Gait & Posture
Bayesian statistical decision theory, Indoor positioning systems (Wireless localization), Ultra-wideband communication systems, Wearable technology
Wearable devices with embedded kinematic sensors including triaxial accelerometers, gyroscopes, and magnetometers are becoming widely used in applications for tracking human movement in domains that include sports, motion gaming, medicine, and wellness. The kinematic sensors can be used to estimate orientation, but can only estimate changes in position over short periods of time. We developed a prototype sensor that includes ultra wideband ranging sensors and kinematic sensors to determine the feasibility of fusing the two sensor technologies to estimate both orientation and position. We used a state space model and applied the unscented Kalman filter to fuse the sensor information. Our results demonstrate that it is possible to estimate orientation and position with less error than is possible with either sensor technology alone. In our experiment we obtained a position root mean square error of 5.2 cm and orientation error of 4.8° over a 15 min recording.
Vasilyev P., Pearson S., El-Gohary M., Aboy M., McNames J. 2017. Inertial and Time-of-Arrival Ranging Sensor Fusion. Gait & Posture, 54:1-7.