Nurgul Karymsakova
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Nurgul Karymsakova has a total of 52 co-authors. Here are the top co-authors by number of shared publications:
- Assiya Boltaboyeva
Al-Farabi Kazakh National University,Faculty of Information Technology and Artificial Intelligence,Almaty,Kazakhstan
4 shared publications - Kassymbek Ozhikenov
Department of Robotics and Technical Means of Automation, Satbayev University, Almaty 050013, Kazakhstan
4 shared publications - Zhanel Baigarayeva
Al-Farabi Kazakh National University,Faculty of Information Technology and Artificial Intelligence,Almaty,Kazakhstan
3 shared publications - Baglan Imanbek
Al-Farabi Kazakh National University,Faculty of Information Technology and Artificial Intelligence,Almaty,Kazakhstan
2 shared publications - Roza Beisembekova
Satbayev University
2 shared publications - Sholpan Zhumagulova
Al-Farabi Kazakh National University, Almaty, Kazakhstan
2 shared publications
Latest Publications
- Open Access
- 11 September 2026
- Published by MDPI AG in Engineering Proceedings
- Vol. 154 (1) , 78
A Discrete Consensus Protocol with Algebraic-Connectivity-Based Fault Tolerance for Decentralized Multi-Agent Communication Networks
Decentralized coordination of multi-agent communication networks underpins autonomous UAV swarms, sensor meshes, and cyber-physical systems that must operate without a single point of failure, yet quantitative design rules linking topology degradation to coordination collapse are still lacking. This paper aims to characterize the fault tolerance of a discrete linear consensus protocol that drives mobile agents’ scalar states toward a common value over a time-varying communication graph. The convergence rate and fault tolerance budget of the protocol are explicitly tied to the algebraic connectivity of the graph Laplacian, and three failure regimes—random independent failures (), sequential staged failures (), and targeted attacks on the highest-degree hubs ()—are analyzed using a full-factorial Monte-Carlo experiment of 1620 runs implemented in Python with the Mesa framework and NetworkX. Critical failure thresholds are 0.42–0.51 for , 0.38–0.48 for , and 0.28–0.38 for ; a Kruskal–Wallis test (p < 0.001) confirms the ordering , providing quantitative design rules for fault-tolerant consensus-driven networks.- 11 September 2026
- Published by MDPI AG in Engineering Proceedings
- Vol. 154 (1) , 78
Open AccessA Discrete Consensus Protocol with Algebraic-Connectivity-Based Fault Tolerance for Decentralized Multi-Agent Communication Networks
Decentralized coordination of multi-agent communication networks underpins autonomous UAV swarms, sensor meshes, and cyber-physical systems that must operate without a single point of failure, yet quantitative design rules linking topology degradation to coordination collapse are still lacking. This paper aims to characterize the fault tolerance of a discrete linear consensus protocol that drives mobile agents’ scalar states toward a common value over a time-varying communication graph. The convergence rate and fault tolerance budget of the protocol are explicitly tied to the algebraic connectivity of the graph Laplacian, and three failure regimes—random independent failures (), sequential staged failures (), and targeted attacks on the highest-degree hubs ()—are analyzed using a full-factorial Monte-Carlo experiment of 1620 runs implemented in Python with the Mesa framework and NetworkX. Critical failure thresholds are 0.42–0.51 for , 0.38–0.48 for , and 0.28–0.38 for ; a Kruskal–Wallis test (p < 0.001) confirms the ordering , providing quantitative design rules for fault-tolerant consensus-driven networks. - Open Access
- 16 July 2026
- Published by JVE International Ltd. in Vibroengineering PROCEDIA
- Vol. 63, 124-131
Intelligent microprocessor system for diagnostics and balancing of rechargeable batteries of various electrochemical types
Contemporary energy storage deployments increasingly incorporate heterogeneous cell chemistries within a single installation, yet most commercial battery management circuits remain optimized for a single electrochemical type. This paper describes a microprocessor-controlled platform capable of autonomous diagnostics and charge equalization across four battery technologies: lithium-ion NMC, lithium iron phosphate (LiFePO₄), nickel-metal hydride (NiMH), and lead-acid. The measurement front-end is a 16-bit delta-sigma converter delivering 46 μV resolution and ±1.4 mV per-cell accuracy, interfaced to an STM32F446RE ARM Cortex-M4 processor at 180 MHz. Equalization combines resistive dissipation for voltage spreads below 50 mV with a switched-capacitor network achieving above 90 % energy-transfer efficiency for larger disparities. State-of-charge tracking couples’ coulomb integration with an adaptive extended Kalman observer, maintaining accuracy within ±3 % even from a 40 % seed-error initialization. Benchtop trials on four-cell series packs of each chemistry recorded inter-cell convergence to ±5 mV within 22 min, SOC root-mean-square error below 3 %, and correct automatic chemistry classification across all laboratory sessions. The novelty of the proposed system is not in the individual use of battery diagnostics or cell balancing, which are well-established topics, but in their integration into a single microprocessor-based multi-chemistry platform capable of automatic electrochemical-type identification, chemistry-specific parameter loading, adaptive SOC/SOH estimation, and hybrid passive-active balancing. This distinguishes the proposed approach from conventional BMS solutions optimized for only one battery chemistry.
