The use of magnetic tracking technology has become increasingly popular in recent years for human motion studies. However, there have been few independent evaluations of how these systems perform. The purpose of this study was to develop a dynamic pendulum calibration method to test the performance of magnetic tracking sensors. A nonmetallic pendulum was constructed and instrumented with a rotary potentiometer. A cube was attached to the distal end of the pendulum so that sensors could be mounted orthogonally. In this manner, it was possible to obtain simultaneous recordings of azimuth, elevation, and roll depending on the sensor mounting orientation relative to the axis of rotation of the pendulum. Sensor data, using Flock of Birds™ sensors, and potentiometer data were collected simultaneously during dynamic pendulum motion at two transmitter distances and then were compared. The results showed excellent trial-to-trial repeatability of 2% or better for the sensors, and high correlations between the sensor and potentiometer data. RMS errors range from about 3 to 10 mm depending on the angular velocity of the pendulum. Angular errors were less than 1 degree RMS for all speeds.
The authors are with the Dept. of Physical Therapy at the University of Maryland School of Medicine, 100 Penn St., Baltimore, MD 21201; K.J. McQuade is also with the Veterans Administration Medical Center in Baltimore.