Reinforcement learning in robotics: A survey

TLDR

Reinforcement learning provides robotics with a framework and tools for sophisticated behaviors, while robotic challenges inspire and validate RL advances, making the interdisciplinary link as promising as physics to mathematics. This survey aims to strengthen the connection between robotics and RL by reviewing key challenges and successes in robot behavior generation. The survey examines how RL contributions reduce domain complexity by exploring algorithms, representations, and prior knowledge to achieve successful robot behaviors. The paper emphasizes the trade‑off between model‑based and model‑free, as well as value‑function and policy‑search methods, illustrating their application on a simple problem and highlighting open questions and future potential.

Abstract

Reinforcement learning offers to robotics a framework and set of tools for the design of sophisticated and hard-to-engineer behaviors. Conversely, the challenges of robotic problems provide both inspiration, impact, and validation for developments in reinforcement learning. The relationship between disciplines has sufficient promise to be likened to that between physics and mathematics. In this article, we attempt to strengthen the links between the two research communities by providing a survey of work in reinforcement learning for behavior generation in robots. We highlight both key challenges in robot reinforcement learning as well as notable successes. We discuss how contributions tamed the complexity of the domain and study the role of algorithms, representations, and prior knowledge in achieving these successes. As a result, a particular focus of our paper lies on the choice between model-based and model-free as well as between value-function-based and policy-search methods. By analyzing a simple problem in some detail we demonstrate how reinforcement learning approaches may be profitably applied, and we note throughout open questions and the tremendous potential for future research.