An approach to restaurant service robot SLAM

Recently service robots are becoming a part of our daily life. There is a huge market for robot since many human friendly service robots are designed. In this paper we describe a system of Simultaneous Localization and Mapping(SLAM) on restaurant service robot using a depth camera and a move base with odometry and gyro. We create a virtual laser scan from the depth camera, which provides depth image to be converted into 2D laser scan. Meanwhile, the color image provided by the depth camera could be used to extract feature descriptions which will be used in the loop closure process. The depth camera act as a low cost laser scanner and a RGB camera in this case. Our method is under particle filter frame. The robot builds map incrementally when exploring new environment. This make it can accurately locate itself in the working environment and complete routine navigation tasks autonomously. We will introduce the fundamental mathematical model and conduct experiments to show it.