CN115291255A

CN115291255A - Distributed GNSS anomaly monitoring method suitable for vehicle-mounted end

Info

Publication number: CN115291255A
Application number: CN202211005975.4A
Authority: CN
Inventors: 赵洪博; 庄忱; 胡闪; 杨旭
Original assignee: Beihang University
Current assignee: Beihang University
Priority date: 2022-08-22
Filing date: 2022-08-22
Publication date: 2022-11-04

Abstract

The invention discloses a distributed GNSS anomaly monitoring method suitable for a vehicle-mounted end, which specifically comprises the following steps: firstly, based on shared GNSS sight distance satellite information in a workshop, a distributed and cooperative GNSS anomaly detection and elimination framework is provided; defining and generating inter-node cooperative reliability information (CRM) by utilizing a pseudo-range double-difference model; and secondly, according to the generated CRM, GNSS abnormal measurement value detection and elimination (FDE) including abnormal detection, CRM screening, LOS satellite prediction, bidirectional search and abnormal elimination are realized. The anomaly detection utilizes pseudo-range consistency detection to judge whether the GNSS is abnormal or not; performing collaborative reliability information screening and line-of-sight satellite prediction according to the abnormal detection result to obtain an alternative LOS satellite set; and finally, designing a two-level search scheme to eliminate abnormal GNSS measured values of the target vehicle. The method can effectively improve the capability of the monitoring algorithm in rejecting the multiple concurrent GNSS abnormalities.

Description

Distributed GNSS anomaly monitoring method suitable for vehicle-mounted end

Technical Field

The invention belongs to the field of satellite navigation, and particularly relates to a distributed GNSS anomaly monitoring method suitable for a vehicle-mounted end.

Background

Reliability of vehicle localization refers to the ability and likelihood of a vehicle localization system to perform localization functions within a certain time and under certain conditions without failure. Ensuring that vehicle location information is reliable is important for vehicle applications. In the traffic sector, applications relating to legal obligations and safety-related applications, corresponding requirements are made on the reliability of navigation positioning. The reliability of vehicle positioning is enhanced from two aspects, namely anomaly monitoring, namely detection and elimination of abnormal signals, measured data and the like, and alarming to a user, namely timely warning the user when the positioning error exceeds the application allowable limit value.

The core idea of the statistical Detection method is to adopt a consistency Detection algorithm to realize autonomous anomaly Detection and rejection (FDE) in a measurement domain or a positioning domain so as to check whether a measured value used for positioning is converged. The method can be divided into top-down abnormal search and bottom-up abnormal search according to an abnormal search strategy, wherein the former searches and checks abnormal measurement values which may appear one by one from a full satellite set, and the latter introduces and verifies the consistency of the rest satellite measurement values one by one on the basis of an assumed healthy satellite set, and the most classical method is a Random Sample consistency algorithm (RANSAC).

The receiver Integrity Monitoring (RAIM) adopts a top-down anomaly search strategy, is a GNSS anomaly Monitoring method for a measurement domain or a positioning domain, and is often applied to aviation users. RAIM does not need to modify a base band of a receiver, does not need additional sensors, map data and other supports, has less calculation amount and high real-time performance, is not limited by a use environment, is still a main means for solving the problems of non-line of sight and multipath of GNSS in an urban environment at present, and is applied to some high-precision RTK receivers.

Whether the RAIM method is a top-down RAIM method or the RANSAC method is a bottom-up RANSAC method, when a plurality of GNSS anomalies are concurrent, errors can occur in the anomaly removing process, namely, FDE passes consistency detection after removing identified anomaly measurement values, but the actual positioning error is still very large. This is because some healthy satellite measurements are rejected incorrectly while some abnormal measurements are still not correctly identified and continue to be used for positioning solution. The probability of the problems occurring in dense urban areas such as urban canyons is high, and a global monitoring method for a vehicle cooperative positioning system is still lacking at present, so that all possible measurement anomalies of the vehicle cooperative positioning system, including GNSS anomalies and inter-node measurement anomalies, can be comprehensively monitored.

Therefore, the distributed GNSS anomaly monitoring method suitable for the vehicle-mounted end is provided, the prior information related to the GNSS measured value quality provided by the adjacent cooperative vehicles is utilized to assist local GNSS anomaly monitoring, and the capability of a monitoring algorithm in eliminating multiple concurrent GNSS anomalies is improved, so that the method has important practical significance.

