Human Behavior Recognition Based On Gait Pressure Characteristics

Nowadays,people's lives have been deeply influenced by the rapid development of Internet technology and electronic technology.In some cases,the accuracy and safety of people's identity authentication have been put forward more stringent requirements,which is why the biometrics at this stage biometric technology has been so widespread concern.Using biometrics to identify human behavior and distinguish human identity is a new trend,gait recognition technology as a new generation of identity technology has an unique advantage,mainly due to the gait by the innate factors and acquired habits of the impact,with relatively easy to imitate,camouflage,stability and specificity advantages.The process of human behavior identification includes three major aspects: human detection,feature extraction and behavior recognition.The method adopted in this paper focuses on the two aspects of feature extraction and behavior recognition,and proposes the corresponding improved algorithm.Through the wearable gauges,the gait information is collected,and a method of adding frequency information to other features is proposed,which can accurately cluster the motion state.Human behavior is not only a single behavior,human behavior is a period of time the movement process,through the analysis of the process of movement and decision-making,gait recognition can be well applied to the field of biometric identification.This paper is based on the wearable insole of single chip and FSR402 sensors,clustered by improving the kernel function parameters of the clustering algorithm,which can be in the unknown state of motion has a good clustering effect,and as a basis for the realization of behavior recognition feature extraction method.Since each individual is different,the threshold of the state of motion in the behavior of each individual varies from person to person.In the process of behavior decision-making,combined with the machine learning algorithm,and match the characteristics and training samples to analyze and make the process of human behavior.