|Topic:||Random Bin Picking|
|Description:||The new generation of so-called collaborative robots allow the use of small robotic arms without them being isolated from human workers. Such an example of collaborative robot is the YuMi robot, a dual 7-axis arms robot designed for precise manipulation of small parts and available in our laboratory.
For further acceptance of such robots in the industry, some methods and sensor systems have to be developed to allow them to pick parts without the position of the parts being known in advance, just as humans do.
The aim of the project is to implement algorithms for the detection of known parts and their picking out of a bin. Part of the work will consist in determining the best compromise between sensor type/quality and achieved results. Different methodologies will have to be tested and asserted, image feature matching, 3D model matching, machine learning with existing algorithms. The development of small mechanical devices may also be necessary (gripper or other).