不整地を移動する車両型ロボットを想定したSLAMの評価

抄録

Rescue robots are expected to be used to search for buildings at disaster sites. A technology called SLAM is used for autonomous movement of robots. Since rescue robots are mainly expected to be used on uneven terrain, the SLAM used is also required to have high performance on uneven terrain. Therefore, in this study, we evaluated the performance of SLAM, which is currently proposed, on uneven terrain. In this study, four SLAMs were evaluated. The evaluation experiment is performed on a simulation, and the robot moves in a total of 19 patterns of environment. SLAM is evaluated using the error between the created map and the reference environment. From the results of the experiment, the accuracy of the uneven environment was lower than that of the flat environment. Karto SLAM showed the best performance in most environments. In addition, Hector SLAM constructed a map with stable accuracy.