Robust localization method based on free-space observation model using 3D-map

This paper proposes robust localization method using a free-space observation model with a 3D map for a particle filter. The free-space observation model judges whether the free space of a laser beam of a sensor is appropriate. The observation model realizes highly robust localization in outdoor environments with many unknown obstacles. This paper describes the free-space observation model and the increase in the calculation speed of the observation model. The proposed localization method achieves 1-km localization in outdoor environments with many unknown objects.