- 16 July 2026
- Published by JVE International Ltd. in Vibroengineering PROCEDIA
- Vol. 63, 124-131
Open AccessIntelligent microprocessor system for diagnostics and balancing of rechargeable batteries of various electrochemical types
Contemporary energy storage deployments increasingly incorporate heterogeneous cell chemistries within a single installation, yet most commercial battery management circuits remain optimized for a single electrochemical type. This paper describes a microprocessor-controlled platform capable of autonomous diagnostics and charge equalization across four battery technologies: lithium-ion NMC, lithium iron phosphate (LiFePO₄), nickel-metal hydride (NiMH), and lead-acid. The measurement front-end is a 16-bit delta-sigma converter delivering 46 μV resolution and ±1.4 mV per-cell accuracy, interfaced to an STM32F446RE ARM Cortex-M4 processor at 180 MHz. Equalization combines resistive dissipation for voltage spreads below 50 mV with a switched-capacitor network achieving above 90 % energy-transfer efficiency for larger disparities. State-of-charge tracking couples’ coulomb integration with an adaptive extended Kalman observer, maintaining accuracy within ±3 % even from a 40 % seed-error initialization. Benchtop trials on four-cell series packs of each chemistry recorded inter-cell convergence to ±5 mV within 22 min, SOC root-mean-square error below 3 %, and correct automatic chemistry classification across all laboratory sessions. The novelty of the proposed system is not in the individual use of battery diagnostics or cell balancing, which are well-established topics, but in their integration into a single microprocessor-based multi-chemistry platform capable of automatic electrochemical-type identification, chemistry-specific parameter loading, adaptive SOC/SOH estimation, and hybrid passive-active balancing. This distinguishes the proposed approach from conventional BMS solutions optimized for only one battery chemistry.
- 30 June 2026
- Published by Toraighyrov University in Bulletin of Toraighyrov University. Energetics series
- No. 2, 2026,p. 191-204
ОЦЕНКА НАДЕЖНОСТИ УСТРОЙСТВ ЖЕЛЕЗНОДОРОЖНОЙ АВТОМАТИКИ ПУТЕМ АППРОКСИМАЦИИ СТАТИСТИЧЕСКИХ ХАРАКТЕРИСТИК ОТКАЗОВ
Статья посвящена оценке эксплуатационной надежности устройств железнодорожной автоматики и телемеханики (ЖАТ) в условиях перехода к предиктивному техническому обслуживанию. Целью работы является определение критических временных интервалов наработки устройств на основе аппроксимации статистических характеристик их отказов. В ходе исследования проведен расчет вероятности безотказной работы P*(t), плотности распределения f*(t) и интенсивности отказов λ*(t) на интервале до 1500 часов. Анализ полученных эмпирических данных и построенных графиков позволил идентифицировать точку перегиба в диапазоне 1100-1150 часов, которая характеризует переход системы из стадии нормальной эксплуатации в фазу интенсивного износа. Установлено, что после преодоления порога в 1100 часов интенсивность отказов возрастает в 2,6 раза, достигая значения 1,212 к концу рассматриваемого периода, а вероятность безотказной работы снижается до 0,426. Сделан вывод, что рубеж в 1100 часов является оптимальным дедлайном для проведения планово-предупредительных ремонтов. Практическая значимость работы заключается в возможности использования полученных зависимостей для прогнозирования остаточного ресурса оборудования ЖАТ в рамках цифровизации транспортной инфраструктуры, что позволит существенно повысить безопасность движения поездов и оптимизировать эксплуатационные расходы. Ключевые слова: железнодорожная автоматика и телемеханика, надежность, износ, интенсивность отказов, критическая наработка.
