Research On Radio Tomography Imaging Model Based On RSS

Traditional localization needs target to carry auxiliary equipment.However,in some scenarios,such as the search and rescue of trapped people,the fall detection for the elderly,target will not carry equipment in advance or it is inconvenient to carry equipment.Device-free passive location(DFPL)is widely used in many scenarios because it does not need the target to carry auxiliary equipment.Radio Tomographic Imaging(RTI)is one of the common methods to achieve DFPL.The core idea of RTI is to establish a relationship model between the target and the Received Signal Strength(RSS)of the sensor in the monitoring area changing with the change of the target’s position.Then,the shadow fading distribution of each pixel in the monitoring area is obtained by regularization inversion,and the target is located.Therefore,how establishing a relationship model between target’s position and the RSS is one of the key factors to improve the quality of the RTI reconstructed image and the localization accuracy,which is also the focus of this paper.In order to improve the quality of the RTI reconstructed image and accuracy,an improved elliptic weight model based on the first Fresnel zone theory of antenna is proposed in this thesis.The elliptic weight model is more consistent with the theory,and the division of the weight of the pixel in the elliptic region is more precise.The main research contents of this thesis include:(1)The RTI-based location system used in this experiment is constructed from the two aspects of hardware and software.In the hardware aspect,the experiment environment is built with CC2530 nodes,and in the software aspect,the token protocol used in the node communication is analyzed.Through the experiment,the changes of the Received Signal Strength are compared when the monitoring environment is empty and when there is a target,which verifies that the location system based on RTI can collect data normally.(2)The relationship between the communication link with RSS fading and the target position is studied.Based on the first Fresnel zone theory of the antenna,an improved elliptic weight model is proposed,and the feasibility of the model is verified by the experiment.The data collected in the experiment are processed with missing period filling,converted from Hexadecimal to decimal and offset to get the true value of RSS,and the location performance of the model is evaluated by RTI imaging of the RSS.For the moving target,a method of corrected outliers and weighted filtering is used to fit its imaging trajectory.(3)The result of imaging by the improved elliptic weight model still has some noise.In this thesis,the RTI reconstruction image is denoised by the Genetic Algorithm based on Otsu segmentation algorithm with the RTI.After many experiments,the RSS of the target is collected in different positions.These data are used as the input for RTI imaging.The experimental results show that the localization accuracy of the improved elliptic weight model is higher than that of the adaptive elliptic weight model based on distance attenuation.Moreover,the method of corrected outliers and weighted filtering to fit the RTI imaging trajectory can improve the localization accuracy,and the use of the Genetic Algorithm based on Otsu segmentation algorithm to extract the RTI foreground can improve the image quality and localization accuracy.It has certain research significance for the development of Radio Tomographic Imaging.