Three-Dimensional Hand Trajectory Estimation from MIMUs during a Hand-Tapping Task

Wearable inertial sensors offer a practical solution to evaluate motor impairments affecting upper-limb function in neurological conditions, although accurate 3D trajectory reconstruction remains challenging. This pilot study compares two approaches, Analytical Double Integration (ADI) and Model-Based Optimization (MBO), for hand trajectory estimation during a hand-tapping task. Both methods showed good agreement with a ground truth. ADI provided more consistent accuracy across conditions, while MBO yielded reliable estimates in controlled tasks, supporting further validation in clinically relevant scenarios.