Automated Guided Vehicle Collision Avoidance with Reinforcement Learning

Abstract

Industry 4.0 introduces a new approach to manufacturing, focusing on connecting and automating machines. Automated vehicles that can operate in a dynamic environment, will enable the move towards Industry 4.0 to require less manual effort. The approach will be to create a simulation of an environment, with several Automated Guided Vehicles, that use a reinforcement learning system to avoid collisions, and continuously learn and adapt to changes in the environment. The expected result would be agents that are able to adapt and learn in the environment.