Balancing Robot Navigation with Virtual Map and Virtual Sensor

Abstract

The need for autonomous robots has recently increased, along with the rapid development of robotics technology. The ability to navigate without the need of human intervention, is one of the advantages of autonomous robots. To make an autonomous robot, a robot with high efficiency, flexibility, and a reliable navigation system are needed. High efficiency and flexibility can be handled with the use of a two-wheeled self-balancing robot. A reliable navigation system can be achieved by using the ROS (robot operating system) platform. The research produces virtual robots, virtual maps, and virtual sensors on simulations in the GAZEBO application, which can be visualized in RVIZ. Map development and navigation system usage, run in the ROS system, producing speed data that sent to GAZEBO simulations and balancing robot in the real world. In this paper, the experiments are divide into two, simulation and real, with two different destinations coordinates, straight (1.0, 0.0) and curved (2.2, -1.0). The tracking simulation test shows that the virtual robot can reach the first destination, with errors averages -0.084 m on X-axis and -0.01 m on Y-axis. The second destination gives error averages -0.052 m on X-axis and -0.05 m on Y-axis. The real tracking test shows the balancing robot can receive speed data from the ROS system, to move towards the destination point based on the virtual map and virtual sensor. The real tracking test gives an error averages 0.046 m on X-axis and 0.02 m on Y-axis, in the first destination. On the second destination, the error averages are 0.044 m on X-axis and 0.38 m on Y-axis. The experiments show that the robot can go to the destination point autonomously with virtual map and virtual sensor.