CN115327596A - Method for restraining RTK positioning by road plane - Google Patents

Abstract

The invention discloses a method for restraining RTK positioning by using a road plane, which specifically comprises the following steps: firstly, constructing a road plane according to positioning data of actual cooperative vehicles, wherein plane fitting based on singular value decomposition and plane detection based on fitting residual errors are mainly realized; designing a plane constraint model according to the constructed and detected plane and by combining other information of the vehicle; secondly, designing a floating point estimator of RTK (real-time kinematic) by adopting a GNSS carrier difference technology, and introducing plane information of workshop cooperation into a floating point resolving process, wherein a state vector, an observation model and a related matrix of a Kalman filter are set according to a GNSS double-difference observation model and a plane constraint model, so as to obtain a floating point ambiguity; and finally, solving the integer ambiguity of the double-difference carrier phase by using a classical LAMBDA algorithm. The invention provides a plane construction method by utilizing a positioning solution of a cooperative vehicle, so that the RTK positioning result is restrained, the floating solution is corrected on a positioning domain, the precision of the floating solution is improved, and the success rate of carrier phase ambiguity fixing is further improved.

Description

Method for restraining RTK positioning by road plane

Technical Field

The invention belongs to the field of satellite navigation, and particularly relates to a method for restraining RTK positioning by using a road plane.

Background

The precise dynamic Real-Time relative positioning of the global satellite navigation system, called RTK (Real Time Kinematic), is realized by eliminating error sources such as satellite orbit error, atmospheric propagation delay error, satellite and receiver clock error and the like by making difference between receivers and satellites, and then utilizing high-precision carrier phase observed quantity to realize positioning with centimeter-level or even millimeter-level precision. RTK positioning has the characteristics of high precision, high reliability and high real-time performance, and has wide application in civil and military fields. For example, RTK can be applied to geodetic surveying, deformation monitoring of buildings and infrastructures, lane-level high-precision automobile navigation, unmanned automobile navigation, aircraft attitude determination, precise approach landing of airplanes, precise approach landing of shipboard aircrafts and the like. The key to achieving RTK positioning is to correctly resolve the carrier-phase integer ambiguity.

The cooperative positioning is a positioning technology based on a vehicle ad hoc network, which is emerging with the vehicle networking. When the vehicles are co-located, the local measurement data can be utilized, and the measurement data shared by the vehicles can also be utilized. Compared with a positioning method of a single vehicle with multiple sensors, the cooperative positioning method has the following advantages: the measured value from the cooperative node can improve the geometric accuracy factor and improve the positioning accuracy; the number of observed dimensions and positioning equations can be increased by additional measured values, and the positioning availability is improved; most of the cooperative positioning methods do not need expensive equipment such as laser radar and the like, mainly utilize vehicle-mounted communication and low-cost distance measurement units, and realize very low cost; and complicated image processing and data matching are not involved, the occupied vehicle-mounted resources are small, and the positioning real-time performance can be ensured. The existing cooperative positioning method is realized on the basis of workshop measurement, and no cooperative positioning method can be realized by separating from the workshop measurement at present. However, due to the constraints of security, privacy, authority, equipment cost and other factors, not all cooperative vehicles have the capability of providing raw measurement information, and the vehicle to be located may not be able to obtain the measurement data of the cooperative vehicles. In addition, if the cooperative measurement data has a large delay in the transmission process, the measurement data cannot be used for positioning calculation at the current time.

Different from the measurement domain cooperative positioning method, the positioning domain cooperative positioning method realizes positioning calculation only by using the position data of cooperative vehicles. Therefore, the cooperative positioning method of the positioning domain is provided, the positioning solution of the cooperative vehicle is utilized to restrain the RTK positioning result, and the problem that the cooperative positioning method cannot be used when the workshop measurement is invalid is solved.

