Indoor Person Device-Free Localization Technology Research For Ultra-Wideband Network

With the sustained development of information technology,location-based services have penetrated into various industries.Indoor localization is effective guarantee to achieve whole field of location services.Relative to the outdoor channel,indoor objects and building itself can lead to multipath transmission of signal.Therefore,indoor propagation channel is more complex.Multipath interference is an important factor leading to inaccurate indoor localization.Ultra-wideband technology with ultrahigh temporal resolution,as well as penetrate or bypass obstacles,etc.,is more suitable for high-precision localization of indoor objects and person.Nowadays,most of the researches focus on ultra-wideband device-based localization.In this dissertation,we mainly study the device-free localization problem of UWB.For the problem of indoor device-free person localization in the ultra-wideband network,based on the study of the current UWB multipath channel model,this dissertation considered the special properties of people such as periodic breathing and heartbeat.This low frequency motion may cause signal transmission path where the target at has a time-varying characteristic.Considering the signal multipath propagation in the environment and time-varying characteristic of channel above-mentioned,a multipath impulse response channel model was proposed.Based on the time-domain uniform diffraction theory and ray-tracing,a deterministic model was constructed,in which the human object was modeled as a rectangular blade,and signal transmission where the target being was modeled as diffraction.According to the proposed UWB channel model,a novel localization algorithm based on time of arrival of UWB signal was proposed,achieved location of single target and double targets.First,by using a priori knowledge of the received signal when there were no targets in the environment,we extracted the received signal related to people;Second,the time of arrival for the signal transmission path where the people at was estimated.Finally,Ranging from various paths of the received signal were calculated,weighted least squares algorithm was used for target position estimation.Based on proposed deterministic model,we established a simulation scenario for indoor devicefree person localization.Simulation results show that 90% location result of single target is lower than 0.3m,root mean square error is 0.7m;for double targets,the 78% location result is lower than 2m,root mean square error 2.1m.The algorithm achieves the single target and the double targets localization well.