Multi-sensors data fusion system for fall detection

This paper presents a data fusion architecture whose goal consists in improving home support of the elderly mainly by detecting a potential fall. Among the different fusion modules composing this architecture, we focus on the human posture recognition system based on the fusion of visual features using a fuzzy logic algorithm. The posture, combined with other information (location, duration), enables to monitor the people state. Experimental results illustrate the robustness and the efficiency of the proposed approach.