Disclosure of Invention

The invention aims to improve the abnormal rejection capability of GNSS abnormal concurrence. Although the method based on statistical detection has the capability of detecting and rejecting GNSS anomalies in real time, multiple concurrent GNSS anomalies may cause the method based on consistency detection to converge to a group of consistent but wrong satellite sets, and because GNSS anomaly measurement values are not correctly rejected, the positioning error is often very large, and in this case, the monitoring system will give dangerous misleading information, that is, the error of the actual positioning result is large but the user is informed that the measurement values are not anomalous.

In order to solve the problems, the invention provides a collaborative GNSS anomaly detection and elimination method, which is based on self-defined collaborative Reliability information (CRM) and is suitable for a GNSS positioning system in an interconnected vehicle.

The invention provides a distributed GNSS anomaly monitoring method suitable for a vehicle-mounted end, which comprises the following implementation steps of:

the method comprises the following steps: defining new workshop sharing information, namely collaborative reliability information (CRM), wherein the CRM comprises three parts, namely a timestamp, vehicle position coordinates at the moment, and a line-of-sight (LOS) satellite set identified by an information provider;

step two: according to the CRM generated in the first step, a Fault Detection and Exclusion (FDE) method of GNSS abnormal measured values assisted by collaborative reliability information is designed, so that multiple concurrent GNSS abnormalities can be more effectively identified, and the possibility of errors in consistency Detection is reduced;

wherein, in step one, the "definition of new workshop sharing information — collaborative reliability information (CRM)" is performed as follows:

s11, judging whether the satellite NLOS exists or not by using a double-difference measurement model in the LOS satellite set identified by the information provider in the CRM. Firstly, pseudo-range measurement models between a vehicle and a satellite and between a reference station and the satellite are given, and the specific formula is as follows:

where the lower subscript u denotes the vehicle, r denotes the reference station, and s denotes the satellite.

Representing the geometric distance between the vehicle and the satellite,

is the geometric distance, x, between the reference station and the satellite _u ＝(x _u ,y _u ,z _u ) Position, x, of the pointed vehicle u _r ＝(x _r ,y _r ,z _r ) Is the position of the reference station, x _s ＝(x ^s ,y ^s ,z ^s ) The position of the satellite s is referred to herein as the geocentric geostationary coordinate system. δ t _u 、δt _r And δ t ^s Respectively, a vehicle receiver clock error, a reference station receiver clock error, and a satellite clock error. When the relative distance between the vehicle and the reference station does not exceed 15km, the vehicle and the reference stationThe atmospheric time delays encountered by the reference stations can be considered similar, I ^s And T ^s Respectively, the ionospheric delay and the tropospheric delay associated with the satellite.

Is the multipath error, ε, that the vehicle u may face _ρ Is the measurement residual and can be considered as additive white gaussian noise. Since the reference station is usually placed in an open environment, it is affected by multipath only to a negligible extent.

S12, obtaining a double-difference measurement model according to the pseudo-range measurement model of S21, and calculating to obtain a double-difference pseudo-range residual error detection quantity:

the double difference measurement model can be expressed as the following formula:

here, b _ur ＝x _u -x _r Is a baseline vector between the reference station and the vehicle, and the upper corner mark represents two satellites participating in double-difference _i And s _j ，

Is vehicle to satellite s _i Direction unit vector, residual error

Can be considered zero mean gaussian noise. According to the above formula, double-differenced pseudorange residual detection quantities can be calculated:

s13, setting a proper threshold value T _LOS The threshold depends on the distribution of double-difference residual detection amount under the LOS assumption and the acceptable false alarm probability, and whether a satellite is influenced by multipath can be judged by comparing the double-difference residual detection amount with the set threshold:

all satellites are classified into an Unknown group, i.e., a set of Unknown satellites, at initialization. If the absolute value of the residual error detection quantity is less than the set threshold value, two satellites participating in double differences _i And s _j Are classified into a LOS satellite set. If the absolute value of the residual error detection quantity is greater than the threshold value, at least one satellite in the double-difference process is influenced by multipath, in this case, if one satellite in the double-difference is already classified into a LOS satellite set, it indicates that the other satellite is influenced by multipath and should be classified into a Non-line-of-sight (NLOS) satellite set. By analogy, when the system has traversed all combinations of satellites, the satellites are eventually sorted into sets of LOS or NLOS satellites.

