A Responsible Window

Machine Learning Physical Computing Projects

A Responsible Window

A Responsible Window is a prototype developed during the Trainable User Interface course at Copenhagen Institute of Interaction Design and the concept takes life thanks to Copenhagen itself and his very unpredictable weather that makes hard to balance the temperature inside a room especially when it comes deciding to keep a window open or not, for how long and how much.

It is also true that comfort is not the same for everyone, so the team thought: how might we optimize the control of environmental comfort inside a room, specific for every user, with few simple interactions?

Focusing on the simple gesture of opening or closing a window, gesture that happen quite often in Copenhagen due to its weather, the team came up with A Responsible Window: a trainable device that can be added to a common window and that uses different sensors to track various conditions of the environment inside your room, like temperature, humidity, quality of air and by using open weather APIs A Responsible Window knows weather conditions outside including temperature, humidity and whether it’s raining outside or not. Then the device combines the collected data from the sensors and APIs and connects them with your behaviour on how and when you were opening and closing the window and it acts according to the learnings.

When the person triggers the device by opening or closing the window, A Responsible Window learns from it and stores the values of the environment into different classes of a machine learning algorithm. After enough training, the device will be self-sufficient and will open or close the window depending on your preferences. The way to control the device is simple and elegant, instead of having buttons or any conventional controls a person just has to teach it by opening and closing the window.

A visual feedback is provided by a LEDs ring on top that informs the user when the device is learning or is changing the window position in order to balance the environmental comfort of the room. It is also possible to turn the device on or off any time if you want or not the device to behave on its own by simply turning the top part of the device.

The user can keep training the device every time he needs to improve a certain condition or to teach a new one.

Because all the learning happens inside the device A Responsible Window provides the opportunity to explore the space of local intelligence and smart devices which aren’t connected to the global network and creating challenges regarding privacy.