Sway analysis and fall prediction method based on spatio-temporal sliding window technique

As people age, they become more fragile and exhibit difficulties in maintaining their gait and balance. Their state of fragility increases their vulnerability to fall incidents. Various analysis methods were developed to detect the abnormality of human gait and balance, and estimate the risk of falling. In this paper, we present a method to estimate the falling risk and alert the patient when a fall is about to happen. The proposed method consists in monitoring and analyzing the amount of sway of the center of mass in the medial-lateral plane by computing the center of pressure displacement at the foot plantar surface. Our proposed method uses the spatio-temporal sliding window processing to generate fall alarms and estimate the falling risk. The method was validated via a two-phase experimental protocol with five young adults who performed a walk of 20 stances with simulated sways using an instrumented shoe with resistive pressure sensors. The threshold of the normal walk THN and the risk level RL of the altered walk are determined as well as the risk of falling. The method can be applied in real-life and clinical settings with real-time processing.