Reinforcement Learning pro ovládání robotů

Abstract

The main focus of this thesis is to design and test methods for training robots using a selected simulation environment. For this purpose, the CoppeliaSim environment, together with the PyRep add-on, has been chosen. The specific task is to use reinforcement learning methods to teach a six-legged robot to navigate to a target on different types of terrain, for which the TD3 algorithm is employed.