An Industrial Intelligent Autonomous Recycling System using a Deep Learning Approach

In this paper, an industrial intelligent autonomous recycling system is proposed for sorting recycling objects accord- ing to their own materials using deep learning. In order to adapt the system to be used in the real environment, movable robotic arm equipped with multifunctional grasping tool and mobile platform are integrated. The system structure is composed of four functionalities that are 1) object perception, 2) object clas- sification, 3) motion planning, and 4) navigation. To demonstrate the viability of the system, results for grasping, classification, and navigation are evaluated in the new industrial platform.