S14, and combining the vehicles u obtained in S21 and S22 _q The set of LOS satellites in epoch t is denoted as

And the timestamp (namely the CRM generation epoch) and the position of the vehicle when the CRM is generated (namely the effective position of the CRM) are combined to form a piece of collaborative reliability information:

here, the number of the first and second electrodes,

in particular to a cooperative vehicle u at the moment t _q Sent to the target vehicle u ₀ The CRM information from different cooperative vehicles at different time points is stored in the target vehicle so as to be used for subsequent cooperative GNSS anomaly detection and elimination.

In step two, a failure Detection and Exclusion method (FDE) for GNSS abnormal measurement value assisted by cooperative reliability information is designed according to the CRM generated in step one, and the method includes:

and S21, performing pseudo-range consistency detection to judge whether the GNSS abnormality exists. Performing point positioning by using all local available pseudo range measurement values to obtain a coarse positioning result, performing consistency detection on the local GNSS pseudo range measurement values, inputting a coarse point positioning solution to a CRM screening part if an abnormal measurement value is detected, inputting an original pseudo range measurement value to an abnormal elimination part, and not performing abnormal elimination if no abnormal measurement value exists; (ii) a

And S22, screening the cooperative reliability information and predicting the sight distance satellite. If abnormal measurement values are detected in the S21 step, available CRM (customer relationship) needs to be further screened, and the LOS satellite set of the current epoch of the target vehicle is predicted by using the CRM. The aim of the method is to filter out invalid CRM information, extract LOS satellite sets in all valid CRM, and use the obtained satellite sets as prior information for initializing abnormal retrieval;

and S23, eliminating abnormal GNSS measured values based on the cooperative reliability information. A two-level search scheme is designed to eliminate abnormal GNSS measurement values of a target vehicle, and first-level search firstly needs to be combined with a local observation satellite set SV at the current moment ^all With alternative LOS satellite sets

Generating two sets of satellites SV _1,m And SV _2,m (ii) a The second-stage search generates a group of to-be-verified set SV for each alternative LOS set according to the locally observed satellite _1,m And SV _2,m The second level of search requires additional retrieval of possible NLOS satellites in SV1, m;

according to the design of the invention, the invention provides a distributed and cooperative GNSS anomaly detection and elimination framework, which is suitable for any positioning system related to GNSS in interconnected vehicles, and can be a single-vehicle GNSS positioning system, a GNSS combined positioning system or the cooperative GNSS positioning system provided by the invention.

According to the design of the invention, the invention defines a new workshop sharing information, namely collaborative reliability information (CRM), which is composed of a sight distance satellite set identified by a broadcaster, a time stamp and a measurement time position of the broadcaster. Considering that vehicles driving through the same place in a short time face similar and even identical non-line-of-sight and multi-path satellites, satellite health state historical information provided by a front vehicle can be regarded as a priori information for predicting satellite health states of a rear vehicle, and CRM can be used for carrying and transmitting the satellite health states related to time and space, so that the range of abnormal search is restricted.

According to the design of the invention, the invention designs a bidirectional search strategy to quickly and accurately eliminate GNSS abnormal measured values, compared with the RAIM, RANSAC and other methods of blind search, the method can more effectively identify a plurality of concurrent GNSS abnormalities, and reduce the possibility of errors in consistency detection.

Drawings

FIG. 1 is a diagram illustrating the problem of error consistency in the conventional RAIM method.

FIG. 2 is a diagram of a CRM-based GNSS anomalous measurement detection and culling implementation architecture.

Figure 3 is a schematic of CRM screening process.

FIG. 4 is a flow chart of a first stage search algorithm in CRM-based anomaly rejection.

FIG. 5 is a flow chart of a second level bi-directional search algorithm in CRM-based exception rejection.

Detailed Description

So that the manner in which the features, objects, and functions of the invention are attained and can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to the embodiments thereof which are illustrated in the appended drawings.

