Vision based control and simulation of a spherical rolling robot based on ROS and Gazebo

This paper presents a process of vision based controller implementation in order to control a spherical rolling robot. A Linear Quadratic Gaussian (LQG) controller is applied as a complementary controller for the proposed spherical rolling robot. The LQG controller is used to synchronize the robot's motors. The simulation results, which are obtained from a Co-simulation between MSC ADAMS and MATLAB software validated the prosperity of the proposed LQG controller in term of synchronizing the aforesaid robot's motors, which lead the robot to have a better performance in path tracking and decreasing the deviation rate from the desired path. In order to find a short path between the obstacles, which are set in the robot's workspace, a potential field algorithm is designed and implemented on the robot workspace in Gazebo simulator. As stated by the simulation results, the performance of the spherical rolling robot under study in terms of autonomous path tracking is considerably improved. The mentioned performance is considered according to the vision control which is done in Gazebo simulator. Furthermore, the simulation results corroborate that using the LQG and vision based controllers simultaneously, improved the performance of the aforesaid robot significantly in terms of steering and obstacle avoidance.