On magnetometer heading updates for inertial pedestrian navigation system

Abstract

A magnetometer is often used to aid heading estimation of a low-cost Inertial Pedestrian Navigation System (IPNS) without which the latter will not be able to accurately estimate heading for more than a few seconds, even with the help of Zero Velocity Update (ZVU). Heading measurements from the magnetometer are typically integrated with gyro heading in an estimation filter such as Kalman Filter (KF) — to best estimate the true IPNS heading, resulting in a better positioning accuracy. However indoors the reliability of these measurements is often questionable because of the magnetic disturbances that can disrupt the measurements. To solve this problem, a filtering method is often used to select the best measurements. However, the importance of the frequency of these measurement updates has not been highlighted. In this paper, the impact of frequency of magnetometer updates on the overall accuracy of the navigation system is presented. The paper starts by discussing the use of a magnetometer in a low-cost IPNS. An exemplary filter to extract reliable heading measurements from the magnetometer is then described. From real field trial results, it will be shown that even if reliable heading measurements may be obtained indoors, it is still insufficient to increase the positioning accuracy of the low-cost IPNS unless it is reliable on every epoch.