Research On High Resolution WiFi Imaging Technology

Imaging sensing is widely used in daily life,including indoor imaging,security check,environmental detection,etc.WiFi-based Sensing technology has the advantages of non-invasive,wide coverage and low cost,which has become a hot topic in recent years.Due to factors such as WiFi’s low frequency band,narrow bandwidth,and non-adjustable signal,the accuracy of imaging results is low and cannot meet daily needs.Focusing on the limitations of commercial WiFi devices in daily imaging,we conducts super-resolution imaging research by using WiFi Channel State Information(CSI)to demonstrate indoor static target imaging and motion detection.To solve the problem of low imaging resolution accuracy due.to the limited bandwidth of WiFi,we propose a WiFi high-resolution indoor static target imaging algorithm by the angular spectrum diffraction theory of signals.First,the CSI data is collected and preprocessed by a simulated antenna array on the receiving surface using the time-space accumulation of the scanning antenna.Then,the frequency-domain signal of the receiving surface is extracted from the CSI ratio of dynamic scanning antenna and static reference antenna,and the noise of the original CSI data is eliminated.Finally,we reconstruct the contour image of the target object and-perform target localization,and the frequency domain signal of the receiving surface mapped by CSI is transmitted through the diffraction propagation model.Our imaging algorithm is not limited by bandwidth.And its accuracy depends on the wavelength and the sampling aperture size of the receiving surface,achieving super-resolution imaging.Experimental results show that the indoor static target imaging algorithm achieves clear imaging results of a metal cross at a distance of 1.5 meters between the transmitter and receiver,with a target positioning accuracy error of 9 cm.Aiming at the problem of the issue of moving objects affecting imaging sensing in the environment,we proposes a WiFi motion detection algorithm based on multi-dimensional features of CSI.First,multiple sets of CSI data are collected under static and dynamic environments and preprocessed.Then,a WiFi signal propagation model is constructed in the moving environment,and multi-dimensional feature indicators on the amplitude and phase of CSI in the frequency and time domains are established to correlate with moving objects.The algorithm is able to quickly detect moving objects in the environment without the need for a wearable device.Finally,we a support vector machine to threshold classify the above features for a real-time motion detection classifier.This motion detection can quickly detect moving objects in space without the need for wearable devices.Experiments show that with the inputs of multidimensional feature,the overall detection accuracy of both motion and static reached 97%.