Anatomy Calibration of Inertial Measurement Unit Using a Principle Component Analysis

Abstract

Recently, as a result of advances in technology, micro-electro-mechanical system (MEMS)based inertial measurement units (IMUs) have frequently been used for human motion analyses. While the output signal from an IMU is based on the sensor coordinate system (CS), it is necessary to express the signal in the body segment CS for clinical meaning. The offset between the sensor CS and anatomical CS needs to be determined. In this study, a calibration method that determines the orientation offset between a body segment and an IMU attached to the body segment was proposed. The angular velocity was measured while the body segment was rotated and processed using a principle component analysis to form the basis of the orientation offset. The proposed method was applied to calibrate an IMU attached to the forearm and evaluated for five young healthy subjects. Its performance showed reasonable repeatability (mean < 1.4°, SD < 0.8°).