Exploring reward shaping in discrete and continuous action spaces: A deep reinforcement learning study on Turtlebot3 | Synapse