### Page Content

### to Navigation

# Autonomous use of real-time data

Autonomous decision making for smart buildings is fundamentally based on a mathematical description of the building and its interactions, which can be used to compute predictions and hence also optimal decisions. Those mathematical models can be obtained using physical knowledge, machine learning techniques or even psychological models that describe human behavior – and combinations of all of them.

Regardless of the method used to generate a model, for the success of smart buildings it is necessary that the developed methods for optimal operation can be adapted according to the real-time data that is obtained. Such adaptation is necessary because systems evolve over time, uncertainty in the form of weather or human interaction changes, and plug loads also vary significantly with time. Parameter estimation and machine learning techniques will be used to perform a continuous adaptation, which is mandatory for the adoption of smart buildings approaches.