Text this: An Improved Dynamic Window Approach Based on Reinforcement Learning for the Trajectory Planning of Automated Guided Vehicles