Vision-less Bin-Picking for Small Parts Feeding

摘要

Part feeding remains a vital yet troublesome task. This is especially true for flexible manufacturing systems, where classical methods are less applicable due to constraints from small batch sizes and high part variance. In this paper, a novel approach to part feeding is presented, which is targeted at small parts, such as nuts and bolts, that commonly occur in assembly tasks. The approach is an alternative bin picking solution, where parts stored unordered in bins, are oriented and presented for further manipulation using a specially designed tool. The tool is manifested a scoop, which a robot can use to scoop parts from the bins without the use of external sensory input such as vision. Using solely mechanical orienting principles and the motion of the robot, the parts are sorted so only parts in a desired orientation remain in the scoop. The approach is tested on a number of different parts to benchmark its performance.