CMU-GPR Dataset: Ground Penetrating Radar Dataset for Robot Localization\n and Mapping

摘要

There has been exciting recent progress in using radar as a sensor for robot\nnavigation due to its increased robustness to varying environmental conditions.\nHowever, within these different radar perception systems, ground penetrating\nradar (GPR) remains under-explored. By measuring structures beneath the ground,\nGPR can provide stable features that are less variant to ambient weather,\nscene, and lighting changes, making it a compelling choice for long-term\nspatio-temporal mapping. In this work, we present the CMU-GPR dataset--an\nopen-source ground penetrating radar dataset for research in subsurface-aided\nperception for robot navigation. In total, the dataset contains 15 distinct\ntrajectory sequences in 3 GPS-denied, indoor environments. Measurements from a\nGPR, wheel encoder, RGB camera, and inertial measurement unit were collected\nwith ground truth positions from a robotic total station. In addition to the\ndataset, we also provide utility code to convert raw GPR data into processed\nimages. This paper describes our recording platform, the data format, utility\nscripts, and proposed methods for using this data.\n