- 30 June 2026
- Published by Toraighyrov University in Bulletin of Toraighyrov University. Energetics series
- No. 2, 2026,p. 191-204
ОЦЕНКА НАДЕЖНОСТИ УСТРОЙСТВ ЖЕЛЕЗНОДОРОЖНОЙ АВТОМАТИКИ ПУТЕМ АППРОКСИМАЦИИ СТАТИСТИЧЕСКИХ ХАРАКТЕРИСТИК ОТКАЗОВ
Статья посвящена оценке эксплуатационной надежности устройств железнодорожной автоматики и телемеханики (ЖАТ) в условиях перехода к предиктивному техническому обслуживанию. Целью работы является определение критических временных интервалов наработки устройств на основе аппроксимации статистических характеристик их отказов. В ходе исследования проведен расчет вероятности безотказной работы P*(t), плотности распределения f*(t) и интенсивности отказов λ*(t) на интервале до 1500 часов. Анализ полученных эмпирических данных и построенных графиков позволил идентифицировать точку перегиба в диапазоне 1100-1150 часов, которая характеризует переход системы из стадии нормальной эксплуатации в фазу интенсивного износа. Установлено, что после преодоления порога в 1100 часов интенсивность отказов возрастает в 2,6 раза, достигая значения 1,212 к концу рассматриваемого периода, а вероятность безотказной работы снижается до 0,426. Сделан вывод, что рубеж в 1100 часов является оптимальным дедлайном для проведения планово-предупредительных ремонтов. Практическая значимость работы заключается в возможности использования полученных зависимостей для прогнозирования остаточного ресурса оборудования ЖАТ в рамках цифровизации транспортной инфраструктуры, что позволит существенно повысить безопасность движения поездов и оптимизировать эксплуатационные расходы. Ключевые слова: железнодорожная автоматика и телемеханика, надежность, износ, интенсивность отказов, критическая наработка.
- Open Access
- 27 March 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (7) , 3232
Stochastic Sirs Modeling of Greenhouse Strawberry Infections and Integration with Computer Vision-Based Mobile Spraying Robot
Viral and fungal diseases of greenhouse strawberries lead to significant crop losses, while traditional uniform spraying schemes do not account for the actual distribution of infection foci or changes in the microclimate. This paper proposes an integrated system for greenhouse farms that combines a stochastic SIRS model of the epidemic process with a microclimate-dependent infection coefficient βeff(t), a computer vision module based on a lightweight YOLOv10n detector, and a mobile sprayer robot. For three sets of parameters corresponding to moderate infection, outbreak, and suppression scenarios, ensemble simulations are performed (100 realizations per scenario). The results show that the maximum number of infected plants reaches approximately 690 out of 1000 in the outbreak scenario and only about 28 out of 1000 in the suppression scenario, reflecting the effect of timely microclimate correction and local spraying. The YOLOv10n detector is used as a sensor to determine the proportion of affected plants I(0)/N and provides automatic formation of the initial conditions of the population model. The resulting forecasts then serve as the basis for selecting one of three operating modes for the spraying robot (observation, microclimate correction, local treatment). Unlike existing works that consider disease detection, epidemiological models, or robotic spraying separately, this paper proposes a unified closed-loop scheme of “computer vision—stochastic model—mobile robot,” linking detection quality with epidemic process forecasting and treatment strategy. In this study, the feasibility of the proposed system was examined through numerical simulations, detector-level performance evaluation, and offline image-based integrated validation of the detector-to-decision workflow. Full closed-loop experiments in a real greenhouse environment are planned for future work.
