A Novel Pose Estimation Scheme for Permanent Magnet Marker Using Rotating Frame Method

Abstract

This thesis presents a new localization scheme for permanent magnet marker(PMM) with a low dimensional optimization process. Existing PMM localization methods using hall sensors were basically optimized by a 6-dimension optimization parameter vector. The previous method based on the ideal equation made it impossible to apply the PMM to a real environment with external environmental changes such as temperature changes outside of the ideal experimental environment. Additionally, it needs large computational costs and thus shows low real-time performance. In this study, by using a new optimization function called the Rotating Frame Method(RFM), it is shown that PMM localization which has high accuracy and real-time performance is possible with low-dimension optimization parameters corresponding to the roll and pitch values of the rotation of sensor's coordinate frame. Since the PMM localization method is currently used as a base technology for capsule-type robots, soft sensors, and soft robots, the newly proposed pose estimation scheme suggest a new method for their actual environment.