3D Mapping of Multi-floor Buildings Based on Sensor Fusion

Abstract

In this paper, we design a highly accurate method for 3D mapping of multi-floor buildings. The basic idea is to combine the laser range sensor for metric mapping and barometric pressure sensor for detecting floor transition and kinect sensor for collection 3D environment information. Meanwhile, we adopt the Monte Carlo localization in 2D map to improve the accuracy of the localization. Finally, the barometric pressure is used to merge the 3D map of the multi-floor buildings. By using the information collected by a real robot in a typical multi-floor environment, the method is tested and compared with some other approaches, the results show that the method is efficient and has a better result in 3D mapping of multi-floor buildings.