- 27 March 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (7) , 3232
Open AccessStochastic Sirs Modeling of Greenhouse Strawberry Infections and Integration with Computer Vision-Based Mobile Spraying Robot
Viral and fungal diseases of greenhouse strawberries lead to significant crop losses, while traditional uniform spraying schemes do not account for the actual distribution of infection foci or changes in the microclimate. This paper proposes an integrated system for greenhouse farms that combines a stochastic SIRS model of the epidemic process with a microclimate-dependent infection coefficient βeff(t), a computer vision module based on a lightweight YOLOv10n detector, and a mobile sprayer robot. For three sets of parameters corresponding to moderate infection, outbreak, and suppression scenarios, ensemble simulations are performed (100 realizations per scenario). The results show that the maximum number of infected plants reaches approximately 690 out of 1000 in the outbreak scenario and only about 28 out of 1000 in the suppression scenario, reflecting the effect of timely microclimate correction and local spraying. The YOLOv10n detector is used as a sensor to determine the proportion of affected plants I(0)/N and provides automatic formation of the initial conditions of the population model. The resulting forecasts then serve as the basis for selecting one of three operating modes for the spraying robot (observation, microclimate correction, local treatment). Unlike existing works that consider disease detection, epidemiological models, or robotic spraying separately, this paper proposes a unified closed-loop scheme of “computer vision—stochastic model—mobile robot,” linking detection quality with epidemic process forecasting and treatment strategy. In this study, the feasibility of the proposed system was examined through numerical simulations, detector-level performance evaluation, and offline image-based integrated validation of the detector-to-decision workflow. Full closed-loop experiments in a real greenhouse environment are planned for future work.
- Open Access
- 24 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (5) , 2185
Modeling a Railway Section to Assess the Effectiveness of Fixed- and Moving-Block Systems
This study was carried out to determine the reliability of the methods of transmission of information about the location of a train to the train control center using a digital radio channel and a method based on Distributed Acoustic Sensing (DAS) technology. The study results were obtained based on the MATLAB R2024b model and showed resistance to external noise in fiber-optic communication with DAS technology. The proposed information transmission method allows the joint use of fixed- and moving-block section concepts in train traffic control systems. The effectiveness of the joint application of the concept of fixed and moving-block sections was analyzed using OpenTrack V1.10 microscopic simulation using the parameters of the operating railway section Kurozek-Ekpindi-Jarsu, locomotive TE33A series, trains No. 3002 and No. 3004, and the interval control system. The obtained research results in the form of a diagram showed the effectiveness of the proposed method of duplicating the concept of moving- and fixed-block sections. The reduction in inter-train intervals contributes to increasing the capacity of the railway line and is the key to the economic efficiency of railway transport.
- 24 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (5) , 2185
Open AccessModeling a Railway Section to Assess the Effectiveness of Fixed- and Moving-Block Systems
This study was carried out to determine the reliability of the methods of transmission of information about the location of a train to the train control center using a digital radio channel and a method based on Distributed Acoustic Sensing (DAS) technology. The study results were obtained based on the MATLAB R2024b model and showed resistance to external noise in fiber-optic communication with DAS technology. The proposed information transmission method allows the joint use of fixed- and moving-block section concepts in train traffic control systems. The effectiveness of the joint application of the concept of fixed and moving-block sections was analyzed using OpenTrack V1.10 microscopic simulation using the parameters of the operating railway section Kurozek-Ekpindi-Jarsu, locomotive TE33A series, trains No. 3002 and No. 3004, and the interval control system. The obtained research results in the form of a diagram showed the effectiveness of the proposed method of duplicating the concept of moving- and fixed-block sections. The reduction in inter-train intervals contributes to increasing the capacity of the railway line and is the key to the economic efficiency of railway transport.
