The application concerns a collaborative assembly station, in which ABB's Yumi robot and an operator perform synergistically pre-assembly operations of a thermostatic valve head. In Man-robot collaboration, the presence of a unit dedicated to detecting not only the presence, but above all the operator's intent, is essential. The cognitive perception device produced by Smart Robots was installed to equip the robot with artificial intelligence giving it the faculty to choose which phases of the processing to carry out, according to the recognition and prediction of the operator’s actions. Thanks to the artificial intelligence algorithms developed by Politecnico di Milano, the collaborative robot is able to modify in line its work cycle and adapt itself to the operator. Finally, the activities entrusted to man can be supported by a wearable augmented reality device that shows, through holographic projections, the assembly instructions or the next pieces to be recovered and their position.
The recognition of operator movements makes the system immune from propagation of any errors in the assembly cycle. The system is in fact able to recognize immediately, or even in a predictive way, an operator’s error in the sequence of assembly operations.
The transformation of the workstation, previously completely manual, in the collaborative mode made possible a reduction of the cycle time and enabled the application to respond more efficiently to the fluctuations of the takt-time. In a parallel study and presented at the Incom 2018 (IFAC Symposium on Information Control Problems in Manufacturing) the same algorithms showed a reduction in cycle time from 10 to 17%.
Compared to a fully automated solution, the collaborative solution made it possible to maintain substantially unchanged the foot-print of the station (2 sqm). The configuration times and, above all, the application reconfigurations were minimized, thanks to the possibility of making a format change or a change to the layout in a few minutes through the recognition of QR codes in the various areas used by the operator.
Finally, the use of augmented reality greatly speeds up the process of learning the job of the operator, especially in case of reduced experience or in situations of continuous job rotations, thus facilitating the commissioning of policies to reduce alienation of the worker.