Architecture for autonomous navigation of inspection robots on energy underground distribution cables

Abstract

This work presents an architecture developed to navigation system in real and simulated robots designed to inspect underground pipes with energy distribution cables. The platform based on this architecture permits to develop and validate the navigation system in a simulation environment and, subsequently, to test and validate in the real environment. In this sense, the platform includes aspects related with sensors simulation, computational vision and planning applied in this kind of environment. The computation platform integrates Hough Transform and Artificial Neural Networks to detect obstacles and infer the best action in the environment.