- Open Access
- 27 January 2026
- Published by MDPI AG in Algorithms
- Vol. 19 (2) , 99
Architecting the Orthopedical Clinical AI Pipeline: A Review of Integrating Foundation Models and FHIR for Agentic Clinical Assistants and Digital Twins
The exponential growth of multimodal orthopedic data, ranging from longitudinal Electronic Health Records to high-resolution musculoskeletal imaging, has rendered manual analysis insufficient. This has established Large Language Models (LLMs) as algorithmically necessary for managing healthcare complexity. However, their deployment in high-stakes surgical environments presents a fundamental algorithmic paradox: while generic foundation models possess vast reasoning capabilities, they often lack the precise, protocol-driven domain knowledge required for safe orthopedic decision support. This review provides a structured synthesis of the emerging algorithmic frameworks required to build modern clinical AI assistants. We deconstruct current methodologies into their core components: large-language-model adaptation, multimodal data fusion, and standardized data interoperability pipelines. Rather than proposing a single proprietary architecture, we analyze how recent literature connects specific algorithmic choices such as the trade-offs between full fine-tuning and Low-Rank Adaptation to their computational costs and factual reliability. Furthermore, we examine the theoretical architectures required for ‘agentic’ capabilities, where AI systems integrate outputs from deep convolutional neural networks and biosensors. The review concludes by outlining the unresolved challenges in algorithmic bias, security, and interoperability that must be addressed to transition these technologies from research prototypes to scalable clinical solutions.
- 27 January 2026
- Published by MDPI AG in Algorithms
- Vol. 19 (2) , 99
Open AccessArchitecting the Orthopedical Clinical AI Pipeline: A Review of Integrating Foundation Models and FHIR for Agentic Clinical Assistants and Digital Twins
The exponential growth of multimodal orthopedic data, ranging from longitudinal Electronic Health Records to high-resolution musculoskeletal imaging, has rendered manual analysis insufficient. This has established Large Language Models (LLMs) as algorithmically necessary for managing healthcare complexity. However, their deployment in high-stakes surgical environments presents a fundamental algorithmic paradox: while generic foundation models possess vast reasoning capabilities, they often lack the precise, protocol-driven domain knowledge required for safe orthopedic decision support. This review provides a structured synthesis of the emerging algorithmic frameworks required to build modern clinical AI assistants. We deconstruct current methodologies into their core components: large-language-model adaptation, multimodal data fusion, and standardized data interoperability pipelines. Rather than proposing a single proprietary architecture, we analyze how recent literature connects specific algorithmic choices such as the trade-offs between full fine-tuning and Low-Rank Adaptation to their computational costs and factual reliability. Furthermore, we examine the theoretical architectures required for ‘agentic’ capabilities, where AI systems integrate outputs from deep convolutional neural networks and biosensors. The review concludes by outlining the unresolved challenges in algorithmic bias, security, and interoperability that must be addressed to transition these technologies from research prototypes to scalable clinical solutions.
- Open Access
- 1 December 2025
- Published by Institute of Advanced Engineering and Science in International Journal of Electrical and Computer Engineering (IJECE)
- Vol. 15 (6) , 5954-5967
Machine learning-based classification of local muscle fatigue using electromyography signals for enhanced rehabilitation outcomes
Muscle fatigue is a key factor affecting rehabilitation progress, safety, and patient engagement. Accurate detection of fatigue during physical activity remains a challenge, particularly in clinical and remote settings. This study presents the development of an Internet of things-based system for classifying local muscle fatigue using surface electromyography (EMG) signals and machine learning. A wearable device was used to collect real-time EMG data and subjective fatigue ratings from 10 healthy participants during sustained isometric grip exercises. Feature extraction was performed on-device, and the data were transmitted wirelessly for analysis. Machine learning models including logistic regression, decision tree (DT), random forest, and extreme gradient boosting (XGBoost) were trained to classify fatigue states. The DT model achieved the highest accuracy of 90.7%, with a precision of 90.7% and a recall of 90.9%. SHAP analysis revealed time under load, smoking, and alcohol use as the most influential factors in fatigue classification. These results show that wearable EMG devices combined with smart algorithms are effective for real-time fatigue monitoring during rehabilitation.
