LEVEL, DOWNHILL AND UPHILL WALKING IDENTIFICATION USING NEURAL NETWORKS

Body accelerations during human walking are recorded by a portable mea suring device. A new method for parameterising body accelerations is i ntroduced. The parameters are presented to a Kohonen neural network cl assifier and the feasibility of identification and dissociation of lev el and walking on a gradient is demonstrated. The most important and o riginal aspect of this classification is its ability to identify the g radient of walking performed in free-living conditions from walking tr ained on a treadmill.