A Novel Algorithm for the Calibration of Inertial/Magnetic Sensors: Application to a Body Sensor Network

Knowledge about the orientation of body segments is a vital source of information. It can be used to estimate human posture in a dynamic environment with multiple applications ranging from artifact or disturbance detection to rehabilitation robotics. The development and test of inertial measurement units requires calibration and a proper reference that can be provided by a motion tracking system. We present a novel approach for the fast calibration of inertial/magnetic sensors and the spatial synchronization with a reference system. The calibration procedure is tested using a Body Sensor Network-based inertial measurement unit and a Vicon tracking system to observe body limb orientation during normal gait on a treadmill. The proposed algorithm reduces the errors incorporated by misalignment of the reference frames and decalibration in the utilized experimental setup by 8.8°.