Visual Guidance of an Autonomous Robot Using Machine Learning

The aim of this work is to create a method to compute the steer direction of an autonomous robot, moving in a road-like environment. It uses artificial neural networks to learn behaviours based on examples from human drivers, replicating and sometimes even improving human-like behaviours. Artificial neural networks perform very fast computations, which make them well suited for real-time navigation. They also have the possibility to perceive information that was undetected by humans and therefore could not be coded in classical programs, improving the overall performance for the steering of an autonomous mobile robot.