Monocular camera-based go-to-goal control of holonomic robots

Visual servoing in autonomous mobile robots has received significant interest in recent times as navigation of the robots can be performed without the use of external sensors, such as GPS or beacons. A commonly performed task by such robots is to move to target locations in their working environment. By fixing these locations with unique well-known visual markers, such as a “plus” symbol in a helipad, using on-board cameras to identify these visual markers, and on-board hardware to process the images, the robots can be made to navigate to them. In this paper, we propose a simple and efficient vision-based “go-to-goal” (GtG) algorithm for a holonomic mobile robot equipped with a single monocular camera. The goal is identified by a marker with known dimensions and is calibrated for the camera. With this information, the GtG algorithm enables the robot to move to this marker from any position in the environment. Simulation results, derived from experiments performed in the ROS environment, demonstrate its operation and features.