This project examines the dynamics of a livestock production house and the requirements for the poultry inside. The goal is to design an MIMO controller capable of keeping the climate inside the production house at references specified by the farmer. For this project the internal temperature is defined as the most important state and it should be kept at reference with a 2°C margin of error.

The project is made in collaboration with SKOV A/S, who has supplied data from a production house located at Mosegaarden as well as providing an Air Physical Laboratory to be used as a test facility for the MIMO controller.

A non-linear model is derived for the production house, where data from the Mosegaarden and from the laboratory is used for parameter estimation. Using the model an Extended Kalman Filter (EKF) is designed and used as an state observer. For the controller an Linear Quadratic Regulator (LQR) and Linear Quadratic Integrator (LQI) is designed.

Testing in the laboratory resulted in a successful test, where the LQI kept the temperature at reference inside the 2°C margin of error.