Reinforcement Learning in Robotics: from Theory to Real-World Applications

EasyChair Preprint 12537

Abstract

Reinforcement learning (RL) has emerged as a powerful framework for training autonomous robotic systems to perform complex tasks in real-world environments. This paper provides an overview of RL techniques and their application to robotics, spanning from theoretical foundations to practical implementations. We discuss key concepts in RL, including value functions, policy optimization, and exploration-exploitation trade-offs, and explore how these techniques can be adapted to robotic control problems. Furthermore, we review recent advancements in RL algorithms, such as deep reinforcement learning (DRL), and discuss their implications for robotics. Finally, we highlight real-world applications of RL in robotics, ranging from manipulation and navigation tasks to autonomous driving and robot-assisted surgery. Through a comprehensive analysis, this paper aims to provide insights into the potential of RL for advancing the capabilities of robotic systems in diverse application domains.

Keyphrases: Reinforcement Learning, Robotics, autonomous systems, policy optimization, value functions

@booklet{EasyChair:12537,
  author    = {Lia Don},
  title     = {Reinforcement Learning in Robotics: from Theory to Real-World Applications},
  howpublished = {EasyChair Preprint 12537},
  year      = {EasyChair, 2024}}