Helping a Robot to be “Autonomous”: The Expertise of a (Human) Roboticist in a Manufacturing Plant

The creation of robotic systems demands the formalization of how-to-do rules. However, professional workers interact with the world in a way that goes beyond formal rules, such as when facing unforeseen and context-dependent events. The solution to replace human tasks by robotic systems consists in the creation of “micro-worlds”, which presuppose controlled environments, with fixed rules, in which robots are able to operate successfully. Accordingly, such micro-worlds must be designed, built, supervised, maintained and optimation by (human) roboticists who make sense of what robots must and must not do. To address this interaction between them, this article discusses cases that show how experienced workers use their perceptual skills to anticipate and solve problems on the robots under their supervision. Through the analysis of the “course of action” (Theureau, 2004) of real events, its goal is to show how human activity is directed by getting a sense of the situation during the interaction with the machines and how the context influences such sense. As a result, it contributes with Industry 4.0 in its aim to increase automation to maximum power, enabling robots to perform increasingly complex functions. On the other hand, it argues against its assumption that human performance in automated environments is a residual problem to be solved – i.e., eliminated. The actual challenge is how to design micro-worlds that enable and enhance the human-robot integration in the shopfloor for guaranteeing quality, safety and efficiency.