Research On A Robust Radio Frequency Tomography Localization Method

Device-free localization(DFL)is a positioning technology that does not require any positioning device to be carried by the positioning target,and does not need to actively participate in the positioning process.Since it overcomes the disadvantages of most existing positioning technologies that require the target to be positioned must carry a positioning device and active cooperation with the positioning system,DFL technology can play a special role in positioning areas that cannot be achieved by traditional positioning methods,such as personnel search and rescue,illegal intrusion detection,animal habit monitoring,and elderly care in special circumstances.Compared with existing DFL methods based on cameras,ultra-wideband radar,infrared and ultrasonic technologies,DFL technology based on wireless sensor networks has advantages of low cost,good versatility,and can penetrate walls and smoke,so it has become one of the current research hotspots in the field of DFL.Among the many DFL methods,radio tomographic imaging(RTI)positioning technology has received more attention due to its intuitiveness and has become one of the current mainstream methods of DFL.In this method,several wireless transceiver nodes are arranged at the edge of the target area,and the influence of the target is reflected according to the changes in the received signal strength(RSS)of each node,and then use the inversion image method to show the positioning results for reference to the medical CT idea.The core principle of the traditional RTI method is to construct the relationship between the change of RSS value and the target position through a prior weight model,and then use the regularization method to reverse the performance of the position image.However,due to the characteristics of radio waves that are easily affected by uncertain factors such as environmental changes,non-line-of-sight,and multipath effects,the RSS value will fluctuate significantly over time in the actual measurement process,which makes priori weight model difficult to accurately describe the actual signal changes.Therefore,there are inevitably complex background noises in the RTI imaging results,and sometimes even false targets appear on the images,which brings great difficulties to the identification of real targets,and significantly reduces the RTI imaging effect and positioning accuracy.Therefore,how to effectively eliminate or reduce the influence of false targets and background noise in RTI imaging results has become a hot spot in RTI research in recent years.In response to this problem,based on the in-depth analysis of existing RTI methods,this article proposes several robust radio tomographic imaging methods for static targets and dynamic targets,which can effectively remove or reduce the influence of false targets and background noise.The main work and research results are as follows:(1)For static target positioning,this paper proposes a robust RTI location method based on background noise separation based on the characteristic that the position of a static target will not change within a period of time.This method uses multiple rounds of measurement data to perform imaging separately in the first place,and then uses robust principal component analysis(RPCA)technology to extract the common components affected by the target in the multiple rounds of imaging results,and at the same time,the RSS changes caused by non-target influencing factors are removed as background noise,which greatly improves the imaging quality and positioning accuracy.(2)Although the above method can effectively remove the influence of false targets and background noise,it requires multiple imaging,so it requires a large amount of calculation.In order to overcome this shortcoming,for static target positioning,this paper proposes an RTI positioning method based on wireless link stability analysis.This method selects a discriminant index that can represent the overall status of a set of data,and directly selects the most stable and the least noisy round data for RTI positioning,the imaging quality and positioning accuracy are greatly improved,and since only one imaging is required,the amount of calculation is also significantly reduced.(3)According to the characteristics of the dynamic target changing position instantaneously,this paper proposes an RTI positioning algorithm based on single-round data.Different from the existing RTI methods that only use the brightness information of the image to locate,this algorithm uses the brightness information and shape information of the target image together to detect and locate the target,which effectively solves the problem of pseudo-target interference that is difficult to overcome in the traditional RTI location.It not only improves the positioning accuracy in the single target location,but also has good performance in the multi-target location.Finally,the research work of this thesis is summarized,and the shortcomings of the work in the thesis are analyzed.At the same time,the improvement ideas are put forward for the shortcomings in order to improve in the future research.