Miniaturized Low-Power Wearable System for Human Motion Tacking Incorporating Monocular Camera and Inertial Sensor Data Fusion for Health Applications

摘要

Human motion tracking systems play an important role in many application spaces where the 3D pose of the body, or its parts, need to be tracked over time, e.g. fitness, rehabilitation, etc. While it is considered a solved problem in expensive infrastructure-intensive setups, it remains an active research area in low-cost and low-power applications. In this work, we propose our wearable motion tracking system that leverages the advances in the state-of-the-art in system miniaturization and integration. Our proposed system integrates a monocular camera and an inertial motion sensor (IMU) in a single wearable unit. Moreover, it can perform a six degree-of-freedom (DOF) pose detection and tracking with our novel data fusion algorithms and a smart coupling with the ambient environment.