Deep-Reinforcement-Learning-Based Semantic Navigation of Mobile Robots in Dynamic Environments

Mobile robots have gained increased importance within industrial tasks such\nas commissioning, delivery or operation in hazardous environments. The ability\nto autonomously navigate safely especially within dynamic environments, is\nparamount in industrial mobile robotics. Current navigation methods depend on\npreexisting static maps and are error-prone in dynamic environments.\nFurthermore, for safety reasons, they often rely on hand-crafted safety\nguidelines, which makes the system less flexible and slow. Visual based\nnavigation and high level semantics bear the potential to enhance the safety of\npath planing by creating links the agent can reason about for a more flexible\nnavigation. On this account, we propose a reinforcement learning based local\nnavigation system which learns navigation behavior based solely on visual\nobservations to cope with highly dynamic environments. Therefore, we develop a\nsimple yet efficient simulator - ARENA2D - which is able to generate highly\nrandomized training environments and provide semantic information to train our\nagent. We demonstrate enhanced results in terms of safety and robustness over a\ntraditional baseline approach based on the dynamic window approach.\n