Combination of Deep Reinforcement Learning and Pure-Pursuit Algorithm for Adaptive Autonomous Mobile Robot Navigation

In the context of real-world environments, mobile robots commonly encounter the challenge of accurately tracking desired paths while navigating through complex terrains. The ability to efficiently track paths holds significant importance for various tasks, including autonomous exploration, object manipulation, and precise localization. This paper proposes a novel approach that combines the Deep Q-Network (DQN) and the pure pursuit algorithm (PPA) to enhance the navigation capabilities of mobile robots. The integration of DQN and the pure pursuit algorithm offers complementary advantages. DQN enables robots to learn complex decision-making policies directly from raw sensor data, allowing them to adapt to dynamic and uncertain environments. On the other hand, the pure pursuit algorithm provides a well-established method for controlling the motion of the robot along a desired path, ensuring accurate and smooth trajectory tracking. The effectiveness of our approach is assessed by conducting extensive simulations and real-world experiments. The results demonstrate the effectiveness of combining the DQN algorithm with the pure pursuit method in enhancing the navigation capabilities of differential-drive mobile robots.