Research On Highly Accurate Localization Algorithm Of Wireless Sensor Networks Under Mixed LOS/NLOS Scenarios

The prosperity of the country,the rapid development of society and the remarkable improvement of people’s living standards are inseparable from the development and progression of science and technology.Among many scientific technologies,wireless sensor positioning technology which can provide location information service has made important contributions in the fields of industrial safety production and order of social life.There are four main localization modes in traditional wireless sensor networks:AOA,TOA,TDOA and RSS.Among them,TO A and TDOA are widely used.Regardless of which positioning mode is used,the environment can be divided into LOS scenario and NLOS scenario according to whether there is occlusion in the transmission process of the signal.The traditional localization algorithms in LOS scenario are mainly LS algorithm,CWLS algorithm,Taylor algorithm,convex optimization method and Kalman filter algorithm.These algorithms have achieved good positioning performance in TOA positioning mode,but have the disadvantage of unstable performance for TDOA positioning mode.For NLOS environment,the main localization algorithm is convex optimization method.However,under mixed LOS/NLOS or NLOS scenarios,the convex optimization method without any prior information has the disadvantage of low positioning accuracy,while the convex optimization method with certain prior information has the advantage of high accuracy,but it has the disadvantage of unrealistic assumptions and cannot be applied to the industrial field.As is known to all,NLOS error has time-varying characteristic,which makes it difficult to model accurately.Therefore,it has a great impact on the positioning performance of the system,and the severity will lead to divergence.Therefore,NLOS error has been one of the most important error sources affecting the positioning performance of WSN for decades,and the research of high-precision positioning algorithm in NLOS environment has also been one of the most important core technologies of "stuck neck" in the industry.Aiming at the problems of unstable positioning algorithm,low positioning accuracy and inability to be applied in the industrial field,this paper applies the iterative least squares technique and two-step detection technique to improve the performance of TDOA positioning algorithm,and uses the sparse technique and exhaustive method to solve the high-precision positioning problem under the mixed LOS/NLOS environments.Specifically,the main innovation works of this paper are summarized as follows:1.Aiming at the problem of low positioning accuracy of TDOA algorithm in LOS scenario,an iterative least squares algorithm is proposed.This method has very high positioning accuracy and fast computational speed in four base station environment.2.Aiming at the problem of location divergence of ICWLS algorithm in TDOA mode,a two-step detection method is proposed to identify and deal with location divergence.The performance of this detection method combined with ICWLS algorithm is significantly better than that of ICWLS algorithm.3.In view of the disadvantages of large computation and limited positioning accuracy of positioning algorithm under hybrid sparse LOS/NLOS environments,it is found that the positioning residual vector has sparse characteristic,so a series of sparse positioning algorithms are proposed.The sparse location algorithms achieve a very good location effect in TOA location mode.4.Aiming at the disadvantages of low accuracy and divergent location of the positioning algorithm under mixed non-sparse LOS/NLOS environments,a criterion for selecting LOS base stations is found,and an exhaustive method is proposed.This method has achieved a very good expected effect in both TOA and TDOA modes,and has a very important reference value and significance for industrial application and promotion.