The Augmented Perception: an emerging approach toward resilient manufacturing systems involving robotic agents and digital twin.

This paper addresses the resilience of Industry 5.0 Manufacturing Systems (MS) with mobile robotic agents, focusing on robustness (handling disruptions while maintaining production) and flexibility (adapting to reconfigurations). We propose Augmented Perception (AP), a Digital Twin-based approach that enhances robot perception by integrating virtual elements in the map of the robot.

Three use cases, in simulation and real-world settings, evaluate AP with a mobile robot transporting components. The first tests robustness by introducing an unexpected obstacle, the second assesses the robustness toward dynamic environment, while the third examines flexibility in a reconfigured workshop. Four scenarios are analysed: (1) nominal operation, (2) unforeseen disruption, (3) disruption shared via the Digital Twin (DT), and (4) full workshop remapping. Key performance indicators, Time Spent (TS), Distance Covered (DC) and energy consumption assess AP impact.

Results show that unknown disruptions significantly degrade performance, increasing TS, DC and energy consumption. AP effectively mitigates these effects, restoring near-nominal performance. For flexibility, AP minimizes performance gaps between simulation and real-world execution, demonstrating its value over full, time-intensive remapping.