The basic application scenarios and principles of the present invention are first outlined. GNSS is the most important part of a vehicle positioning system and widely applied to various cooperative positioning methods, on one hand, GNSS can provide an absolute space-time reference for cooperative positioning, and on the other hand, also can provide vehicle ranging for cooperative positioning, abnormal GNSS measurement values will degrade performance of the GNSS cooperative positioning method, and the existence of multiple concurrent GNSS abnormalities may cause a method based on consistency detection to converge to a set of consistent but wrong satellite sets, resulting in failure to achieve expected positioning accuracy, and therefore, special attention needs to be paid to reliability of GNSS measurement. The invention provides a distributed GNSS anomaly monitoring method suitable for a vehicle-mounted end, aiming at the problem of monitoring GNSS anomaly in cooperative positioning.

As shown in FIG. 1, the present invention simplifies a positioning model to describe the problem of "error consistency". In FIG. 1, R represents a measurement of health, and

the abnormal measurement values affected by multipath and the like are shown, and it can be seen that even if the measurement values of the satellite 1 and the satellite 2 are abnormal, the three measurement values can still converge to one point (wrong positioning solution), at this time, the consistency between the measurement values is high, the measurement residual is small, the monitoring system judges that no abnormality exists, but the actual positioning result deviates from the real position.

As shown in fig. 2, how to utilize CRM to achieve fast and accurate GNSS abnormal measurement value Detection and elimination (FDE) is described, which is mainly divided into three parts, namely, abnormal Detection, CRM screening and LOS satellite prediction, bidirectional search, and abnormal elimination. The specific implementation steps are as follows:

the first step is as follows: GNSS anomaly detection method based on consistency detection

The invention adopts pseudo-range consistency detection to judge whether GNSS abnormality exists, the method is derived from classical RAIM, and the consistency detection is usually completed together with single-point positioning based on iterative least square.

Obtaining a single-point positioning solution by iterative least square estimation

After that, the pseudorange residuals may represent:

wherein

Representative of satellites s _i And the geometric distance between the single-point positioning solutions，

Is an estimate of the receiver's clock error,

and

respectively representing satellite clock error, ionospheric delay and tropospheric delay.

If there is a deviation in the pseudo-range measurement value due to multipath or the like, the performance of the least square fitting deteriorates. As a detection statistic, the sum of squares of the pseudorange residuals may be used to evaluate the performance of a least squares fit to determine if the pseudorange measurements for a set of satellites are healthy. Given a set of satellites a, the detection statistics can be calculated as:

wherein, the first and the second end of the pipe are connected with each other,

the pseudorange residual vectors representing the set of satellites, L being the number of satellites in the set of satellites to be detected. If all measurements are healthy, i.e. not affected by multipath, the pseudorange residuals

And conforms to zero mean gaussian distribution. Therefore, in the case of no anomaly, the detection statistic TS (-) should conform to DOF = L-N _state Chi-square distribution of (N) _state The number of unknown state parameters is the number, in the invention, because the GPS and Beidou dual system is used, the unknown state has receiver clock error related to the two systems besides three-dimensional space coordinates, so N _state Is 5. In order to ensure the validity of the coincidence detection, the degree of freedom needs to be 1 or more.

When there is no anomaly, the cumulative distribution function of the detection statistics is as follows:

wherein

A probability density distribution function of the detection statistic when there is no anomaly, this distribution being related to the degree of freedom DOF. Threshold value of chi-square detection

The method can be calculated according to the cumulative distribution function of the detection statistics and the set false alarm probability to obtain:

here, the first and second liquid crystal display panels are,

the probability accumulation distribution function is an inverse function of the probability accumulation distribution function of the detection statistic, if the detection statistic is smaller than the detection threshold, the detection system gives a result of abnormal judgment, otherwise, the detection system gives an abnormal warning. When the system detects an abnormality, a single-point positioning solution of the target vehicle (positioning deviation exists in abnormal measurement values due to existence of multipath and the like) is input to the next step for CRM screening. The false alarm rate is set to 0.1%, the threshold value of the card room detection can be calculated in advance, and the threshold values under different degrees of freedom can be set by adopting a table look-up mode in practical application.

The target vehicle u can be obtained in the step ₀ Coarse single point positioning solution of

The second step is that: collaborative reliability information screening and line-of-sight satellite prediction

If abnormal measurements are detected, available CRM's are screened and used to predict the LOS satellite set for the current epoch of the target vehicle. The goal of this step is to filter out the invalid CRM information and then extract the LOS satellite sets in all valid CRMs, which will be used as a priori information to initiate anomaly retrieval.

