Predicting and preventing mistakes in human-robot collaborative assembly

Abstract

The human-robot collaboration (HRC) in industrial assembly cells leads to great benefits by combining the flexibility of human worker with the accuracy and strength of robot. On the other hand, collaborative works between such different operators can generate risks and faults unknown in current industrial processes, either manual or automatic. To fully exploit the new collaborative paradigm, it is therefore essential to identify these risks before the collaborative robots are introduced in industry and start working together with humans. In the present study the authors analyze a benchmark set of general assembly tasks performed by HRC in a laboratory environment. The analyses are executed with the use of an adapted Process Failure Mode and Effects Analysis (PFMEA) to identify potential mistakes which can be made by human operator and robot. The outcomes are employed to define proper mistake proofing methods to be applied in the HRC assembly work cell.