Mobile Robot Localization and Path Planning in a Picking Robot System Using Kinect Camera in Partially Known Environment

Motion planning, such as mobile manipulation, must not only be able to move safety through their environments, but also be able to manipulate objects in their environments. This paper proposes localization and path planning for a mobile robot (MR) in a picking robot system using Kinect camera in a partially known environment. A Kinect RGB and depth camera is utilized to recognize objects for obstacle avoidance and path planning. To do this task, the followings are done. Firstly, the system configuration of MR is described. Secondly, mathematical kinematic modeling of the MR in the picking robot system is presented. Thirdly, an object detection algorithm based on the Kinect camera is proposed to detect landmark. Extended Kalman Filter (EKF) is used to get the best localization estimation of the MR by combining the encoder positioning result and landmark positions obtained from the Kinect camera. Fourth, D* Lite algorithm is used to generate a path from the start point to the goal point for MR and to avoid unknown obstacles using information obtained from Kinect camera. A controller design for the MR to track the trajectory generated by the D* algorithm based on backstepping method is proposed. Finally, the effectiveness of the proposed algorithms and controller is verified by using experiment. The experimental results show that the MR successfully reaches the goal point with an acceptable small error.