Regresar

Objects detection techniques and deep neural networks for autonomous navigation of robot youbot KUKA with signaling

Abstract:

The processes of exploration and autonomous navigation implemented in ROS allow to generate routes depending on the space that the environment presents, however, it does not take decisions based on specific objects that are in the space since it does not incorporate by default a system of vision by computer that allows it to obtain data. As for the current computer vision systems, there are object detectors based on neural networks, which provide a fast and precise detection of specific objects which allows a better interaction with the environment surrounding the system. The objective of this research is to link the two processes mentioned above, in order to detect objects in the environment where the robot is located and modify the decision making in the route to be taken.