The ability to quantify athletic performance in real-world environments has many applications, including providing immediate feedback to the athlete so he or she can modify his or her technique and evaluating performance degradation due to fatigue or load carriage. In general, athletic performance is well-defined for constrained tasks (e.g., completing a specific obstacle or performing a specific movement), but our work demonstrates that different tasks require different algorithms to take advantage of the specific opportunities for correcting drift in sensor-based estimates of kinematics. Most recently, we have applied our knowledge of using wearable sensors to quantify human movement to investigate the biomechanics of throwing and pitching in real-world environments.
Quantifying performance on a balance beam obstacle using an array of wearable sensors.
Arm orientation angles for a 37-pitch bullpen session. Biomechanical data can be combined with pitch performance data to understand the relationships between pitching mechanics and pitch outcomes.