archivers is carried out. It is shown that the main part of memory in compressed data relates to timestamps. A method for compressing timestamps that consists in delta calculation, Burrows-Wheeler transform (BWT), distance coding (DC) and recursive group coding (RGC) at the final stage is proposed. It is demonstrated that the use of RGC at the final stage provides more efficient coding compared to known methods. It is also shown that the proposed method of timestamps coding produces about twice larger compression ratio than WinRar.