Machine learning gait analysis algorithm for ontogenetic features compensation

From the very first time when knowledge of exact human gait cycle phase was needed, numerous gait identification algorithms were proposed. Because of relatively simple nature in comparison to other phenomenons, some of those methods were based on foot-ground contact zone reactions. Regrettably, human gait is drastically complex process that differs from one person to another, not infrequently significantly, which causes numerous problems in creating its mathematical model. Due to this fact it is hard to create rehabilitation device control system based on traditional algorithms suitable for each person. In this paper we present comparison of signal similarities algorithms and adaptive neuro fuzzy inference system based gait phase classifier designed to counteract ontogenetic characteristics. Data sets composed of foot pressure sensors readings are provided to both algorithms to determine differences in each approach.