Research On Indoor Node Localization For Wireless Sensor Network In NLOS Environment

With the promotion of digitalized China,it has gone all the way to the information age.As the main technical support of the Internet of things,wireless sensor networks have been widely used in the fields of military operations,environmental monitoring,medical and health,industrial production,intelligent home and so on.Localization technology,as one of the most critical technologies in wireless sensor networks,has attracted much attention from researchers.In this dissertation,the indoor environment is the research object.On the basis of understanding the research status at home and abroad,the localization problem of static nodes and dynamic nodes are investigated in depth for wireless sensor networks under indoor NLOS environment.The aim is to realize the localization problem with high accuracy and adaptability in complex environment.The main research contents and achievements are reflected in the following aspects:The localization algorithm of adaptive weighted least squares estimation based on RSS(Received Signal Strength)is proposed since many barriers can cause signal NLOS(Non-line of sight)spread phenomenon in indoor small complex environment.Under the condition of known distribution of error probability density function,the generalized likelihood ratio hypothesis test is used to determine whether the measurement channel is polluted by NLOS.Under the condition of determining the state of signal propagation,the adaptive weighted least squares estimation based on median is used to locate the localization to weaken the effect of NLOS error.Considering the simple and strong global search ability of intelligent optimization algorithm,an NLOS node localization method based on intelligent optimization is proposed.The method of hypothesis testing is used to determine the state of the NLOS and the particle swarm optimization algorithm with a shrinkage factor is used to localization.Since abnormal sample value,the performance of the sample median is better than the sample mean.Therefore,under the condition of LOS and NLOS,the minimum residual cost functions are established by using sample mean and sample median.In order to enhance the global and local search ability of the algorithm,the shrinkage factor is introduced on the basis of particle swarm optimization algorithm,so that the algorithm converges faster,and the location accuracy of the algorithm is obviously improved.Considering that the sensor nodes often present the mobile characteristics,the non-line of sight location algorithm based on the feasible region particle filter of TODA and RSS is proposed for the phenomenon of signal switches between LOS/NLOS.First,a hypothesis testing method based on two ranging models based on TDOA and RSS is used to identify whether there is NLOS phenomenon in measurement signals.Then the location of unknown mobile nodes is estimated by using the feasible region particle filter method considering NLOS measurement information.Due to the reduction of the scope of the feasible region of the particle,the calculation of the algorithm is reduced and the positioning accuracy is improved.Considering that the indoor environment is influenced by external factors,the dynamic changes of the NLOS noise parameters are changed,and an improved Kalman filtering strategy is proposed for the unknown NLOS noise parameters.In the case of LOS,the measured distance is filtered by Kalman,and the modified Bayesian approximation adaptive Kalman filter is designed in the case of NLOS to estimate the mean and covariance of the measured noise to correct the measured values.Finally,residual weighting method is used to locate nodes,which weakens the effect of NLOS error.Considering that the NLOS identification in the positioning algorithm may bring some errors,the interactive multi-model mobile algorithm based on UKF is proposed to solve the problem that the positioning system model is nonlinear and the NLOS measurement noise parameters are unknown.The algorithm uses a parallel structure to input the estimated state of the unknown node to the UKF filter under the LOS channel and the modified UKF filter under the NLOS channel to obtain the state estimation respectively.Finally,according to the weighting factor,the estimated values are fused to get the estimated position.The algorithm can improve the positioning accuracy greatly.This dissertation systematically studies the theory and method of localization in wireless sensor networks under indoor NLOS environment.Based on that,simulation experiments and results analysis of the proposed algorithm are carried out.By comparing and analyzing with other methods,the effectiveness,feasibility and progressiveness of the proposed method are verified.