Simultaneous Localization and Mapping (SLAM) has been a research focus for many years within the robotics community. Leveraging that expertise, we adapted SLAM algorithms to run on a body-worn sensing and computing package. The user wears a Hokuyo LIDAR, Microsoft Kinect, Microstrain IMU, battery, and laptop. A 2D map is built in real-time using iSAM and transmitted wirelessly to a remote workstation. We address issues unique to a body-worn device, including sensor roll and pitch (as a result of the user’s walk) and investigated calibration techniques for non-rigidly mounted sensors.
Maurice F. Fallon, Hordur Johannsson, Jonathan Brookshire, Seth Teller, John Leonard, “Sensor Fusion for Flexible Human-Portable Building-Scale Mapping”, IROS, Algarve, Portugal. October 2012.