Deep Reinforcement Learning for Mobile Robots: Overview and Issues

This study offers a comprehensive overview of the integration and impact of Deep Reinforcement Learning (DRL) in the field of mobile robotics. Focusing on the period from 2000 to 2023, The evolution and influence of DRL are analyzed through a detailed citation analysis, leveraging data from major scholarly databases. The core concepts of Reinforcement Learning (RL) are explored, emphasizing its application in mobile robotics for balancing exploration and exploitation, a critical aspect of autonomous decision-making processes. The paper delves into various applications of DRL in mobile robotics, including autonomous navigation, path planning, and target tracking, highlighting the advancements that have significantly enhanced robotic capabilities. A review of recent developments is also included in the analysis, such as Successor Representations and latent imagination learning strategies, illustrating their contribution to the exploration-exploitation trade-off. Additionally, the advantages of DRL, such as adaptability and learning from experience, alongside challenges like hardware limitations and dynamic environment adaptability. The paper concludes by identifying key future research directions, notably in generalization and transfer learning, robustness to uncertain environments, and the potential role of Graph Neural Networks (GNNs) in further advancing RL for mobile robotics. This overview aims to provide a clear understanding of the current state and future potential of DRL in enhancing the efficiency and intelligence of mobile robotic systems.