<![CDATA[The utilization of technology in sports has become a crucial need to enhance modern coaching efficiency. In pencak silat, particularly within Persaudaraan Setia Hati Terate (PSHT), accurate mastery of basic techniques is essential. However, when students practice independently without supervision, they often struggle to ensure correct hand movements, lowering training quality and increasing injury risks. As a solution, this study develops a real-time, Internet of Things (IoT)-based hand movement monitoring system using Inertial Measurement Unit (IMU) sensors. The sensor data is processed via a Machine Learning approach utilizing a 7-SVM Pipeline architecture. The Support Vector Machine (SVM) algorithm is applied to Model 0 (Movement Classifier) for movement classification, and Models 1–6 (Correctness Classifier) to evaluate quality into "Correct" or "Incorrect". The model is tested using the Leave-One-Subject-Out (LOSO) Cross-Validation method. Results show that Model 0 recognizes movement types with a 63.3% accuracy on unseen subjects. Meanwhile, the correctness models yield varying results; the highest achievement reaches 100% for the Left Jab, whereas the lowest is 55% for the Right Combination due to subjects' biomechanical variations. The results are displayed on a website dashboard as an objective companion tool for independent training while supporting digitalization in preserving pencak silat culture. ]]>
Copyrights © 2026
