Discussion

Smart home research requires an integrated, multidisciplinary approach to the many and varied questions it raises.

The first is: which sensors for which information? Obviously, the greater the number of sensors, the more detailed the information we'll be able to obtain, as in the case of video cameras and scene interpretation algorithms. But it will also be more difficult to install and maintain these technologies in the home, and to get them accepted in the privacy of the home. On the other hand, using technologies that are already accessible is less tricky. This leads us to work on traces (indices) of human activity, at the cost of a major investment in more subtle mathematical tools to identify and classify common scenarios (e.g. hidden Markov models