Corner extraction from indoor environment for mobile robot mapping

In order to construct more accurate map of the environment, it is essential to extract the key features of environment using appropriate sensor. Sensor plays a fundamental role in the detection of key features. For example to extract the corners from the mobile robot´s indoor environment sonar sensor is not appropriate due to Specular reflections and for glass like features laser sensor will provide erroneous result but for corners it will provide more accurate and reliable results. In complex indoor environment identification of concave and convex corner not only reduces the navigational complexity but also reduces the decision time for autonomous mobile robots. This paper addresses the new architecture for corner feature extraction from indoor environment. Second Section of the paper deals with the filtration of the raw information and perform clustering of the laser range information based on the resolution of laser range finder and concerned environment. Third section deals with identification of concave and convex corners in the environment using the laser range finder. Experimentation carried out shows 6.5% improvement in corners detection for the indoor environment as compared to the conventional approach. The computational time is also reduced up to reasonable extent.