Digital twin for adaptation of robots’ behavior in flexible robotic assembly lines

Product personalization is the global trend driving the development of modern production systems. This in turn creates the need to deploy flexible and reconfigurable assembly systems. New type of factories exploiting the capabilities of multiple resources such as human operators and mobile and / or stationary robot assistants are emerging. This paper investigates the use of digital world modelling techniques in such hybrid production systems for enabling system reconfiguration through shared environment and process perception. The suggested digital world model infrastructure involves three main functionalities: a) Virtual representation of the shopfloor, combining multiple sensor data and CAD models. The digital shopfloor is rendered in the 3D environment exploiting the capabilities provided by Robot Operating System (ROS) framework, b) Semantic representation of the world through the implementation of a unified data model for representing the geometrical as well as the workload state, and c) Dynamic update of the digital twin based on real time sensor and resource data coming from the actual shopfloor. The communication and integration layer among the physical and the virtual agents is realized on top of the ROS framework. The developed infrastructure is applied to an industrial case study from the automotive industry focusing on a car’s front axle assembly. Nevertheless, it is designed to be generic enough to be able to address multiple cases.