Design of a smart shoe for reliable gait analysis using state transition theory

Abstract

This paper proposes a novel decision system that can distinguish the following five phases of a normal gait cycle in real-time - stance, heel-off, swing 1, swing 2, and heel-strike - by using a smart shoe. This decision system employs four sensors to evaluate forces exerted by a foot on the shoe insole during ambulation and a gyroscope, attached at the back of the shoe, to measure angular velocity along the sagittal plane. A threshold-based detection algorithm employing state transition theory is proposed to determine the appropriate gait phase at each gait moment using these sensor outputs. The proposed decision system is tested and verified through experimentation by a normal person who is assigned to walk with the smart shoe on an indoor field with smooth floor. The sensor outputs, with gait analysis acquired from the experiment, are shown in this paper.