The method has scientific basis for predicting the LOS satellite of the target vehicle by utilizing the prior information provided by the cooperative vehicle. In urban environment, NLOS and multipath are mainly caused by shielding and reflection of high-rise buildings on two sides of a road, the distribution, density and height of the buildings on two sides of the road on different road sections are different, and the shielding and reflection conditions of GNSS signals are changed along with the driving of vehicles. Meanwhile, the LOS and NLOS conditions of the satellites observed in the adjacent places have similarity due to similar occlusion and reflection conditions, and the characteristic of spatial correlation is also called the spatial correlation of LOS or NLOS. In addition to the spatial location, the LOS/NLOS distribution is also time-dependent, and at the same location, the LOS or NLOS distribution will change gradually with time due to the movement of the satellite, the change is usually slow, and in a short time, the LOS and NLOS distribution of the same location observing the satellite is always kept constant, and the time-dependent characteristic is called as the time dependence of LOS or NLOS.

The LOS/NLOS space-time correlation enables the LOS satellite of the target vehicle to be predicted by the aid of the prior information of the cooperative vehicle to be feasible, and the LOS/NLOS situation similar to that of the cooperative vehicle when the cooperative vehicle runs to a nearby place is likely to be met when the target vehicle runs to the nearby place within a certain time.

(1) CRM to determine LOS/NLOS spatio-temporal correlation requirements

Since we cannot obtain an accurate positioning solution for the target vehicle before the anomaly is correctly eliminated, it is difficult to directly predict the LOS satellite of the target vehicle. One feasible idea is to search all CRM meeting the space-time correlation requirement of LOS/NLOS, store the LOS satellite set in the CRM as alternative solutions for predicting the LOS satellite set, and the alternative solutions are used in the next link one by oneAnd (6) verifying. Because LOS/NLOS has spatiotemporal correlation, not all CRM is effective for predicting local LOS satellites at the current epoch, which is why CRM screening is also required. In the present invention, in the case of the present invention,

is judged by the following formula:

wherein t is _c Represents the current epoch, t is

The time stamp in (1) is stored in the memory,

is a coarse positioning solution for the target vehicle,

is that

The effective position of the middle marker. By comparing the time difference Δ t _CRM Position difference Deltax _CRM Threshold value (eta) corresponding thereto _t And η _x ) The availability of CRM is determined. If Δ t _CRM And Δ x _CRM If the CRM is less than the corresponding threshold value, the CRM is considered to be available, otherwise, the CRM is filtered. We extract all the available LOS satellite sets in the CRM, and after the same LOS satellite set is merged, the rest different satellite sets are regarded as alternative solutions to be used for initializing the abnormal two-way searching process, and the fact that M different alternative solutions are assumed to be in total is assumed.

(2) Determining two threshold values

Referring to fig. 3, a simulation scenario is shown to illustrate how to screen available CRMs using the coarse positioning solution and the set thresholds, where the red five-pointed star represents the real position of the current epoch of the target vehicle, and the red cross represents the coarse positioning solution of the target vehicleDue to the influence of multipath, a large deviation exists between the coarse positioning solution and the real position. The solid circles represent the effective position coordinates of the markers in the CRM, and theoretically the CRM closest to the real position of the target vehicle in the effective position coordinates is the best candidate solution, since the adjacent positions imply strong spatial correlation of LOS/NLOS. Because the CRM cannot be directly found, the invention adopts a method for setting threshold value screening, and searches for the radius eta by taking a coarse positioning solution as an original point _x All valid location coordinates within the CRM are considered available. If the threshold is properly selected, the selected CRM will contain the best alternative solution (i.e., the CRM whose effective position is closest to the real position of the target vehicle). If the threshold is too small, the best CRM may be filtered out, so that the subsequent two-way search cannot find a set of satellites that meet the consistency requirement. If the threshold is too large, the computational burden of the subsequent bidirectional search is increased, because the bidirectional search needs to check the LOS satellite sets in the alternative CRMs one by one. Therefore, it is necessary to select an appropriate threshold η by comprehensively considering the performance of the algorithm and the computational burden _x 。

In subsequent experiments in urban canyons, the maximum single-point positioning horizontal error of the vehicle can reach 50 meters, and in such a scene, if the threshold eta is _x Set to 50 meters, the screened CRM will always contain the best candidate solution, therefore, the invention will be eta _x Set to 50 meters. It is noted that for the region with more serious multipath effect, the coarse positioning error before the exception is not eliminated may be larger, and the threshold value will not be applicable any more, and it is necessary to further increase the threshold value η _x So that the best candidate solution remains in the screened CRM.

