I. Introduction
In recent years, autonomous mobile robots have been used in real-world environments to replace guidance, transportation, and other tasks to reduce the number of personnel. Wheeled autonomous robots are expected to be used in artificial environments with few uneven surfaces, such as sidewalks and parks. However, even on sidewalks, there are uneven surfaces such as Braille blocks and gutter grating. When the robot traverses such uneven surfaces, depending on the uneven surface and the direction in which the robot travels, the swaying of the robot's body may increase, causing distortion of the path and errors in wheel odometry. Therefore, the robot is required to detect and estimate the shape of the uneven surface, and to pass over the uneven surface appropriately so that the sway of the vehicle body is minimized.