Rivet inspection with multi-sensor robotic system

Rivet inspection is a very common task that represents a key activity in many industrial fields ranging from aerospace to civil infrastructure maintenance. Being a repetitive and standardised operation, rivet inspection is a good candidate for the automated process. On the other hand, since a high level of precision and accuracy is required so the human intervention is still predominant. This paper describes a robotic system that is able to automatically perform this kind of inspection and can indicate the status of rivets, thanks to a novel multi-sensor analysis method. The detection system, that is located on a crawler robot, consists of a camera, a laser scan and a micro-hammering system to detect the acoustic response of an exciting mechanical pulse. The collected data are processed and an analysis of the rivets's status is performed. The work, here reported, illustrates the details of the above described method and the experimental results obtained with a preliminary prototype used as a proof of concept.