CN-121995315-A - Reliable receiver subset selection method and system under non-line-of-sight condition

CN121995315ACN 121995315 ACN121995315 ACN 121995315ACN-121995315-A

Abstract

A reliable receiver subset selection method and system under a non-line-of-sight condition belong to the technical field of arrival time difference positioning, and solve the problem that false low residual errors but positioning errors are caused by selecting a subset only depending on minimum fitting residual errors under NLOS conditions in the prior art. The method comprises the steps of constructing a candidate receiver subset, determining a traversing sequence, introducing balance parameters to form an expansion set, screening to determine a reliable newly-increased receiver set, combining the candidate receiver subset and the reliable newly-increased receiver set to obtain a reconstructed receiver set, and carrying out reference consistency check on the reconstructed receiver set to obtain a final reliable receiver set. The method and the device are suitable for the arrival time difference positioning scene.

Inventors

ZHAO YAQIN
DING QINYU
WU LONGWEN
YANG ZHUTIAN
HU XIANGRUI
XU SIYUAN

Assignees

哈尔滨工业大学

Dates

Publication Date: 20260508
Application Date: 20260206

Claims (10)

1. A method of reliable receiver subset selection under non-line-of-sight conditions, comprising the steps of: S1, constructing candidate receiver subsets, selecting any one of the receivers as a reference receiver for each candidate subset, adopting weighted least square cost to represent the fitting performance of the candidate subset, calculating to obtain cost values of the candidate subsets, sequencing all the candidate subsets from small to large according to the cost values, and determining a traversing sequence; S2, introducing balance parameters, selecting any receiver which does not participate in the formation of the current candidate subset based on the sorted candidate subset, and combining the receiver with the current candidate subset to form an extended set, fixing the reference receiver in the extended set, and jointly estimating the position of a transmitter and the balance parameters by using TDOA measurement data in the extended set; S3, combining the candidate receiver subset in the S1 with the reliable newly-increased receiver set obtained in the S2 to form a reconstruction receiver set; s4, carrying out reference consistency check on the reconstructed receiver set to obtain a final reliable receiver set.
2. The method of claim 1, wherein in S1, the candidate receiver subsets are constructed according to a predetermined number, the predetermined number being 4.
3. The method for reliable receiver subset selection under non-line-of-sight conditions of claim 1 wherein in the weighted least squares cost calculation, the residual is the difference between the TDOA measurement and the ideal TDOA value, and the weight is a preset weight coefficient.
4. The method according to claim 1, wherein in S2, the joint estimation problem is solved by at least one optimization algorithm selected from the group consisting of an alternate direction multiplier method and a gradient method.
5. The method for selecting reliable receiver subsets under non-line-of-sight conditions according to claim 1, wherein in S2, the specific judgment logic of the screening is that if the balance parameters corresponding to newly added receivers meet the preset threshold condition, the extended set is determined to contain NLOS polluted receivers and reject the current candidate subset, if the balance parameters corresponding to all newly added receivers meet the reliable condition, the newly added receivers are classified into the reliable newly added receiver set, and if the balance parameters corresponding to the newly added receivers do not meet the reliable condition, the newly added receivers are determined to be influenced by NLOS and are eliminated.
6. A reliable receiver subset selection method under a non-line-of-sight condition according to claim 1 is characterized in that S4 comprises traversing each receiver in a reconstructed receiver set as a reference receiver, performing weighted least squares positioning by using TDOA measurement data of the rest of the receivers in the set to obtain a plurality of position estimation values, calculating an average position of the plurality of position estimation values, determining a maximum deviation degree of each position estimation value relative to the average position, comparing the maximum deviation degree with a preset consistency threshold, determining that the reconstructed receiver set is a reliable receiver subset if a consistency condition is met, outputting the average position as a final positioning result, selecting the next candidate subset according to a traversing sequence of S1 if the consistency condition is not met, repeating S2 to S4 until a reliable receiver subset meeting the consistency condition is obtained as the final reliable receiver set, and outputting a positioning result.
7. A method for reliable receiver subset selection under non-line-of-sight conditions as recited in claim 1, wherein in S2, using TDOA measurement data within the extended set, jointly estimating the transmitter position and the balance parameter is expressed as: , Wherein, the Indicating the location of the transmitter(s), The balance parameter is represented by a value of the balance parameter, The weight is represented by a weight that, , Representing a TDOA measurement model of the device, Representing the coordinates of the receiver and, Representing the signal propagation velocity.
8. A reliable receiver subset selection system in non-line-of-sight conditions, the system being implemented based on a reliable receiver subset selection method in non-line-of-sight conditions as claimed in any one of claims 1-7, the system comprising: the subset construction and cost ordering module is used for constructing candidate receiver subsets, selecting any one of the candidate receivers as a reference receiver for each candidate subset, adopting weighted least square cost to represent the fitting performance of the candidate subset, calculating to obtain cost values of the candidate subsets, ordering all the candidate subsets from small to large according to the cost values, and determining a traversing sequence; The subset expansion and balance parameter detection module is used for introducing balance parameters, wherein the balance parameters are used for describing bias related to reference receivers, selecting any receiver which does not participate in the formation of the current candidate subset based on the sorted candidate subset, and combining the receiver with the current candidate subset to form an expansion set; The reliable receiver set reconstruction module is used for combining the candidate receiver subset in the subset construction and cost sorting module with the reliable newly-added receiver set obtained by the subset expansion and balance parameter detection module to form a reconstructed receiver set; And the reference consistency checking module is used for carrying out reference consistency checking on the reconstructed receiver set to obtain a final reliable receiver set.
9. A computer device, characterized in that it comprises a memory in which a computer program is stored and a processor which, when running the computer program stored in the memory, performs a reliable receiver subset selection method according to any one of claims 1-7 in a non-line-of-sight condition.
10. A computer readable storage medium for storing a computer program for performing the reliable receiver subset selection method under one of the non-line-of-sight conditions of any one of claims 1-7.

