A TDOA Localization Method for Nonline-of-Sight Scenarios

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.

Section: Related Workmentioning

confidence: 99%

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

Cheng

Huang

Xue

et al. 2020

“…In [ 17 ], the authors transform the new robust weighted least squares problem into a non-convex optimization problem through the S-Lemma, so that the optimal target position is obtained. Because the frequency of signal is hard to affect during propagation, the authors in [ 18 ] establish a frequency and position transfer function by linking the field where a receiver is given to the source, which not only mitigates the NLOS effect but also calibrates the propagation channel back to free space, thus the performance is better than the usual TDOA positioning algorithm. For the purpose of avoiding the shortcomings of the robust least square (RLS) methods and reducing the upper limit of the NLOS error, the source location and the NLOS error in the predicted path are jointly estimated in [ 19 ].…”

Section: Related Workmentioning

confidence: 99%

An Indoor Robust Localization Algorithm Based on Data Association Technique

Cheng

Wang

Xue

et al. 2020

As a key technology of the Internet of Things, wireless sensor network (WSN) has been used widely in indoor localization systems. However, when the sensor is transmitting signals, it is affected by the non-line-of-sight (NLOS) transmission, and the accuracy of the positioning result is decreased. Therefore, solving the problem of NLOS positioning has become a major focus for indoor positioning. This paper focuses on solving the problem of NLOS transmission that reduces positioning accuracy in indoor positioning. We divided the anchor nodes into several groups and obtained the position information of the target node for each group through the maximum likelihood estimation (MLE). By identifying the NLOS method, a part of the position estimates polluted by NLOS transmission was discarded. For the position estimates that passed the hypothesis testing, a corresponding poly-probability matrix was established, and the probability of each position estimate from line-of-sight (LOS) and NLOS was calculated. The position of the target was obtained by combining the probability with the position estimate. In addition, we also considered the case where there was no continuous position estimation through hypothesis testing and through the NLOS tracking method to avoid positioning errors. Simulation and experimental results show that the algorithm proposed has higher positioning accuracy and higher robustness than other algorithms.

“…In contrast, range-based approaches determine the location of sensor nodes using various ranging methods and provide an accurate location [19]. Ranging methods can be classified into four categories [20]: Time of Arrival (TOA) [21,22,23,24], Time Difference of Arrival (TDOA) [25,26,27], Angle of Arrival (AOA) [28,29,30] and Received Signal Strength Indicator (RSSI) [31,32,33]. In underwater environments, distance information obtained by the TOA method is usually used, such as wide coverage positioning system (WPS) [34], GPS-less localization protocol (GPS-less) [35], motion-aware sensor localization (MASL) [36], underwater sparse positioning (USP) [37] and 3D underwater localization (3DUL) [38].…”

Section: Related Workmentioning

confidence: 99%

A Mobile Anchor Node Assisted RSSI Localization Scheme in Underwater Wireless Sensor Networks

Sun

Yuan

et al. 2019

In this paper, a mobile anchor node assisted RSSI localization scheme in underwater wireless sensor networks (UWSNs) is proposed, which aims to improve location accuracy and shorten location time. First, to improve location accuracy, we design a support vector regression (SVR) based interpolation method to estimate the projection of sensor nodes on the linear trajectory of the mobile anchor node. The proposed method increases the accuracy of the nonlinear regression model of noisy measured data and synchronously decreases the estimation error caused by the discreteness of measured data. Second, to shorten location time, we develop a curve matching method to obtain the perpendicular distance from sensor nodes to the linear trajectory of the mobile anchor node. The location of the sensor node can be calculated based on the projection and the perpendicular distance. Compared with existing schemes that require the anchor node to travel at least two trajectories, the proposed scheme only needs one-time trajectory to locate sensor nodes, and the location time is shortened with the reduction in the number of trajectories. Finally, simulation results prove that the proposed scheme can obtain more accurate sensor node location in less time compared with the existing schemes.