Research On NLOS Localization And Coverage Hole Healing Methods For Wireless Sensor Network

With the rapid development of the sensor,wireless communication and network technologies,the wireless sensor network(WSN)with multiple functions of information collection and data processing has been widely used and attracted considerable research interest.Location information and network coverage are important topics for wireless sensor network.In this dissertation,we investigate the NLOS(Non-line of sight)localization and coverage holes repairing based on the analysis of domestic and overseas researches.The main contents are shown as follows:A NLOS localization algorithm based on strict residual selection is proposed since the obstacle and moving objects are easy to cause the NLOS propagation in monitoring region.The characters of the residual derived from the NLOS distance measurements are firstly analyzed in detail.The proposed algorithm produces the residuals of the NLOS distance measurements based on the Extend Kalman Filter(EKF)linear regression model.Then the strict residual selection mechanism is proposed to identify the state of the node.Finally,the parallel variable node EKF algorithm is used to estimate the location of the target node.The simulation results prove that the proposed algorithm could effectively inhibit the NLOS error and improve the localization accuracy.And a CSS(Chirp Spread Spectrum)indoor localization system is designed for practical experiments.The experimental results effectively verify the proposed algorithm could restrain the NLOS error.The validity of the algorithm is verified.A localization algorithm based on IMM-PDA(Interacting Multiple Model-Probabilistic Data Association)is proposed since the dynamic of NLOS error is strong in monitoring area.The transition between the LOS and NLOS is described by two state Markov model based on the characteristics of different measurements in LOS environment and NLOS environment.Then the improved probability data association method is proposed to combine and process the estimated location.Finally,the location of node is estimated by using the updated state probability.The simulation results show that the proposed algorithm could obtain the accurate localization in complex environment and improve the localization accuracy effectively.Considering the dynamic of the obstacles in the monitoring region,a localization algorithm based on IKF-GMD(Improved Kalman-Filter Gaussian Mixture Distribution)is proposed.Firstly,the predicted state and the covariance matrix are calculated based on the Kalman prediction.Secondly,a robust NLOS identification algorithm is proposed by analyzing the residual characteristics.For the NLOS measurements,the GMD based on EM(Expectation Maximization)method is proposed to estimate the mean and covariance of the measurements.Finally,multiple modes combination method is proposed to mitigate the NLOS error and achieve the accurate localization.The simulation results and experimental results show that the proposed algorithm does not assume any statistical knowledge of the NLOS error to obtain the accurate location of the node.Considering the NLOS error obeys the different distribution in different environments,a localization algorithm based on the vote selection is proposed.Firstly,a vote selection mechanism is proposed to filter the distance measurements and reserve the reliable measurements since the characteristics of NLOS error.Then the weights of distance measurements are obtained based on the Bayes theorem and a modified probabilistic data association algorithm is proposed to combine the estimated values.The position of the mobile node is finally determined by linear least squares algorithm.The simulation results and experimental results show that the proposed algorithm has better robustness and higher localization accuracy.It's easier to emerge the coverage holes when the network monitors long time.To solve the problem,a coverage holes detecting and healing algorithm based on C-V model is proposed.Firstly,the sensing model based on Neyman-Pearson criterion is employed to calculate the joint detection probability of each node in the monitoring area.Then a novel coverage holes detection algorithm based on the C-V model is proposed,which could obtain the number and size of the holes effectively.Finally,an improved particle swam algorithm is proposed to heal the coverage holes.The simulation results show that the proposed algorithm could effectively improve the coverage of the network and prolong network lifetime.