Artificial Brains Taking Over Your Home

Research in artificial intelligence probably is as old as computation devices itself. Especially in the last few years the interest in artificial intelligence by software companies grew exponentially. But why is that? Probably because of the opportunities in data analysis and therefore user experience artificial intelligence could create. The possibility of using the large amounts of data that are available on the internet to find user behavior patterns no rule based program could ever detect just is astonishing. But why am I talking about all of this with you?

Well, the truth is that I am just thrilled by thus opportunities that lay ahead waiting to be found and developed to enhance peoples — our — lifes.

Now let me tell you about the opportunity I see in artificial intelligence: Home automation! You heard right. I am definitely not the first one with the idea of applying machine learning concepts to home automation. As a matter of fact Michael C Mozer already wrote about this topic in 1998 way before machine learning was the next big thing. Now 18 years later I tried to use the insights researchers have made in the recent time to implement such a system as an iOS application to show that it could be the future of home automation.

The roadmap for the project was very clear from the beginning. The system should use sensor data from any kind of third party home automation system that implements a basic protocol. And consequently use this data to learn the users behavior and therefore save the user from having to fiddle with these different systems to configure them to his beliefs — over and over again for every change in his daily live.

Special thanks also have to go out to Bernard Wessels and the techlab for supporting and working with me on this project. If you do not know techlab or the enera project, you should definitely check out their website and follow them on twitter to stay informed on the latest and greatest in renewable energy and home automation.

If you are interested in the work that has already been done to create a proof of concept you are invited to further look into my research results on Github: