Deep Reinforcement Learning for Autonomous Navigation: Sim-to-Real Transfer on TurtleBots

Deep Reinforcement Learning (Deep-RL) has emerged as a powerful paradigm for enabling autonomous agents to learn complex behaviors in dynamic environments. Despite its significant advancements and applications in robotics, Deep-RL faces substantial challenges when transitioning from simulation to real-world deployment, due to limited resource availability and the large amount of data required for training. To address these issues, this paper evaluates three state-of-the-art continuous control Deep-RL algorithms in the context of autonomous navigation tasks. A structured experimental methodology is used, progressing from high-fidelity simulations to real-world experiments. This work involves more than 120 hours of real-world experiments and shows evidence of a possible gap between on-paper performance and real-world performance of Deep-RL algorithms due to their different computational requirements and assumptions.