Hybrid semantic mapping using door information

We describe hybrid semantic mapping method with classified area information for home environments. The hybrid map contains two map types: a grid map, and a classified area information-in-grid (CAIG) map. The grid and CAIG maps can be used for intelligent navigation (e.g., localization and path planning) and motion selection, respectively. In home environments, a door can be used to divide an area into various sections such as a room or a kitchen. Therefore, we use a grid map of the home environment and door information as main clues to classify the area and to build the hybrid map. The proposed method is verified by various experiments. We show that the robot can build a hybrid semantic map autonomously in home environments.