Building Intelligence Into Intelligent Buildings

Research into intelligent environments and smart houses tends to focus on dedicated applications or general infrastructures for building intelligent environments. There have been few projects that considered how ordinary people would be able to configure and program the intelligent environments that they inhabit. There are many jokes about the inability of ordinary people to program their VCR and if we are not careful, similar jokes may be told in the near future about people who are unable to manage the smart houses that they live in. It is therefore important to develop programming abstractions that allow people to adapt the intelligent environment to their current needs. Ideally, smart houses should go beyond programming and become really intelligent environments, i.e., they should observe the behaviour of their inhabitants and use these observations to adapt the environment to satisfy the anticipated future needs of the inhabitants.

 
The purpose of the BIIIB project (pronounced "B3B") is to investigate a combination of sensor networks and artificial intelligence techniques, which allows the smart house to observe and identify patterns in the user's behaviour. Whenever a user requests a particular action from the smart house, e.g., switching on the lights in a particular room, the intelligent environment will attempt to identify a unique pattern of preceding sensor events, which will allow the smart house to automatically switch on the lights for the user in the future. This means that existing installations, such as light switches, define the API for the smart house and that the user programs her smart house simply by living in it. The project will examine the applicability of different technologies for this purpose, including sensors, sensor networks and artificial intelligence techniques. The BIIIB project will develop simple prototypes that demonstrate the feasibility of the techniques and technologies developed in the project.