Description

Reliable receiver subset selection method and system under non-line-of-sight condition Technical Field The invention belongs to the technical field of arrival time difference positioning, and particularly relates to a reliable receiver subset selection method under a non-line-of-sight condition. Background Time difference of arrival (TDOA) positioning is widely used in various types of positioning systems because absolute time synchronization between the transmitter and receiver networks is not required. However, in non line-of-sight (NLOS) environments, the direct path tends to be blocked, and multipath components introduce additional propagation delay, resulting in TDOA measurements exhibiting forward bias, with significantly reduced positioning accuracy. To mitigate NLOS effects, one common approach is to replace the traditional secondary loss function with a robust loss function to reduce the dominant contribution of large residuals to the estimation results. Typical robust loss functions include Smooth L1, cauchy, geman-McClure, and the like. It is noted that the effectiveness of such methods generally depends on the assumption of "outlier sparseness," and the ability of the robust penalty function to reject outliers will be significantly reduced when the NLOS measurement scale is high. Another class of methods employs explicit bias modeling, treating NLOS effects as parameterized unknown biases, typically by introducing balance parameters or jointly estimating position and bias. Such methods mitigate NLOS effects by estimating and compensating for bias. However, when modeling biases for multiple measurement links simultaneously, the problem easily becomes underdetermined or unrecognizable, thereby increasing the difficulty of optimization. To solve this problem, existing methods limit the resolvable region and guarantee resolvability by imposing stricter spatial constraints, and studies have been made to estimate only the bias of the reference receiver or reference path, thereby avoiding the underdetermined problem. In recent years, robust convex optimization methods based on a minimum and maximum (min-max) framework have also been developed. Such methods typically require an upper bound to the NLOS bias and minimize the worst-case positioning error within the corresponding uncertainty set, thereby obtaining a more stable solution under bias disturbances. Under the framework, robust convex approximation methods for TDOA-based localization under NLOS conditions proposes two approximate solving strategies based on convex relaxation, and the two approximate solving strategies show good stability and robustness under different NLOS conditions. In order to alleviate conservation and performance loss ,"Consistent and accurate TDOA localization in dynamic NLOS environments via prior information independent estimation" caused by overlarge error bound or insufficient constraint of triangular inequality, an improved RLS form is further provided, and balance parameters are introduced and the practicability of the method is improved by utilizing the S-theorem. In addition, there are also methods directed to NLOS identification and measurement selection, distinguishing LOS from NLOS measurements by geometric consistency or signal characteristics, or detecting outliers, and then locating based on a subset of the measurements or selected anchors after screening. The correlation work includes a residual threshold based detection method and a consistency based TDOA subset selection method. The main risk of this type of approach is that the NLOS measurements are preserved by error, and once biased NLOS measurements are incorporated into the positioning solution, the positioning results will deteriorate significantly. However, theoretically, if LOS/NLOS recognition is sufficiently accurate, locating based on a clean measurement set is closest to the ideal LOS scene and therefore has the potential to achieve near optimal accuracy. Although the prior art identification techniques alleviate to some extent the effect of NLOS on TDOA location, the following deficiencies still exist: (1) The receiver subset selection method based on the least fit residual is not reliable enough: The prior art often uses a minimum fit cost or minimum residual criterion to select the subset that minimizes TDOA fit error from among the multiple receiver subsets for positioning. However, under NLOS conditions, biased measurements may geometrically form "false coincidences," and even if the positioning result deviates significantly from the true position, smaller fitting residuals may still be produced, resulting in a wrong subset being selected, reducing positioning reliability. (2) The existing method is difficult to fully utilize the unilateral characteristic of NLOS bias: The extra delay introduced by NLOS propagation has non-negative unilateral characteristics, but many existing methods cannot effectively utilize the prior information in the