This study presents a fundamental development of monitoring system for cervical spine posture designed in long-term observation by neck movements during daily activities. The system utilizes three nine-axis sensors placed at the head, neck (C3), and upper torso (T1) to estimate cervical spine posture by compensating for the wearer body tilt and incorporating segmental contribution rates for each vertebra. Sensor data are processed using Madgwick filter and visualized as three-dimensional motion in Unity programming environment. A comparative analysis between actually recorded motion and simulated posture had suggested that the estimation accuracy may have improved, particularly in axial rotation, by reflecting anatomical characteristics. However, improper alignment of the head sensor with the axis of rotation introduced unintended rotational components with reducing estimation accuracy. Future work will be focused on quantitative validation by comparing estimated data with ground-truth reference data.
GUO et al. (Wed,) studied this question.