- 1 December 2025
- Published by Institute of Advanced Engineering and Science in International Journal of Electrical and Computer Engineering (IJECE)
- Vol. 15 (6) , 5954-5967
Open AccessMachine learning-based classification of local muscle fatigue using electromyography signals for enhanced rehabilitation outcomes
Muscle fatigue is a key factor affecting rehabilitation progress, safety, and patient engagement. Accurate detection of fatigue during physical activity remains a challenge, particularly in clinical and remote settings. This study presents the development of an Internet of things-based system for classifying local muscle fatigue using surface electromyography (EMG) signals and machine learning. A wearable device was used to collect real-time EMG data and subjective fatigue ratings from 10 healthy participants during sustained isometric grip exercises. Feature extraction was performed on-device, and the data were transmitted wirelessly for analysis. Machine learning models including logistic regression, decision tree (DT), random forest, and extreme gradient boosting (XGBoost) were trained to classify fatigue states. The DT model achieved the highest accuracy of 90.7%, with a precision of 90.7% and a recall of 90.9%. SHAP analysis revealed time under load, smoking, and alcohol use as the most influential factors in fatigue classification. These results show that wearable EMG devices combined with smart algorithms are effective for real-time fatigue monitoring during rehabilitation.
- Open Access
- 31 August 2025
- Published by National Scientific Center of Traumatology and Orthopaedics named after Academician N.D. Batpenov in Traumatology and Orthopаedics of Kazakhstan
- Vol. 76 (4) , jto010
AI-Driven Smart Wearable Systems for Personalized Rehabilitation Monitoring
The growing demand for personalized and effective rehabilitation strategies has driven the development of AI-driven smart wearable systems. These systems provide real-time health monitoring, prediction, and adaptive feedback, which helps to improve clinical outcomes in neuro-musculoskeletal and postoperative rehabilitation. This scoping review analyzes 14 peer-reviewed articles published between 2018 and 2025, comprising clinical trials, cohort studies, and engineering applications. The selected studies were identified through structured searches in IEEE, MDPI, and other scholarly databases, based on relevance to AI-enhanced wearable rehabilitation devices. Commonly used AI algorithms include support vector machines (SVM), convolutional neural networks (CNN), and reinforcement learning (RL), enabling functions such as gait analysis, joint movement recognition, muscle activation tracking, and postural control. In addition, integration with IoT sensor networks, cloud-based platforms, and telemedicine interfaces was widely reported. The review finds that AI-enabled wearables significantly improve patient adherence, monitoring accuracy, and personalized therapy delivery. Nonetheless, challenges remain in data security, sensor calibration, interoperability, and long-term user retention. These results confirm that smart wearables play an important role in supporting personalized, data-driven rehabilitation.
- 31 August 2025
- Published by National Scientific Center of Traumatology and Orthopaedics named after Academician N.D. Batpenov in Traumatology and Orthopаedics of Kazakhstan
- Vol. 76 (4) , jto010
Open AccessAI-Driven Smart Wearable Systems for Personalized Rehabilitation Monitoring
The growing demand for personalized and effective rehabilitation strategies has driven the development of AI-driven smart wearable systems. These systems provide real-time health monitoring, prediction, and adaptive feedback, which helps to improve clinical outcomes in neuro-musculoskeletal and postoperative rehabilitation. This scoping review analyzes 14 peer-reviewed articles published between 2018 and 2025, comprising clinical trials, cohort studies, and engineering applications. The selected studies were identified through structured searches in IEEE, MDPI, and other scholarly databases, based on relevance to AI-enhanced wearable rehabilitation devices. Commonly used AI algorithms include support vector machines (SVM), convolutional neural networks (CNN), and reinforcement learning (RL), enabling functions such as gait analysis, joint movement recognition, muscle activation tracking, and postural control. In addition, integration with IoT sensor networks, cloud-based platforms, and telemedicine interfaces was widely reported. The review finds that AI-enabled wearables significantly improve patient adherence, monitoring accuracy, and personalized therapy delivery. Nonetheless, challenges remain in data security, sensor calibration, interoperability, and long-term user retention. These results confirm that smart wearables play an important role in supporting personalized, data-driven rehabilitation.
