Indoor Body Motion Detection Based On WIFI Signal Research And Application

Wi Fi technology based on wireless LAN can not only be used to transmit data,but also can be used to perceive human body movements in the environment.For the WiFi signal,the received signal strength RSS is easy to obtain,but RSS as a coarse-grained information mainly comes from the superposition result at the receiving end during the signal transmission process,and is susceptible to multipath effect and environmental noise,and there is a large fluctuation,poor stability.Currently,a more granular physical channel state information CSI can be obtained on an ordinary commercial WiFi device,which can effectively avoid the effects of multi-path effects and noise.At the same time,the CSI can carry characteristic information reflecting the surrounding environment during transmission,and can be used to detect changes in human motion.The main work of this article includes the following sections:(1)A brief overview of the technology and application of human detection at home and abroad,comparing the characteristics of traditional human detection technology,and pointing out the deficiencies,summarizes the current development of human detection technology based on radio frequency.(2)Analyze the wireless signal propagation model and elaborate the principle of wireless channel response.Based on the comparison of the key indicators RSS and CSI of the existing context-aware WiFi-based technologies,the flaws of RSS and the advantages of CSI are analyzed,and the feasibility of indoor human motion detection based on CSI is proposed.(3)Study the characteristic information of WiFi signal to provide basis for realizing fine-grained human detection.The CSI-Tool acquisition tool is used to extract the physical layer CSI data,and the original CSI data is subjected to local anomaly detection LOF algorithm and Hampel anomaly discrimination method to perform anomaly detection and rejection.At the same time,Butterworth filtering and principal component analysis(PCA)are used for filtering processing.Then,the linear phase transformation technology is used to process the frequency offset from the wireless network card itself,and the phase deviation is corrected.The feature information is extracted from the preprocessed CSI data by using the covariance matrix of amplitude and phase,and the SVM algorithm is used to classify and detect human motion.(4)Set up a CSI-Tool experimental platform,test the acquisition tools,analyze the SNR in different environments,compare the stability of the RSS and CSI,and analyze the values of different parameters in the Hampel anomaly discrimination.The influence of the number of different subcarriers and the number of different data streams on the detection of human motion,and the feature-based information is extracted based on the amplitude,and is used as a classification detection input basis.The three actions of standing up,walking,and sitting down are detected.