ビジュアルオドメトリを用いた屋外二次元移動時の自己位置推定精度検証

抄録

In recent years, infrastructure facilities have been noticeably deteriorating and need to be inspected effectively. On the other hand, the number of workers who inspect them has been decreasing in recent years. In particular, a ground penetrating radar, which is manually moved and measured, is used for inspection of underground buried infrastructure. This work is very time-consuming. Therefore, autonomous mobile robots are required to improve the efficiency of inspection and saving labor. High-accurate self-position estimation is required for autonomous movement. Moreover, high consistency between the position and the position of underground buried objects obtained from the ground penetrating radar is required. In order to automate the inspection of underground infrastructure, which is currently done by pushing carts, it is necessary to improve the accuracy of self-localization of robots. In this study, we verified the accuracy of the self-localization method using a road camera when moving outdoors two-dimensionally long distance.