Design of Multipurpose Sensing System for Human Gait Analysis

In this paper, we propose a multipurpose sensing system and fast phase analysis algorithms to predict and estimate behaviors of human being in advance. The sensing system is designed lightly and simply to minimize data acquisition errors and composed of six absolute encoders, eight FSR sensors, and FPGA board for data processing. The eight FSR sensors are attached at four points of sole of each foot where human weight is largely concentrated. The leg detaching structure that connects up with six light steel links is tuned according to the fittest condition of a subject. Thus, each link can be adjusted according to subject's body length. The phase analysis algorithms have two major categories for analyzing walk behaviors and stand behaviors, and four subordinate phases for standing, lift, swing, and landing. This algorithm considers standing motion and two leg behaviors. Experimental implementations are performed successfully to estimate various human gaits. This acquisition data are efficient for locomotion trajectory generation of biped robots and exoskeleton systems