Only 2D proximity information, gathered from 2D laser range finder, may cause problems with overhanging obstacles(chair, table, etc.). Because a 3D obstacle has different cross sections with height, it can only be detected properly when 3D scan is performed.
We developed 3D sensing scheme which enables localization and 3D obstacle detection simultaneously using only one laser range finder and we developed real-time localization algorithm which runs on RTAI. When mobile robot with tilted laser range finder moves, it rebuilds 3D environment model from 3D scan data and visualize the result in 3D. Visualization program based on Linux GTK+ can present the environment model in top view or in perspective view.
Through corridor navigation, the performance of localization algorithm is measured and reconstructed 3D environment model with obstacles is presented.