Research On Human Behavior Recognition Technology Based On CSI

Behavior recognition has great potential in many fields,such as human-computer interaction,security and so on,which has always been a hot topic in academic and scientific circles.WiFi Based behavior recognition technology overcomes the limitations of traditional behavior recognition technology,and has become the most suitable research scheme in this field.Channel state information(CSI),as a channel attribute of communication transmission path,can reflect the fine-grained characteristics of WiFi signal,and the sensing accuracy is very considerable.Therefore,it is of great significance to design and implement CSI based behavior recognition technology,which is also the main research direction of relevant research teams around the world.This paper focuses on the research of behavior recognition.Aiming at the problems of low recognition accuracy and no wearability in this field,a behavior recognition technology based on WiFi signal and multi feature combination is proposed.The main work and innovation of this paper are as follows:(1)This paper analyzes the principle and definition of behavior recognition;summarizes the current research status and development process of behavior recognition technology;analyzes the shortcomings and limitations of previous technologies;and understands the broad application prospects in this field.The advantages and disadvantages of RSS and CSI are analyzed.Finally,CSI is selected as the base signal.(2)After learning and summarizing the principles and methods of previous behavior recognition technology based on WiFi signal CSI amplitude,this paper proposes the multi feature behavior recognition technology.It includes two feature modules.The third chapter of this paper corresponds to the first module,that is,the age segment recognition based on the Fresnel zone model gait features.Based on the previous gait recognition,we can further identify the user’s age segment,and then push different service content according to the user’s different age segment.(3)The second feature module is a complex gesture recognition module based on Hidden Markov model.This module applies hidden Markov model to gesture recognition,and on the basis of simple gesture recognition,it can recognize complex gestures composed of simple gestures.The combination of the two modules forms the overall technology of this paper,which can be applied to many scenes,including some special scenes with high security level that need real-time feedback and some public indoor scenes,and can provide different prevention and services for different age groups.(4)In this paper,experiments are carried out in three different interference environments,and the overall error rate is less than 15%.The experimental evaluation results show that the proposed technology has considerable effectiveness and robustness.The proposed technology has practical significance in various indoor scenes,and has a certain role in promoting the development of this field.