Indoor 3D path planning using a Kinect V2 sensor

Abstract

This paper proposes an indoor three-dimensional path planning system for the automatic control of robots using a Kinect V2 sensor. In the method, a RGBD SLAM system using Kinect V2 sensor is introduced to identity the environment and obtain the indoor three-dimensional map. At the same time, the rapidly exploring random tree (RRT) is used for a safe and optimal navigation of robot in three-dimensional space. We also propose one variation of the RRT to addresses the optimization problem of traditional RRT. In the end, the experiments are conducted with robot operating system (ROS), and a 3D map that can be directly used for navigation of robots is created. The traditional RRT path planning method and the improved RRT method are also presented based on the 3D map.