Application of artificial neural networks in man’s gait recognition, 3rd Polish Congress of Mechanics & 21st International Conference on Computer Methods in Mechanics, Short Papers, Vol. 2, Gdańsk 2015
ABSTRACT: In this paper the method of human gait recognition based on artificial neural network is presented. Method is based on observation, that each individual has different, unique way of basic movement activity. Hence, parameters representing typical characteristics of gait of each person can be defined. In classical approach the video analysis is used to obtain the set of sequential pictures of the moving person. In this paper, all considered gait parameters are calculated from data which were obtained from dynamometric platforms used to measure the acting forces during the gait. Those gait characteristics are unique for each person, and some of geometrical qualities of curve which represent vertical and anterior-posterior components of force are taken into the account. To implement the recognition process the back-propagation neural network algorithm is used. In experiments, higher gait recognition performances have been achieved.