Research Of Indoor Human Perception Method Based On Channel State Information

The development of building intelligence not only increases the demand for indoor human perception applications,but also puts forward higher requirements for indoor human perception technology.Indoor human perception technology is divided into two types:contact type and non-contact type.Contact indoor human detection technology requires the active cooperation of users,and the application scenarios are limited and affect the user experience.Non-contact indoor human detection technology needs to rely on complex hardware equipment,which is more expensive.This paper proposes a non-contact,low-cost wireless perception-based indoor huamn perception technology.This technology uses the channel state information(CSI)of wireless signals to estimate and authenticate indoor people.The main work is as follows:In order to solve the problems existing in the traditional indoor human perception,this paper analyzes the feasibility of wireless perception technology,and then establishes a new wireless signal acquisition model to provide theoretical basis and data sources for indoor number estimation and identity authentication.In terms of indoor population estimation,in order to improve the utilization rate of CSI data signals and analyze the CSI amplitude and phase characteristics at the same time,an offline data model is established to achieve population classification.Firstly,the CSI data is preprocessed by phase correction and filtering,and the stable data features are extracted using Principal Component Analysis(PCA).Compare the performance of five algorithms:namely Support Vector Machine(SVM),K-Nearest Neighbor(KNN),Artificial Neural Networks(ANN)and AdaBoost based on single-layer decision tree,choose AdaBoost algorithm as the number of people estimated classification algorithm.The experimental results have good robustness and can effectively solve the problem of indoor number estimation.In terms of indoor personnel identity authentication,a new method is proposed to verify the indoor personnel identity according to each user’s unique breathing pattern.The phase change amplitude of the CSI signal can represent the periodic motion of the thoracic cavity caused by breathing.The breathing data can be optimized by using phase difference,sub-carrier selection and filter parameter adjustment techniques,and effective breathing features can be extracted by combining peak-to-peak value and matrix transformation.Dynamic Time Warping(DTW)is used to verify the feasibility of the identity authentication method,and an optimal identity authentication model is established by comparing five different classification algorithms.Eight users are invited to evaluate the method,and the proposed authentication algorithm can achieve an accuracy rate of more than 90%in different experimental scenarios.