For a time dependent threshold η _t Theoretically, the longer the time interval is, the greater the satellite distribution change degree is, and even in the same place, the distribution of LOS/NLOS will change along with the time, so that CRM information has a certain validity period, that is, prior information provided by cooperative vehicles can only be used in a certain time range, and eta is _t Too large of a setting, the "outdated" CRM provides a priori information that is not valid and only burdens the subsequent two-way search. Threshold eta _t Is determined byThe change frequency of the LOS/NLOS distribution of the satellite is changed, namely, the LOS satellite set (or the NLOS satellite set) changes at intervals of time at the same place. The best method for setting the threshold is to collect and analyze GNSS data in a test area, and the other method is to determine the satellite distribution change condition through simulation, so that the actual measurement method is more reliable, but the implementation process is relatively complex and long, and the simulation method is more convenient, therefore, the LOS satellite distribution change in the urban dense environment is analyzed by simulation to set the time threshold value. In the simulation, in order to keep the same with the experimental conditions as much as possible, the average height of the simulated building and the street width refer to an actual measurement scene, and the simulation satellite comprises all existing satellites of the GPS and the Beidou. After simulating 12 hours of satellite data, we found that at least about 180 seconds is required for the LOS satellite to change at the selected test point. Thus, the present invention sets a time-related threshold to 180 seconds to filter out "expired" CRM.

The output of this step: total M mutually different alternative LOS satellite sets

The third step: abnormal GNSS measurement rejection based on collaborative reliability information

(1) First level search

The invention designs a two-level search scheme to eliminate the abnormal GNSS measurement value of the target vehicle, theoretically, if the prior information provided by the cooperative vehicle meets the space-time correlation requirement of LOS/NLOS, an alternative LOS satellite set is selected

At least one satellite set exists, satellites in the set are consistent with LOS satellites observed by a target vehicle at the current moment, at this time, only the alternative LOS sets are verified one by one, the LOS satellite set meeting the consistency requirement is found, and then the rest visible satellites are retrieved.

In the second step, we screen out M alternative LOS satellite sets

The first stage of search firstly needs to combine with the local observation satellite set SV at the current moment ^all With alternative LOS satellite sets

Generating two sets of satellites SV _1,m And SV _2,m ：

Wherein, SV _1,m Is a pending local LOS satellite set, SV _2,m Is SV _1,m At SV ^all The complement in (1) represents the set of pending local NLOS satellites. It is worth noting that even if two vehicles travel in front of and behind the same location, the satellites observed at that location by the target vehicle and the cooperating vehicle may differ, since they may be equipped with different receivers, so that when taking the intersection, the SVs _1,m The number of satellites in (A) may be less than

Number of satellites, i.e. part of LOS satellites not received by the target vehicle, SV _2,m There may also be alternative sets in

Satellites not found in (c). m is initialized to 1, and SV corresponding to the alternative set m is obtained _1,m And SV _2,m Next, the present invention uses a consistency test to verify SV _1,m Whether all the satellites belong to LOS satellite or not, if SV _1,m If the consistency check is not passed, the SV of the next alternative set is continuously verified _1,m . When a certain SV _1,m By consistencyWhen detected, SV is explained _1,m The satellites in the set meet the consistency requirement, all satellites in the set are identified as LOS satellites, at which time we will continue at SV _1,m Corresponding SV _2,m The target vehicle may observe satellites not observed by the cooperating vehicle, while satellites judged as NLOS by the cooperating vehicle may also be converted to LOS satellites over time. In particular, the method of the invention relates to SV _2,m One satellite of (a) is added to the SV one by one _1,m In (1), generate

A new set of satellites is created, and,

is SV _2,m And (3) determining the number of the medium satellites, then performing consistency detection on the new satellite sets, if the satellite sets pass the consistency detection, determining the satellites added to the set as LOS satellites, and determining the satellites added to the satellite sets which do not pass the detection as NLOS satellites, so that the grouping conditions of the LOS satellites and the NLOS satellites in the current epoch of the target vehicle can be obtained:

here, the first and second liquid crystal display panels are,

is shown as SV _2,m TS (-) represents a consistency detection statistic for a group of satellites,

is the threshold for consistency detection.

