Enhancing Prosthetic Hand Control: A Study on IMU Sensor-Based Machine Learning for Precise Hand Orientation Classification

Abstract

The functionality of the human upper extremity is a cornerstone of daily life, enabling essential activities and environmental interactions. This study underscores the importance of using Inertial Measurement Unit (IMU) sensors and deploying machine learning models to enhance accuracy in classifying different hand orientations. As such, our work is an initial step towards empowering individuals who have experienced limb loss with a pathway to increased interaction with the environment and object-grasping skills and speed. We harnessed IMU sensors to collect data from various positions of a collaborative robot (7 -axis) during grasping tasks with diverse objects and orientations. Leveraging a combined machine learning model, we achieved an impressive classification accuracy of 99.8%.