A Robust Localization Algorithm Based on NLOS Identification and Classification Filtering for Wireless Sensor Network

Abstract

With the rapid development of information and communication technology, the wireless sensor network (WSN) has shown broad application prospects in a growing number of fields. The non-line-of-sight (NLOS) problem is the main challenge to WSN localization, which seriously reduces the positioning accuracy. In this paper, a robust localization algorithm based on NLOS identification and classification filtering for WSN is proposed to solve this problem. It is difficult to use a single filter to filter out NLOS noise in all cases since NLOS cases are extremely complicated in real scenarios. Therefore, in order to improve the robustness, we first propose a NLOS identification strategy to detect the severity of NLOS, and then NLOS situations are divided into two categories according to the severity: mild NLOS and severe NLOS. Secondly, classification filtering is performed to obtain respective position estimates. An extended Kalman filter is applied to filter line-of-sight (LOS) noise. For mild NLOS, the large outliers are clipped by the redescending score function in the robust extended Kalman filter, yielding superior performance. For severe NLOS, a severe NLOS mitigation algorithm based on LOS reconstruction is proposed, in which the average value of NLOS error is estimated and the measurements are reconstructed and corrected for subsequent positioning. Finally, an interactive multiple model algorithm is employed to obtain the final positioning result by weighting the position estimation of LOS and NLOS. Simulation and experimental results show that the proposed algorithm can effectively suppress NLOS error and obtain higher positioning accuracy when compared with existing algorithms.

A Novel NLOS Suppression Algorithm for Indoor Location based on FCM and REKF

“…Paper [4] uses the association gate to identify whether a deviation is acceptable, by discarding the NLOS measurements the localization can attach high accuracy. Paper [5] classifies the environment into LOS, mixture of LOS and NLOS, NLOS three conditions while paper [6] classifies the NLOS environment into soft and hard two parts. These more detailed classification leads to the more precise processing which shows significant improvements in the simulation results.…”

A Distance Increment Smoothing Method and Its Application on the Detection of NLOS in the Cooperative Positioning

“…However, the algorithm still needs noise statistics, which are unknown. Chen et al [ 34 ] proposed a robust algorithm using NLOS recognition and classification, dividing NLOS into light and heavy NLOS. The light NLOS was truncated by robust filtering, while the line–of–sight reconstruction estimated the heavy NLOS, but this method relied on an established known platform.…”

Constrained L1-Norm Minimization Method for Range-Based Source Localization under Mixed Sparse LOS/NLOS Environments

“…If the intersection area becomes larger, the coordinates of MS will be difficult to be estimated and the corresponding positioning performance will be degraded. Additionally, note that Gaussian noise, NLOS error and layout of BSs are the most important factors that affect the size of the intersection area [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]. Among them, NLOS error may be the most influential factor because it can directly lead to inaccurate measurements, resulting in inaccurate localization results or even divergence .…”

“…Additionally, note that Gaussian noise, NLOS error and layout of BSs are the most important factors that affect the size of the intersection area [ 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 ]. Among them, NLOS error may be the most influential factor because it can directly lead to inaccurate measurements, resulting in inaccurate localization results or even divergence [ 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. Therefore, it is of great practical significance to study how to improve the positioning performance and degrade the effects of NLOS errors without requiring the priori information of NLOS status and NLOS errors.…”

“…To reduce the effects of NLOS errors and enhance the positioning performance, lots of methods have been widely studied in [ 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. In [ 10 ], a fast identification of NLOS error using eigenvector (FINE) algorithm is proposed by using autocorrelation matrix eigenvalues of residuals to identify the NLOS paths.…”