ROS navigation STACK on the B21r mobile robot.

Abstract

Perception is an imperative task for a mobile robot to be able to navigate autonomously inits environment. The robot can perceive its surrounding through one or more sensors; then, create a representative map of the environment and locate itself on it in real time. The propose of this project is to build a surveillance strategy to enable a mobile robot to navigate and supervise an interior environment. It consists in the implementation of a perception techniques, localization and navigation strategy on a surveillance robot. Hence, this project deals with ROS Navigation Stack which is powerful for mobile robots to move from place to place. We worked with B21r mobile robot of CDTA equipped with light detection and ranging (LIDAR) sensor and D435i depth camera. First, our work consists of building a map of the hall of CDTA (the Advanced Technologies Development Center) using Lidarsensor and Simultaneous Localization and Mapping techniques. Then, we localized our robot inside that map using AMCL, therefore, our B21r mobile robot is well localized. Second, we created a path so that our robot reaches its goal or target and creates a new path in case the robot faces obstacles that were not present in the old map. However, our robot was not able to correctly deal with its last task consisting in tracking the planned path; even that, its position is updated enabling it to achieve its desired destination, due to mechanical problem of robot wheel. The work is implemented using ROS melodic and Rviz tool.