12. CTI-Jahrestagung


im Gesamtfahrzeugkonzept CTI - Car Training Institute Jahrestagung10. und 11. Oktober 2019, Köln

Artificial intelligence for personalized climate control

Dr. Ing. Christine Junior, IAV GmbH

Using an innovative operating concept developed by IAV for automatic climate control in vehicles, the number of user inputs is to be reduced to an absolute minimum. To achieve this goal, the preferences of users for different environmental conditions must be known. From the input of the user regarding his thermal comfort, it is possible to record these preferences or to draw corresponding conclusions.

In order to personalize and individualize climate control, adaptive algorithms can be used. These algorithms are able to learn from user input and lead to the desired degree of individualization. In a first step, IAV used the regression of specific characteristic curves in combination with the HMI concept mentioned above in an automatic climate control system. In addition, initial results with KI approaches were obtained.

A comparison of the methods reveals decisive advantages of the use of artificial intelligence. It lies in the fact that a manual production of connections of different information, which characterize an environmental situation, is void, since this is already covered by the algorithm. In addition, there is no need to select one or more suitable functions to describe the characteristic curves to be adapted.

Furthermore, it is not necessary for the automatic climate control to be adapted in such a way that the applied characteristic curves are varied subsequently. Accordingly, the software does not need to be equipped with the appropriate capabilities. The Q-Table learning method can be used as an additional module with only minor changes to existing climate software and achieves a strong adaptation performance after only a short time or a few user inputs.