Omnidirectional Mobile Robot Path Finding Using Deep Deterministic Policy Gradient for Real Robot Control

摘要

Recently, workers are in short supply in the distribution industry. Therefore, the objective of this study is to develop an autonomous mobile robot that can search for a path to the goal while avoiding static and dynamic obstacles to support workers in a warehouse. In this study, we apply five learning types in a hybrid static and dynamic environment in a simulation environment as a preparation for learning in an actual machine and verify the effectiveness of these learning methods.