Loading...
RRT and Octomap based 3D environment exploration and mapping
Koirala, Koshish
Koirala, Koshish
Description
A Master of Science thesis in Mechatronics Engineering by Koshish Koirala entitled, āRRT and Octomap Based 3D Environment Exploration and Mappingā, submitted in May 2020. Thesis advisor is Dr. Shayok Mukhopadhyay. Soft copy is available (Thesis, Approval Signatures, Completion Certificate, and AUS Archives Consent Form).
Abstract
For autonomous navigation in an unknown environment, all robots need a map of the environment along with the robot's current location on the map. This work builds on existing rapidly-exploring random tree (RRT)-based exploration techniques and considers the mapping of three dimensional (3D) space using an Octomap. There have been previous efforts using multiple RRTs to produce map of 2D space. This work aims to utilize such efforts along with a 2D scanner mounted on a servo actuated tilting joint, to effectively generate a map of 3D space. The Gazebo simulator, present in the Robot Operating System (ROS), is used for simulations whereas the ROS Visualizer (RVIZ) is used for visualization. This work presents a 3D mapping strategy based on Octomaps and OctTrees whereas 2D Simultaneous Localization and Mapping (SLAM) running on ROS based Gmapping techniques provides robot localization information. A cusĀtom servo-based tilting platform has been designed and used for scanning 3D spaces with a traditional 2D laser scanner. The appropriate frame-transformations required to integrate such data onto the frame of a map being explored, are also shown. Further, the effects of running the proposed RRT based exploration algorithm with only a global detector, are compared explicitly with the case of using a local detector along with the global detector, and the improvements are summarized. Several experimental runs, and their results are presented to show the effectiveness of the proposed map exploration strategy.