- Open Access
- 18 June 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (12) , 6840
A Review of Innovative Medical Rehabilitation Systems with Scalable AI-Assisted Platforms for Sensor-Based Recovery Monitoring
Artificial intelligence (AI) and machine learning (ML) have introduced new approaches to medical rehabilitation. These technological advances facilitate the development of large-scale adaptive rehabilitation platforms that can be tailored to individual patients. This review focuses on key technologies, including AI-driven rehabilitation planning, IoT-based patient monitoring, and Large Language Model (LLM)-powered virtual assistants for patient support. This review analyzes existing systems and examines how technologies can be combined to create comprehensive rehabilitation platforms that provide personalized care. For this purpose, a targeted literature search was conducted across leading scientific databases, including Scopus, Google Scholar, and IEEE Xplore. This process resulted in the selection of key peer-reviewed articles published between 2018 and 2025 for a detailed analysis. These studies highlight the latest trends and developments in medical rehabilitation, showcasing how digital technologies can transform rehabilitation processes and support patients. This review illustrates that AI, the IoT, and LLM-based virtual assistants hold significant promise for addressing current healthcare challenges through their ability to enhance, personalize, and streamline patient care.
- 18 June 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (12) , 6840
Open AccessA Review of Innovative Medical Rehabilitation Systems with Scalable AI-Assisted Platforms for Sensor-Based Recovery Monitoring
Artificial intelligence (AI) and machine learning (ML) have introduced new approaches to medical rehabilitation. These technological advances facilitate the development of large-scale adaptive rehabilitation platforms that can be tailored to individual patients. This review focuses on key technologies, including AI-driven rehabilitation planning, IoT-based patient monitoring, and Large Language Model (LLM)-powered virtual assistants for patient support. This review analyzes existing systems and examines how technologies can be combined to create comprehensive rehabilitation platforms that provide personalized care. For this purpose, a targeted literature search was conducted across leading scientific databases, including Scopus, Google Scholar, and IEEE Xplore. This process resulted in the selection of key peer-reviewed articles published between 2018 and 2025 for a detailed analysis. These studies highlight the latest trends and developments in medical rehabilitation, showcasing how digital technologies can transform rehabilitation processes and support patients. This review illustrates that AI, the IoT, and LLM-based virtual assistants hold significant promise for addressing current healthcare challenges through their ability to enhance, personalize, and streamline patient care.
- Open Access
- 31 January 2024
- Published by JSC Academy of Logistics and Transport in Вестник КазАТК
- Vol. 130 (1) , 240-251
ЖҮРЕК-ҚАН ТАМЫРЛАРЫ АУРУЛАРЫН ДӘЛ БОЛЖАУ ҮШІН МАШИНАЛЫҚ ОҚЫТУ ӘДІСТЕРІН ҚОЛДАНУ
