Virtual Commissioning of 6 DoF Pose Estimation and Robotic Bin Picking Systems for Industrial Parts

Abstract

The robotic handling of unsorted parts, also known as bin picking, is still a challenge in industrial automation. Many systems require physical testing and training, resulting in a long commissioning phase. In this paper, a concept for the testing and performance evaluation of bin picking systems based on industrial simulation software is introduced. The system utilizes physics simulation, virtual stereographic image rendering and kinematics as well as robot controller simulation. This enables a virtual software/hardware-in-the-loop verification and optimization of bin picking system performance. The concept is validated on major bin picking sub processes in a real-world industrial scenario in electronics production. It is found that the system correctly maps major effects influencing real bin picking systems. After concluding, further application fields for this new simulation concept are discussed.