Automotive research institute IAV GmbH and the German Research Center for Artificial Intelligence (DFKI) are joining forces to open the Research Laboratory for Learning from Test Data (FLaP) in Kaiserslautern. The research carried out here will initially focus on powertrains – on engine-control, to be precise. Modern engine control units (ECUs) are today capable of controlling over 50,000 parameters for injection valves, turbochargers, catalytic converters, etc. through on-board and self-diagnosis systems – in other words: all settings that ultimately determine engine performance, consumption, and wear. Until now, the control unit has always been set during production, normally never to be adjusted again.
Now, though, AI systems are set to leverage machine learning to record consumption, wear, and a whole host of other data – and to adjust the settings accordingly based on this data. Interesting ethical consequence: AI will in future have to pay the price for variable exhaust emissions. The data collected should also help better predict wear. At the same time, the collaboration also wants to trial visualization options, which should in turn make data more manageable for future product developments.