A Hybrid Feature based Mapping for Indoor Service Robots

Abstract

In this paper, we present a hybrid feature based mapping for robots operating in indoor environments. We combine environment features such as line segments and corners to construct 2D map for the purpose of robot navigation. Our method uses an extended Kalman filter algorithm for fusing the hybrid features into a single feature map. The proposed method works very well in different indoor scenarios with varying level of obstacles and provide feature rich map for the purpose of autonomous robot navigation. The algorithms in the present work is tested in simulation and real environments.