Disclosure of Invention

The invention aims to construct a road plane by utilizing a positioning solution of a cooperative vehicle so as to restrict the result of RTK positioning. RTK and N-RTK positioning technology based on GNSS carrier phase measurement can provide centimeter-level positioning solution for vehicle users under the condition of successfully fixing carrier phase ambiguity, however, the technology depends on satellite pseudo range and carrier phase measurement values with sufficient quantity and good observation quality, in an urban dense environment, the satellite observation degradation can be caused by the shielding of tall buildings, bridges and even trees, when the quantity of healthy satellite measurement values is reduced, the floating solution precision of RTK and N-RTK can be reduced, the carrier phase ambiguity is difficult to fix, the overall positioning precision of the vehicle is degraded due to the improvement of the floating solution ratio and the reduction of the floating solution precision in the positioning solution, and the positioning precision requirement of high-level vehicle application is difficult to meet.

The invention provides a cooperative positioning method of a positioning domain to solve the problem that the cooperative positioning method is unavailable when workshop measurement fails, and a road plane is constructed by utilizing a positioning solution of a cooperative vehicle so as to restrict an RTK positioning result.

The invention provides a method for constraining RTK positioning by using a road plane, which comprises the following implementation steps of:

the method comprises the following steps: the road plane construction method based on the cooperative vehicle position information comprises a plane fitting method based on singular value decomposition and a plane detection method based on fitting residual errors.

Step two: and (4) providing a plane constraint high-precision positioning method according to the fitted road plane obtained in the step one, designing a plane constraint-based floating point estimator, and utilizing a plane constraint RTK floating point estimation process.

Step three: and resolving the carrier phase ambiguity by utilizing an LAMBDA algorithm to obtain a high-precision fixed solution.

The road plane construction method based on the cooperative vehicle position information in the step one comprises road plane construction, plane detection and vehicle antenna height compensation, and comprises the following steps:

and S11, processing the cooperative vehicle positioning data and the antenna height. The plane construction process and the positioning calculation process uniformly use an ECEF coordinate system. In order to eliminate the influence of antenna height, a vehicle positioning result is converted into an ENU coordinate system before plane fitting, the coordinate of a phase center of a cooperative vehicle GNSS antenna under the ENU coordinate system is obtained through calculation, and the projection coordinate of the vehicle antenna on a road is obtained;

and S12, calculating a base line between the projection position of the vehicle on the road surface and the position of the base station. Converting ENU coordinates into ECEF coordinates by using a coordinate system conversion matrix S', and converting b ″ _i,k Expressed as a base line between the projection position of the vehicle on the road surface and the position of the base station under an ECEF coordinate system, wherein k represents the kth epoch, and i represents the position information of the ith vehicle which can obtain the cooperative vehicle;

and S13, selecting proper vehicle data. The baseline position information calculated in S12 is used to generate a Set of point sets for fitting a plane _fitting,k ＝{b″ _1,k ,...,b″ _N,k }. In the invention, the cooperative vehicle positioning data of a plurality of continuous epochs are adopted to fit a plane, specifically, the data of 5 continuous epochs closest to the current time is selected by default and Set is used _fitting To represent a set of points made up of the location data of these epochs;

and S14, fitting a plane alpha where the vehicle runs. The invention adopts a classical least square method to fit a plane on which the vehicle runs.

And S15, selecting the optimal solution of the fitting plane. The invention provides a method for searching an optimal solution by using singular value decomposition;

s15, for the follow-up GNSS positioning of the target vehicle, the antenna height of the target vehicle needs to be introduced into a road plane equation, and a plane beta which is parallel to the road surface and passes through the GNSS antenna phase center of the target vehicle is further constructed;

s16, in order to verify whether the fitted planes can meet the requirement that the vehicle runs on the same road surface and the positioning solution is accurate enough, the usability of the planes is detected by adopting the average residual error of plane fitting, and whether one fitted plane can be judged by comparing the fitting residual error with a set threshold value;

wherein, in the step two, the method for designing the floating point estimator based on the plane constraint and utilizing the plane constraint RTK floating point estimation process comprises the following steps:

s21, taking the plane parameters obtained by utilizing the position information of the cooperative vehicle as new observed quantities, and combining the new observed quantities with a GNSS double-difference observation model;

s22, constructing a state vector to be estimated of Kalman filtering, wherein the state vector comprises three-dimensional position, speed and acceleration and ambiguity of double-difference carrier phase, namely

