Consistent map building in petrochemical complexes for firefighter robots using SLAM based on GPS and LIDAR

• DOI: 10.1186/s40648-018-0104-z
• Publication Date: 2018-04-04T05:18:44Z
• Authors: Abu Ubaidah Shams...; Kazunori Ohno; Ryunosuke Hamada; Shotaro Kojima; Thomas Westfechtel; Takahiro Suzuki; Yoshito Okada; Satoshi Tadokoro; Jun Fujita; Hisanori Amano
• Keywords: Ranging; Ground truth

The objective of this study was to achieve simultaneous localization and mapping (SLAM) firefighter robots for petrochemical complexes. Consistency the SLAM map is important because human operators compare with aerial images identify target positions on map. global positioning system (GPS) enables increased consistency. Therefore, paper describes two Rao-Blackwellized particle filters (RBPFs) based GPS light detection ranging (LIDAR) as solutions. Fast-SLAM 1.0 2.0 were used in grid maps RBPFs study. We herein propose use combine LIDAR. difference between original proposed method log-likelihood function GPS; combination implemented using a probabilistic mathematics formulation. methods evaluated sensor data measured real complex Japan size from 550–380 m. RTK-GPS measurement had an availability 56%. Our results showed that LIDAR dense produced best results. There significant improvement alignment data, mean square root error 0.65 To evaluate consistency, accurate 3D point cloud by Faro Focus (± 3 mm) ground truth. Building sizes compared; minimum errors 0.17 0.08 m oil refinery management building area sparse layout large tanks, respectively. Consequently, consistent map, which also (from Google Maps), built reproduced consistency ten runs variance ± 0.3 accuracy 0.52 low RTK-Fix-GPS environment, factory similar complexes 20.9% availability.