Using computer gaming models to understand the behavior of industrial machines

The exponential growth of audiovisual and textual data is aggressively leading to the creation of tools and systems capable of selecting and recognizing given patterns of information. Such urge is at the basis of the development of machine learning techniques, techniques that are today employed to analyze all kinds of information. Computer gaming is, with no doubt, one of the fields where such techniques have been put to good use most pervasively. For example, in order to provide better opponents in human vs. machine games, machine learning techniques are being extensively utilized to model how players behave during gaming sessions. These same ideas, however, could be put to good use when realizing completely different systems and applications, as for example when analyzing how industrial machines behave while running complex tasks. We here show that techniques that have been well accepted and deployed through entertainment technologies can also be applied to industrial ones. In particular, we will show how it is possible to model the operation of an industrial machine as a game in order to assess whether the machine is following its regular behavior path. The result of this study is that it is possible to model also complex machines by patterns where divergences from such patterns can represent indicators of malfunctioning or unexpected working states.