Robotic bin-picking system based on a revolving vision system

Abstract

This paper is focused on a complex robotic process called bin-picking. It is a very known and well-researched topic in the field of robotic manipulation where the robot is expected to grasp non-oriented objects from a bin and placed them to locations within the following process flow. Although we present here a functional case study of a bin-picking system based on a modified revolving vision system, it is still a challenge to design such processes to be implemented in industry environment. The main problem is to determine the cycle time and a reliability of a potential implementation. The approach to this problem we have used so far deals with a feasibility testing procedures with real objects in a robotic lab. The apparent drawback is the fact that this procedure is quite time-consuming. Aware of this limitation, we also pay attention to some efforts focused on simulation of even bin-picking systems in a virtual environment: some prospective ideas to cover this issue of a virtual simulation and optimization of robotic processes are included in this paper.