Application Development Framework for Smart Rooms

 

A smart-room is an example of context awareness and ubiquitous computing scenario. The main idea is to use the
room’s sensors (cameras and microphones) and some computers to recognize the activity in the room and react
based on it, i.e., the computers should fit the human environment instead of forcing humans to enter theirs.
By nature a smart-room application has several important differences from a desktop one. From the point of view of
interaction, desktop applications mainly use GUIs with mouses and keyboards, but the inputs of the smart-room
applications are recognition algorithms, e.g. generic image and audio recognition or specific face, voice, speech,
gesture, activity recognition, etc... Also the outputs are not just displays but more “actuators” like talking heads,
different types of alarms or notifications, controllers of the room devices, personalized launchers of external desktop
applications, etc.. From the point of view of architecture there are also important differences. The sensors,
processing computers, actuators and saved data are distributed systems and need high transfer bandwidth, and
sessions (i.e. users entering in the room) are single user, multi-user or both at same time.
The main objective of this project is to design and implement a generic framework to develop smart-room
applications in a similar way than VR-Juggler6 or Ogre7 do it for virtual reality and 3D applications respectively. We
aim at developing a generic and scalable framework supporting the definition of components or abstract layers
(sensors acquisition, recognition algorithms, actuators, stored data, controllers...), implicit communication
(signalization) system between components, centralized control and visualization, support for online executions and
offline simulations for complete applications and for individual components, shared database for all the components
and all users. In order to test the framework one or two demo applications will be also developed.