Predictive Analytics in Robotic Industry

Modern breakthroughs have enabled the massive use of automation for addressing the ever changing automation market demands. To address an ever changing market and volatile requirements, every production process should be, if possible, optimised, reducing costs and wastes. At the same time, reconfiguration enabling flexibility through software automation and, in turn, resiliency is much pursued in the context of advanced manufacturing. One sector where automation is widely and deeply applied is the automotive one. For many years, the number of cars globally sold per year is above the 50 million at production rate up to 60 jobs (i.e., cars) per hour and even more. Therefore, even a rare and unexpected production stop has a significant financial impact in terms of production losses for the manufacturer. That is why predictive maintenance is the key aim of the SERENA project. The purpose of this chapter is to present the SERENA approach in the context of a robotic test-bed related to the automotive sector. As described in this chapter, the approach proposed is able to predict the Remaining Useful Life of an industrial robot key component, propagate that to the maintenance schedule system and the visualisation tool (including augmented and virtual reality maintenance procedures as well). The entire software system is designed to run on a containerised infrastructure, providing thus flexibility, scalability and fault tolerance.