Realization of the Cognitive Architecture LIDA in the Building Automation Domain

Abstract

In the building automation domain the maximum comfort wants to be achieved; however energy efficiency has to be maximized, and both of them cannot be optimized at the same time. Cognitive agents allow solving multi-goal problems like this, having a background motivation acting in order to fulfill its given goal. A building automation agent that controls an air conditioning system will act according to these two goals and take the proper decisions to obtain the desired conditions. Even though there are several cognitive architectures, there is not yet a framework for a building automation cognitive agent. In this thesis, the LIDA architecture will be tested in a building automation problem in order to say whether it is viable for this purpose. Here, it is shown a basic implementation of a LIDA-based cognitive agent in a building automation problem. Some experiments have been made to show the results on different scenarios and prove the viability of LIDA in the building automation domain. It is also shown that some of the characteristics of LIDA complement well with the building automation problems. With these results a more complicated cognitive agent could be developed in the building automation domain. Despite existing other cognitive architectures that could be as viable as LIDA or even more, the good performance of it motivate the development of a more sophisticated software to show an actual implementation of LIDA in building automation.