Résumé
A capacitive proximity capture device SensFloor(1) was installed in the HUman at homeprojecT(2) apartment in Montpellier, South of France. Activations related to participants’movements were continuously captured and recorded by this smart floor. This low-cost Floorpresent the issue of spatial precision of collected signals. Actually, the least spatial elementis a triangle of 25x50 cm. Moreover, the activation signal is a capacitive one that is notproportional to the weight of the object. As a consequence, it is challenging to organize thisspace-temporal signals into human behavioral events such as static position, trample, walk,alone or more persons into the apartment and so on. These events will be in the foundationof defining Human@Home metrics. Trajectories were detected, identified and reconstructedby the Walk@Home algorithm(3). In the core of this algorithm is a space-temporal windowthat scans the raw signals and organize them into a dynamic graph containing the eventualtrajectories. Even then, the result is an approximation of the movement of the center of thegravity of a human.