In the fields of urban search and rescue (USAR), it is important that crawler-type mobile robots explore occluded areas (collapsed buildings, underground shopping centers, etc.) in preference to rescue workers from the point of view of safety. To map such an occluded environment, it is important for robots to localize their position and pose. In this paper, we propose two three-dimensional localization algorithms for crawler-type mobile robots. The algorithm is based on "three-dimensional scan-matching" using the three-dimensional laser range finder information. Several experiments using our crawler-type mobile robot verifies the validity and limitation of this method in a simulated disaster environment.