If SV does not exist in the first search _1,m Through consistency detection, a second-level exception search needs to be entered. This occurs because the LOS/NLOS changes over time, and although in the last step we set a time threshold to filter out as much "outdated" CRMs as possible, the possibility that the LOS/NLOS may change in a short time cannot be excluded, which leads to SV _1,m There are still individual NLOS satellites in the system, making consistency detection impossible.

(2) Second level search

The second-level searching scheme provided by the invention introduces a top-down retrieval process on the basis of bottom-up searching, thereby forming bidirectional searching, and is shown in figure 5.

The second level of search is very similar to the first level of search, and a group of SVs to be verified are generated for each alternative LOS set according to the locally observed satellite _1,m And SV _2,m . Unlike the first level search, the second level search requires additional retrieval of SVs _1,m Possible NLOS satellites. In particular, we derive from SV _1,m Removing one satellite to generate a series of subsets, then performing consistency detection on the subsets, if the subset passes the consistency detection, all satellites in the subset are judged to be LOS satellites, and if more than one subset passes the consistency detection, selecting the subset with the minimum detection statistic for subsequent bottom-up retrieval, wherein the specific process can be represented by the following formula:

wherein

Is expressed as SV _1,m In the process of satellite navigation, a satellite,

is expressed as SV _1,m Total number of satellites in (a).

Is a newer set of alternative SVs _1,m And inputting the data into a subsequent retrieval process. If the current SV _1,m And if no subset passes consistency detection, continuously detecting the next alternative LOS set until all alternative LOS sets are searched, and at the moment, the system needs to give an alarm to the user in time to prompt that the abnormal GNSS measurement value cannot be eliminated and the positioning result may have larger deviation.

After obtaining the updated set

Then, the subsequent retrieval process is consistent with the first-stage search, namely a bottom-up search strategy is adopted, so that the first half part of the second-stage search belongs to top-down abnormal rejection, and the second half part continuously classifies the rest satellites in a bottom-up mode, so that the second-stage search is also called bidirectional search.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art to which the present application pertains. Any modification, equivalent replacement, improvement or the like made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A distributed GNSS anomaly monitoring method suitable for a vehicle-mounted end is characterized in that for a GNSS anomaly concurrency problem, cooperative reliability information (CRM) is customized, and prior information related to GNSS measurement value quality provided by adjacent cooperative vehicles is used for assisting local GNSS anomaly monitoring. The method comprises the following steps:

the method comprises the following steps: defining and generating cooperative reliability information (CRM) among nodes, and determining a line-of-sight (LOS) satellite set identified by an information provider by a design algorithm;

step two: collaborative reliability information screening and line-of-sight satellite prediction;

step three: and eliminating abnormal GNSS measured values of the cooperative reliability information, and designing a two-level search scheme to eliminate the abnormal GNSS measured values of the target vehicle.

2. The method for monitoring the anomaly of the distributed GNSS on the vehicle according to claim 1, wherein the LOS satellite set identified by the information provider in the CRM is generated as follows:

s11, obtaining a double-difference measurement model according to pseudo-range measurement models between vehicles and satellites and between a reference station and the satellites, and calculating to obtain double-difference pseudo-range residual error detection quantity;

s12, setting a proper threshold value T _LOS The threshold depends on the distribution of double-difference residual detection amount under the LOS assumption and the acceptable false alarm probability, and whether a satellite is influenced by multipath can be judged by comparing the double-difference residual detection amount with the set threshold:

s13, and combining the cooperative vehicles u obtained in S11 and S12 _q LOS satellite set at epoch t is marked as

And associate it with a timestamp (i.e., CRM generated calendar)Element) and the location of the vehicle at the time of CRM generation (i.e., the effective location of CRM), together form a piece of collaborative reliability information:

3. the collaborative reliability information of claim 2

Characterised in that it is specified to be operated by a cooperating vehicle u at time t _q Sent to the target vehicle u ₀ The CRM information from different cooperative vehicles at different moments is stored in the target vehicle, so that subsequent cooperative GNSS anomaly detection and elimination are facilitated.

