Towards wearable platform for accurate unconstrained trunk motion tracking using inertial and strain sensors data fusion

The thesis focused on the development of a wearable motion tracking platform employing fiber strain sensors and inertial measurement units through a data fusion algorithm. The development of a smart sleeveless shirt for measuring the kinematic angles of the trunk in complicated 3-dimensional movements was demonstrated. Fiber strain sensors were integrated into the fabric as the sensing element of the system. Furthermore, a novel method for obtaining the kinematic data of joints based on the data from wearable sensors was proposed. More specifically, the proposed method uses the data from two gyroscopes and the smart shirt strain sensors in a combined machine learning-unscented Kalman filter (UKF) data fusion approach to track the three-dimensional movements of a joint accurately. The suggested technique thus avoids the common problems associated with extracting the movement information from accelerometer and magnetometer readings in the presence of disturbances. A study with 12 participants performing an exhaustive set of simple to complex trunk movements was conducted to investigate the performance of the developed algorithm. The results of this study demonstrated that the data fusion algorithm could significantly improve the accuracy of motion tracking in complicated 3-dimensional movements. Future work requires coherently combining both types of sensors in a wearable platform for full-body motion tracking so that the proposed algorithm can be tested in a variety of daily living activities.