Gait Analysis Algorithms Based On The Body Sensor Networks

Following social progress and development, people are more concerned about their own health. Especially, aging problem becomes more serious, and current pension services and nursing staff can’t meet people’s needs. China’s major medical monitoring mode is mainly based on large-scale hospital equipment, which increases the burden of hospitals, and is very inconvenient. The introduction of new medical monitoring mode to improve the quality of medical care becomes an urgent need. The rise of Body Sensor Network (BSN) technology in the world can effectively assist hospitals. And the acquisition of physiological signals collected from patients at home helps to achieve the purpose of remote medical monitoring.A gait is a walking posture of human body. And gaits of human contain a lot of information which can reflect the status of human health. So the analysis of spatial and temporal parameters of gaits is very important in medicine. The traditional gait analysis methods are mainly based on the video, pressure sensors or large gait simulation platform. But these methods are not conducive to the promotion of telemedicine. With the rapid development of micro-electro-mechanical systems (MEMS) technology, acceleration sensors and gyroscopes have attracted a lot of attentions due to the low-cost, lightweight, low power consumption and small size, and they have been applied to the motion analysis.In this paper, a set of quantitative gait analysis algorithms based on inertial measurement module (IMU) which contains a three-axis acceleration sensor and a three-axis gyro are introduced. The inertial measurement module fixed to human’s ankle collects the inertial signals from human motion. And the signals will be analyzed through several certain algorithms to get the desired gait parameters. The signal analysis process mainly includes signal preprocessing, gait cycle segment, gait parameter calculation and error compensation. The wavelet decomposition technique is used in the signal preprocessing part, and a knowledge-based algorithm was used in the gait cycle segment part. To calculate the posture of a foot, the quaternion-base Strap-down Inertial Navigation technology is used. And the spherical interpolation technique and linear interpolation techniques are used to achieve foot posture and foot speed error compensation.