Correlations between end point foot trajectories and inertial sensor data

Abstract

Inertial measurement units (IMU) comprising accelerometers and gyroscopes have recently found use in a wide range of motion analysis applications. Even though the technique of inferring secondary parameters such as velocity, displacement and angular rotation is greatly influenced by sensor noise, IMUs continue to find interest in movement studies due to other attractive properties such as light weight, wearability, compactness and low cost. In this study, the motivation was to determine the relationship between the raw foot inertial sensor data and the key end point foot trajectories. Correlations were calculated between the sensor data and the end points obtained from an optoelectronic motion analysis system, i.e. Optotrak Certus, NDI. It was found that vertical acceleration of the toe 0.06 s prior to toe-off were the most correlated with the end point foot trajectories. During toe-off, the amount of rotation of the distal foot was also significantly correlated with the first maximum - m×1 of the foot clearance. Foot rotation direction however varied between subjects tested suggesting different gait initiation strategies. The results are encouraging because they indicate that raw inertial sensor measurements are correlated with vertical foot end point trajectories, and could be used to detect the risk of falling.