A recent main issue in home appliance production is constituted by the so called “smart device”, i.e. a device with a high level of self-adaptability to different operational conditions. For example, one would like to design washing machines which are able to switch between full-loaded and half-loaded program without any human interaction, or ovens which are able to decide autonomously the cooking time of a certain dish without weighing the food.
The goal of this project is to develop smart devices without excessive modifications of existing home appliance. In this respect, smart devices should make decisions based on the trend of quantities (drum speed in the washing machine, temperature profile in the oven, etc.) which are already measured in existing devices.
In this context, it is a difficult evaluation task to retrieve relevant information, since the relation between unknown and measured quantities is typically implicit and nonlinear. Ad hoc black-box identification techniques are thus required.
A new two-stage identification technique has been worked out for this purpose. Theoretical investigations as well as validation by simulation are completed; the validation by means of experimental data is currently underway.