Research On Easy Falling Gait Feature And Early Warning

Walking is a kind of limb movement which is coordinated with muscle,skeleton and nervous system.The body's posture and movement pattern are defined to gait when walking.With the increasing of the population,the body falling has become a major threat to people's healthy lives,and the abnormal gait has received more and more attention.The study of gait feature can be used for medical health,sports competition,human machine interface,identification and so on.With the development of microelectronics technology,it is possible to create a portable gait analysis equipment with small size,low power consumption and multifunction.In this paper,a gait data acquisition system which can gather the pressure,acceleration and angular velocity of foot based on wireless microcontroller is designed and implemented.The data acquisition system is embedded in a sole,which is small and easy to be used.It has high using comfort with no burden on daily wearing.In order to observe and store the gait data,an application is developed so that real-time gait data can be displayed on the cellphone.An experiment of gait data acquisition was performed by using the gait data acquisition system.Data of various types of gait including standing,normal walking,unstable walking and falled were collected.After the feature analysing,the easy falling gait feature are confirmed.The recognition of easy-to-fall gait is realized by constructing a LSTM based gait recognition model.This model is integrated into application to realize the real-time detection of easy-to-fall gait,achieving the role of the fall warning.Finally,a test was carried out in the actual scene,which verified that the gait acquisition system could collect gait data in real-time,and the gait recognition model could successfully recognized the easy-to-fall gait.The result shows that the system can perform warning before falling.