Abstract
We propose a robust self-localization method against seasonal environmental changes. In environments where plants are widespread, high-precision maps used for self-localization with scan matching using LiDAR can become outdated due to seasonal variations of plants, resulting in deterioration of the accuracy of self-localization. In this study, we propose a method of creating a high-precision map containing artificial structures that do not change with seasons and self-localization algorithm using the map. The experimental results show that the accuracy of self-localization is maintained through different seasons by the proposed method.
