A Robust Position and Posture Measurement System Using Visual Markers and an Inertia Measurement Unit

Abstract

Automatic control of mobile robots and robot arms requires techniques for estimating the position and orientation of objects with high accuracy and robustness. Although methods using markers or machine learning have been developed, general-purpose and highly accurate estimations have not been realized. Our team developed a high-accuracy visual marker “LentiMark” for high-accuracy estimations of position and posture, but data are lost when the camera cannot obtain marker images. We therefore developed the Marker-IMU system for integrating visual markers with an inertia measurement unit (IMU). When cameras cannot acquire the image of a visual marker, any missing data are restored from IMU data. However, when calculating positions from acceleration sensor values, sensor error increases estimation error. Therefore, we developed a method for error correction using measurements from before and after the missing data. Evaluation experiments confirm that missing data can be estimated using the proposed method.