System of Robotic Systems for Crack Predictive Maintenance

摘要

In the built heritage, maintenance represents a not insignificant part of the costs of restoration and repair. Especially if this maintenance is carried out aperiodically. In this paper, we discuss a concept of System of Systems Engineering (SoSE), based on autonomous and collaborative robotization techniques. The latter is intended to detect, localise and repair microscopic cracks on flat wall surfaces. This was made possible by collaborating between several operational and managerial independent drones, one of which is intended to automatically detect cracks using the deep learning technique offered nowadays by Artificial Intelligence (AI). A second flying-manipulator robot, communicating and collaborative, manages to localise the cracks and deposit material continuously on the pruning shears of the crack. The whole concept works in a collaborative way and with an evolutionary development, in the presence of hazards and faults on one of the robotic system components. Experiments demonstrate the passage from the theory, in terms of architecture of the SoSE concept, as well as its robustness.