Жүрек-қан тамырлары ауруларын анықтау және болжау медицина саласындағы негізгі міндеттер болып табылады. Бұл процестер дұрыс жіктеуді қамтамасыз етеді, бұл өз кезегінде кардиологтарға пациенттерге қажетті емдеуді қамтамасыз етуге көмектеседі. Медицина саласында машиналық оқыту қосымшаларын қолданудың артуы байқалады, өйткені олар деректерді талдау негізінде заңдылықтарды анықтай алады. Жүрек-қан тамырлары ауруларын жіктеуде машиналық оқытуды қолдану қате диагноздар санын азайту арқылы дәрігердің жұмысын жеңілдетуге көмектеседі. Бұл зерттеудің мақсаты - өлімді азайту үшін жүрек-қан тамырлары ауруларын дәл болжай алатын модель жасау. Бұл мақалада жіктеу дәлдігін жақсарту мақсатында k-mod кластерлеу әдісі ұсынылды. Зерттеуде random forest (RF), decision tree classifier (DT), multilayer perceptron (MP) және XGBoost (XGB) сияқты әртүрлі модельдер қолданылды. Нәтижелерді оңтайландыру үшін GridSearchCV көмегімен қолданбалы модель параметрлерін баптау жүргізілді. Әзірленген модель 70 000 жазбадан тұратын Kaggle нақты деректер жиынтығына қолданылды. Модельдерді оқыту 80:20 қатынасында бөлінген мәліметтерде жүргізілді. Алынған дәлдік нәтижелері келесідей: decision tree үшін - 86,37% (айқас тексерумен) және 86,53% (онсыз), XGBoost үшін-86,87% (айқас тексерумен) және 87,02% (онсыз), random forest үшін-87,05% (айқас тексерумен) және 86,92% (онсыз), multilayer perceptron үшін - 87,28% (айқас тексерумен) және 86,94% (онсыз). Ұсынылған модельдер үшін AUC (қисық астындағы аймақ) мәндері келесідей: decision tree: 0,94, XGBoost: 0,95, random forest: 0,95, multilayer perceptron: 0,95. Зерттеу негізінде multilayer perceptron айқас тексеруді қолдана отырып, дәлдік тұрғысынан барлық басқа алгоритмдерден асып түсті деген қорытынды жасауға болады. Бұл модель дәлдіктің ең жоғары деңгейіне 87,28% жетті.
- 31 January 2024
- Published by JSC Academy of Logistics and Transport in Вестник КазАТК
- Vol. 130 (1) , 240-251
Open AccessЖҮРЕК-ҚАН ТАМЫРЛАРЫ АУРУЛАРЫН ДӘЛ БОЛЖАУ ҮШІН МАШИНАЛЫҚ ОҚЫТУ ӘДІСТЕРІН ҚОЛДАНУ
Жүрек-қан тамырлары ауруларын анықтау және болжау медицина саласындағы негізгі міндеттер болып табылады. Бұл процестер дұрыс жіктеуді қамтамасыз етеді, бұл өз кезегінде кардиологтарға пациенттерге қажетті емдеуді қамтамасыз етуге көмектеседі. Медицина саласында машиналық оқыту қосымшаларын қолданудың артуы байқалады, өйткені олар деректерді талдау негізінде заңдылықтарды анықтай алады. Жүрек-қан тамырлары ауруларын жіктеуде машиналық оқытуды қолдану қате диагноздар санын азайту арқылы дәрігердің жұмысын жеңілдетуге көмектеседі. Бұл зерттеудің мақсаты - өлімді азайту үшін жүрек-қан тамырлары ауруларын дәл болжай алатын модель жасау. Бұл мақалада жіктеу дәлдігін жақсарту мақсатында k-mod кластерлеу әдісі ұсынылды. Зерттеуде random forest (RF), decision tree classifier (DT), multilayer perceptron (MP) және XGBoost (XGB) сияқты әртүрлі модельдер қолданылды. Нәтижелерді оңтайландыру үшін GridSearchCV көмегімен қолданбалы модель параметрлерін баптау жүргізілді. Әзірленген модель 70 000 жазбадан тұратын Kaggle нақты деректер жиынтығына қолданылды. Модельдерді оқыту 80:20 қатынасында бөлінген мәліметтерде жүргізілді. Алынған дәлдік нәтижелері келесідей: decision tree үшін - 86,37% (айқас тексерумен) және 86,53% (онсыз), XGBoost үшін-86,87% (айқас тексерумен) және 87,02% (онсыз), random forest үшін-87,05% (айқас тексерумен) және 86,92% (онсыз), multilayer perceptron үшін - 87,28% (айқас тексерумен) және 86,94% (онсыз). Ұсынылған модельдер үшін AUC (қисық астындағы аймақ) мәндері келесідей: decision tree: 0,94, XGBoost: 0,95, random forest: 0,95, multilayer perceptron: 0,95. Зерттеу негізінде multilayer perceptron айқас тексеруді қолдана отырып, дәлдік тұрғысынан барлық басқа алгоритмдерден асып түсті деген қорытынды жасауға болады. Бұл модель дәлдіктің ең жоғары деңгейіне 87,28% жетті.