Modular Wireless Inertial Motion Capture System with Self-Calibration

Wearable inertial motion capture systems enable movement analysis outside the laboratory, thereby promising valuable data collection in natural environments. We present an open-source motion capture system consisting of a variable number of wireless IMU sensor modules and a modular software framework providing OpenSim compatibility. System calibration is achieved by automatic sensor-to-segment alignment (S2SA) using a weighted least squares approach, enabling self-calibration based on arbitrary movements and ongoing recalibration. The S2SA approach is evaluated in a simulation and a proof-of-concept walking experiment with a human participant. Both in the simulation and in the experiment, all initializations converge, demonstrating the robustness of the automatic S2SA approach. Position and orientation deviations are below 0.09 m and 1.28/spl degree/, respectively. Lower limb joint angles exhibit root mean square errors ranging from 2.1/spl degree/ to 7.2/spl degree/ compared to a state-of-the-art S2SA method, while Pearson correlation coefficients are larger than 0.91 for all joints. Altogether, the results demonstrate the potential of self-calibration based on arbitrary movements, thereby enhancing the motion capture system's usability and paving the way towards clinical applicability.