This study investigates the application of sensor fusion and machine learning techniques for seated movement detection using inertial measurement unit (IMU) and electromyography (EMG) datasets. The IMU dataset includes user information, timestamps, and tri-axial accelerometer data, with movement activities as target labels. The EMG dataset comprises multiple channels of muscle activity readings, with labelled classes indicating specific movements. Our approach integrates K-Nearest Neighbours (KNN), Decision Tree, Random Forest, and a stacking classifier to classify and predict seated movements accurately. By leveraging sensor fusion from IMU and EMG data, this research aims to enhance the precision and reliability of movement detection, crucial for developing assistive technologies like trunk orthosis systems.

Keywords: -Nearest Neighbours (KNN), Decision Tree, Random Forest, and a Stacking classifier.