Darkhan Zholtayev

Darkhan Zholtayev

Auto-generated profile, based on publicly available information

Darkhan Zholtayev

Nazarbayev University, Kazakhstan

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications

30 September 2026 in Scientific Journal of Astana IT University
Verified
Journal Article
Open access
PDF file
DZDarkhan Zholtayev
Darkhan Zholtayev
TMTemirlan Meiramkhanov
Temirlan Meiramkhanov
AOAiman Ozhikenova
Aiman Ozhikenova
Жадыра Нурдаулетовна АлимбаеваЖадыра Нурдаулетовна Алимбаева
Жадыра Нурдаулетовна Алимбаева
AOAssylbek Ozhiken
Assylbek Ozhiken
BABeibit Abdikenov
Beibit Abdikenov
Reliable control of upper-limb prostheses remains challenging because muscle signals can change between sessions and wearable EMG sensors may suffer from poor electrode contact. This study presents a real-time prosthetic hand control platform based on an OYMotion gForce Pro armband and a five-finger Dynamixel-actuated prosthetic hand. The system is designed to work without an external dataset, using only a short user-specific recording session and brief calibration before real-time use — specifically, five main training repetitions plus four supplementary calibration repetitions per gesture — intentionally minimising the calibration burden to enable practical single-session deployment. Five hand gestures—power grasp, open hand, OK gesture, index pointing, and thumbs-up—were classified using several machine-learning and ensemble-based methods. All models were evaluated under the same live data stream to ensure a fair comparison. Six recording sessions collected across five able-bodied participants on different recording days were included in the final statistical analysis. For every session, the main and supplementary training samples were combined to train session-specific models before live evaluation. Across six independent recording sessions from one able-bodied participant, Linear Discriminant Analysis achieved the best average live accuracy of 85.9%, followed closely by majority voting and adaptive ensemble methods. In contrast, the LSTM model showed lower performance, reaching 70.2% accuracy. The results indicate that, for low-data real-time prosthetic control, ensemble-based methods preserved near-best average accuracy while reducing reliance on any single classifier assumption, making them a more robust deployment choice than individual classifiers. Statistical analysis confirmed that session-to-session variability was the dominant factor in system performance (Kendall's W = 0.249) rather than classifier choice (W = 0.043), suggesting that practical reliability depends more on robust calibration and signal monitoring than on model selection.

28 October 2026 in Traumatology and Orthopаedics of Kazakhstan
Verified
Journal Article
Open access
PDF file
AAAyaulym Assylbekova
Ayaulym Assylbekova
DZDarkhan Zholtayev
Darkhan Zholtayev
AOAiman Ozhikenova
Aiman Ozhikenova
Жадыра Нурдаулетовна АлимбаеваЖадыра Нурдаулетовна Алимбаева
Жадыра Нурдаулетовна Алимбаева
EBErik Baizhunussov
Erik Baizhunussov
AOAssylbek Ozhiken
Assylbek Ozhiken
Objective. To assess participant-specific recognition of grasp type and subjectively scaled effort using surface electromyography in adults without upper-limb amputation, and unintended outputs during ordinary movements. This study is a preliminary step toward prosthetic control for people with congenital or acquired upper-limb absence. Methods. Four adults without upper-limb amputation performed eight command states: rest, open hand, and power grasp and tip pinch at low, medium and high verbally instructed effort. Effort categories were subjective; mechanical force was not measured. A separate linear discriminant analysis model was trained for each participant using time-domain features from consecutive signal windows, followed by causal majority voting over twenty windows. Evaluation comprised held-out trials separated from model fitting, newly recorded command trials processed causally, and ordinary movements without control intent; a confidence threshold of 0.60 was applied retrospectively to the latter. All recordings were obtained during one visit. No physical prosthesis was actuated. Results were summarized descriptively. Results. On held-out trials, mean balanced accuracy was 0.810 for raw windows, 0.823 after majority voting, and 0.875 for trial-level decisions. Among outputs with correctly recognized grasp type, effort recognition was more accurate for power grasp (0.970) than tip pinch (0.706). On newly recorded command trials, mean balanced accuracy was 0.902 and trial accuracy was 0.939; at the confidence threshold, coverage was 0.925 and end-to-end accuracy was 0.862. During ordinary movements, the mean proportion of unintended active outputs was 35.6% and decreased to 22.3% after confidence rejection, which retained 81.3% of outputs. The longest uninterrupted active-output episode after rejection lasted 7.25 seconds. Conclusion. The participant-specific decoder recognized instructed grasp and effort categories under controlled conditions. Unintended outputs during ordinary movements remain a concern for potential prosthetic use. These results do not establish clinical benefit or prosthesis control; evaluation with intended users, physical devices and functional tasks is needed.