Towards situated public displays as multicast systems

Within organizations diverse asynchronous, unreliable multicast channels to push information chunks already exist. These usually include posters, newsletters, intranet pages, print media and talks. Situated public displays have the potential to complement these channels, because their properties of being situated and public make them superior to conventional information channels in certain situations. In order to present the information that is most interesting to users and thus improve the efficiency of information distribution, we propose to learn a model of the spatio-temporal behaviour of users, find clusters belonging to certain organizational subgroups within this model, gather profiles for some of the users and use the estimated profiles of clusters to optimize the information presentation on the displays.