Research and Development

Applied engineering and sensor-technology development that bridges biomechanical expertise with robotics and computer vision.

Model-based motion capture with depth sensors

During my time in the Biomechanics of Human Movement and Ergonomics Lab at the University of Extremadura, I collaborated with the university’s robotics and artificial-vision research group on the development of markerless, depth-sensor-based human motion capture.

Low-cost RGB-D sensors (such as the Microsoft Kinect / PrimeSense) paired with open frameworks like OpenNI made human tracking accessible without invasive marker-based equipment, but the plausibility of the resulting body poses was not guaranteed — sensor noise and tracking errors could produce anatomically implausible postures. Drawing on my background in kinematics and biomechanical modeling, I contributed to a model-based pose generator that reinforces raw OpenNI tracking data with kinematic constraints, removes implausible postures, reduces sensor noise, and adapts to each individual’s body dimensions — improving pose accuracy over standard OpenNI tracking at minimal added computational cost.

This line of work reflects a broader interest in translating biomechanical and motor-control principles into practical sensing and engineering solutions, alongside my core research in neurorehabilitation and adapted physical activity.