Research On Indoor NLOS Multi-station Collaborative Location Technology

With the development and popularization of information technology.The amount of people that demand for location services is increasing.Nowadays,indoor activites are the trend that adopted by most of people.The importance of indoor location information is self-evident which promoted indoor localizaiton technology.However,most of them with high precision rely on the Line-of-Sight(LOS)path,in the complex indoor environments,the LOS path may be blocked,and thus those localizaiton technologies based on the LOS environment will become unavailable.Therefore,it is very important to investigate indoor localization based on Non-Line-of-Sight(NLOS)propgatagiton.However,it is hard to realize it in the existing networks.The main reason is that the phase error makes the absolute Time of Flight(TOF)unavailable in the wireless networks.To this end,this thesis proposes an indoor NLOS multi-station cooperative positioning algorithm based on the messages the scatterers imply,which fully takes advantage of the deployment information of the indoor scene as well as the scatterers distribution feature to estimate the target location.The main research contents are as follows:Firstly,based on the propagation characteristics of NLOS path,the influence of indoor structure information on the scatterer distribution is analyzed,and then the location information of the scatterer is determined to construct the target location equation.Specifically,the layout information of multiple Access Points(AP)and the priori information of the house structure are used to determine the fuzzy area of the scatterer location and the feasible region of the target location.Then the ambigutiy area of the scatterer is further reduced based on the Angle of Arrival(AOA)of each reflection path,and then the scatterer is selected in the area.Based on this,the target localizaiton objective equations are constructed by utilizing the selected scatterer location,AP location and differential TOF observations.Then,the problem of target localization is converted into an optimization problem,and the Genetic Algorithm(GA)is employed to roughly estimate the location of the target.Then,the Leverberg Maquardt(LM)algorithm is used to optimize the target location.Based on this,the clustering algorithm is further used to optimize the localization results,and the factors that affect the localization accuracy contained in the model are analyzed.Finally,the simulation results show that the localization method designed in this thesis has a localization error of 1.66 meters with a confidence of 67%.In order to further verify the effectiveness of the algorithm in actual scenarios,this thesis uses Wireless Insite to carry out a real environment simulation test.The experimental results show that the localization error with a confidence of 67% is 1.72 meters.Therefore,the localizaiton method proposed in this thesis using multiple APs can achieve the target meter-level localizaiton in indoor NLOS environments.