Comparison of Multiple Reinforcement Learning and Deep Reinforcement Learning Methods for the Task Aimed at Achieving the Goal

Reinforcement Learning (RL) and Deep (DRL) methods are a promising approach to solving complex tasks in the real world with physical robots. In this paper, we compare several reinforcement learning (Q-Learning, SARSA) deep (Deep Q-Network, Sarsa) for task aimed at achieving specific goal using robotics arm UR3. The main optimization problem of experiment is find best solution each RL/DRL scenario minimize Euclidean distance accuracy error smooth resulting path by Bézier spline method. simulation word applications controlled Robot Operating System (ROS). environment implemented OpenAI Gym library which uses RVIZ tool Gazebo 3D modeling dynamics kinematics.