Research On Human Detection Method Of Speed Adaptive Indoor Based On Channel Status Information

Mobile human detection plays an important role in building monitoring,cultural relic protection,boundary detection and detection of survivors in fire or earthquake.Traditional mobile human detection is based on special physical hardware and has great limitations in deployment and application.In today's era,with the rapid development of WLAN technology,there is a new research direction in the field of passive mobile human detection in the indoor environment.Due to the coarse granularity of wireless signal strength,the detection method based on wireless signal strength will affect the collected information under the influence of indoor multipath effect,resulting in the inability to accurately detect the presence of a human body and make the indoor equipment irrelevant Passive body detection accuracy can not be achieved to a satisfactory level.In order to improve the detection accuracy of moving human indoor indoor,this paper proposes the use of fine-grained channel state information as detection information,and further studies from two aspects of channel state information denoising processing and speed adaptive human detection.Noise has an unavoidable influence on the collected channel state information.Generally,the noise processing of the channel state information is based on the Fourier transform filter for denoising.However,due to the flaw of the time-frequency transform characteristic of the Fourier transform itself is not suitable for the unstable signal,but for the human body to detect the channel state information Is inevitably composed of unstable signals,which makes the human detection based on Fourier transform denoising effect is not ideal.In a few studies,the noise of channel state information is processed by means of wavelet denoising,but the denoising effect of wavelet denoising can be further improved.In order to improve the effect of wavelet denoising,this paper aims to optimize the existing wavelet threshold processing methods,and combines the default processing methods of the wavelet denoising,the hard threshold processing methods and the soft threshold processing methods respectively,and proposes a gradual threshold processing Method.Compared with the default soft thresholding method and the hard thresholding method,it is proved that the gradual thresholding method has a better denoising effect on the channel state information detected by the human body,which is very suitable for the mobile human body detection based on the channel state information.In traditional research methods of channel state information,the average value of a certain subcarrier or all subcarriers is usually used as the detection data.In speed-adaptive human detection,if the human body slowly moves in the detection environment,the impact on the channel state information may be small,resulting in a lower detection rate.In this paper,in-depth study on the above issues,proposed based on the dynamic selection of sub-carrier speed adaptive indoor human detection method(SAHD).In this method,the subcarriers are selected dynamically by detecting the fluctuation of subcarriers at the same time as the detection data,and the eigenvalues of the channel state information are selected for analysis.The appropriate criteria is selected Detection of the human body movement.Through contrast experiments,it is verified that SAHD is more accurate than traditional detection methods in human detection at different moving speeds.