Abstract
Reinforcement learning is a behavior learning method by which autonomous agents acquire action rules to adapt to unknown environments. The objective of the agent is to maximize the sum of the received reward. Many studies in reinforcement learning have used only computer simulations in their experiments. In this study, reinforcement learning is applied to a six-legged robot with CPG (Central Pattern Generator) in the real environment. The objective for this robot is to acquire the efficient forward movement. We aim to realize that a six-legged robot with CPG acquires the efficient walk patterns by reinforcement learning.
