Feasibility of a Wearable, Sensor-based Motion Tracking System

Abstract

The objective of this study was to develop and evaluate the feasibility of a wearable, sensor-based motion tracking system that provides an economical and quantitative means of recording upper limb motion for physical rehabilitation. The tracking system is comprised of a wirelessly connected network of inertial measurement units (IMUs), each containing a gyroscope and an accelerometer. Two IMUs were rigidly attached to each subject's forearm and upper arm. A trajectorizing algorithm was developed to estimate the three dimensional upper limb motion based on the measurements of the IMUs. A major advantage of the algorithm is that it allows the IMUs to be attached with arbitrary orientation to each limb and no manual anthropomorphic measurements need to be performed. By recording specific, known motions, the sensors can be calibrated with respect to their orientation in space and with respect to their orientation relative to their respective body segments. During the experiment, healthy subjects performed elbow flexion-extension motions that were recorded using the IMUs. To validate the system including the accuracy of recorded data and the correctness of the trajectorizing algorithm, an optical motion capture system was also used to record the same motions. Results showed that the proposed motion tracking system measured the elbow joint angles of the flexion-extension motions with high consistency with the measurements obtained from the optical motion capture system. Statistical analysis showed that joint angles between two systems are highly correlated. The error of elbow joint angles measured by our system yielded small root mean square error (RMSE) and small median absolute deviation (MAD). These results suggest that an IMU-based (more specifically, a gyroscope-based) motion tracking system can be realistically used to accurately track a patient's motion without the need of numerous sensors or an overly complicated set-up.