4. The method of claim 1, wherein the LOS satellite of the target vehicle is predicted according to the prior information of the cooperative vehicle by using the space-time correlation of LOS/NLOS, and the LOS satellite set in all valid CRM is extracted. The specific process is as follows:

s41, searching all CRM meeting the space-time correlation requirement of LOS/NLOS, and storing an LOS satellite set in the CRM as an alternative solution for predicting the LOS satellite set:

is determined by the following equation:

wherein t is _c Represents the current epoch, t is

Time of (1)The time stamp is a time stamp,

is a coarse positioning solution for the target vehicle,

is that

The effective position of the middle marker. By comparing the time difference Δ t _CRM Position difference Δ x _CRM Threshold value (eta) corresponding thereto _t And η _x ) The availability of CRM is determined.

S42, determining a spatial correlation threshold value eta _x The method of setting threshold value screening is adopted, and the rough positioning solution is taken as the original point to search the radius eta _x All valid location coordinates within the CRM are considered available.

S43, determining a time-dependent threshold eta _t Threshold value eta _t The size of (2) depends on the change frequency of the LOS/NLOS distribution of the satellite, namely, the LOS satellite set (or the NLOS satellite set) changes once every long time at the same place. And (4) analyzing the distribution change of LOS satellites in the urban dense environment by adopting simulation to set a time threshold value.

5. The method of claim 4

The method is characterized in that LOS satellite sets in all available CRM are extracted, and after the same LOS satellite sets are combined, the rest satellite sets which are different from each other are regarded as alternative solutions to be used for initializing an abnormal bidirectional search process.

6. Determining a spatial correlation threshold η according to claim 4 at S42 _x The method is characterized in that the algorithm performance and the operation burden are comprehensively considered:

if the threshold is too small, the best CRM may be filtered out, so that the subsequent two-way search cannot find a set of satellites that meet the consistency requirement.

If the threshold is too large, the computational burden of the subsequent bidirectional search will be increased.

7. Determining a time-dependent threshold η as claimed in claim 4 at S43 _t The time-related threshold is set to 180 seconds based on the selected test point to filter out "expired" CRMs.

8. The method as claimed in claim 1, further comprising a third step of designing a two-level search scheme to eliminate abnormal GNSS measurement values of the target vehicle. The specific process is as follows:

and S81, designing a bottom-up searching scheme which is called as first-level searching. Firstly, a local observation satellite set SV at the current moment is combined ^all With alternative LOS satellite sets

Generation of two satellite sets SV _1,m And SV _2,m ：

Wherein, SV _1,m Is a pending local LOS satellite set, SV _2,m Is SV _1,m At SV ^all The complement in (1) represents the set of pending local NLOS satellites.

S82, on the basis of bottom-up search, a top-down retrieval process, namely a second-level search scheme, is introduced, so that bidirectional search is formed. Characterised by the need for additional retrieval of SVs _1,m Possible NLOS satellites.

9. The bottom-up search scheme of claim 9S 91 wherein a consistency check is used to verify SVs _1,m Whether all the satellites belong to LOS satellite or not, if SV _1,m The SV of the next alternative set is continuously verified without passing consistency check _1,m . When a certain SV _1,m When the consistency is detected, SV is illustrated _1,m The satellites in (1) meet the consistency requirement, and all satellites in the set are identified as LOS satellites, continuing at the SV _1,m Corresponding SV _2,m Potential healthy satellites are retrieved. Finally SV _2,m One satellite of one by one is added to the SV _1,m In (1), generate

A new set of satellites is created, and,

is SV _2,m And (4) determining the number of the middle satellites, then performing consistency detection on the new satellite sets, if the satellite sets pass the consistency detection, determining the satellites added to the set as LOS satellites, and determining the satellites added to the satellite sets which do not pass the detection as NLOS satellites, so as to finally obtain the grouping conditions of the LOS satellites and the NLOS satellites of the current epoch of the target vehicle.

10. The top-down search process of claim 9, wherein the top half is a top-down outlier culling, and the bottom half continues to sort the remaining satellites in a bottom-up manner. The first half from SV _1,m Removing one satellite to generate a series of subsets, then carrying out consistency detection on the subsets, if the subset passes the consistency detection, all the satellites in the subset are judged as LOS satellites, and if more than one subset passes the consistency detection, selecting the subset with the minimum detection statistic. The second half, after obtaining the updated set, keeps the retrieval process consistent with the first level search.