Wherein the first three terms represent baseline, velocity and acceleration in the ECEF coordinate system, the last term is a double-difference ambiguity vector, the dimensionality of which depends on the number of double-difference carrier phase measurements;

s23, constructing an observation vector of Kalman filtering, wherein the observation vector comprises a pseudo-range double-difference measurement value of a GNSS, a carrier phase double-difference measurement value and parameters of a plane equation;

s23, constructing an observation matrix H, H ₀ And H _G Are all observation sub-matrices related to the double difference measurement, and H _P Is an observation submatrix related to plane constraints, H _P Element P in (1) _A 、P _B And P _C Is the plane normal vector parameter obtained in step 1;

s24, estimating a process noise matrix Q and a measurement noise matrix R, wherein the matrix Q is a constant matrix, and estimating and updating one of the noise matrices, namely the R matrix;

s25, the estimator has a risk of divergence, i.e. when the Q matrix is negative-timed or the R matrix is semi-negative-timed, and occasionally occurs in areas of severe multipath. When the above situation occurs, the invention adopts the traditional Kalman filtering noise matrix setting method

The "resolving carrier phase ambiguity by using the LAMBDA algorithm to obtain a high-precision fixed solution" described in the third step belongs to a classical method, and is not described in detail in the present invention.

Through the steps, the road plane is constructed by utilizing the positioning solution of the cooperative vehicle, the RTK positioning result is restrained, the introduced road information can restrain the positioning results of the vehicle in the horizontal and vertical directions at the same time, the floating solution is corrected on the positioning domain, the precision of the floating solution is improved, and the success rate of fixing the carrier phase ambiguity is further improved.

According to the design of the invention, the position data of the cooperative vehicle is obtained by only utilizing the workshop communication equipment, the realization cost is low, meanwhile, the method has stronger expandability, and can be combined with a multi-sensor method to further improve the positioning precision and the usability.

According to the design of the invention, the cooperative positioning can be realized only by using the historical position data of the cooperative vehicle without the GNSS original data of the cooperative vehicle, the workshop measurement data needing time synchronization such as the workshop wireless ranging value and the like, and the application range is wider and the usability is higher compared with the existing measurement domain cooperative positioning method.

According to the design of the invention, the method provided by the invention does not need to store and process map information with huge data volume, and utilizes cooperative vehicle positioning data at the latest moment. The high-precision map also has the capability of providing road plane model data, but the high-precision map information needs to be maintained and updated continuously to correct the change of the road along with the time, so the invention has strong environmental adaptability and small resource consumption.

Drawings

Fig. 1 is a diagram of a positioning process of an embodiment of the present invention.

Fig. 2 is a diagram of a method architecture of an embodiment of the present invention.

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 scenario and core principles of the present invention are first outlined. The vehicles can share respective real-time position information through a V2X link when running on a road with a flat road surface, specifically, the position refers to the position of the phase center of the GNSS main antenna of the vehicle, and the vehicles can share height information of the respective GNSS antenna from the ground besides the positions of the vehicles. When a vehicle receives the position information and antenna height information sent in conjunction with the vehicle, the vehicle can use these data to fit a plane on which the vehicle is traveling. In order to facilitate understanding of the implementation process of the method, a target vehicle is defined as a research object in the invention, and the rest vehicles are regarded as collaborators of the target vehicle.

As shown in fig. 2, the method architecture diagram of the present invention can be divided into two parts, namely, a plane construction based on cooperative data and a high-precision positioning based on plane constraint, wherein the high-precision positioning based on plane constraint can be further divided into floating point estimation and integer ambiguity resolution. The plane construction is the basis for realizing the method, and comprises road plane construction, plane detection and vehicle antenna height compensation, and the output of the step is three parameters of the plane. The floating point estimation under the plane constraint is the core of the method, the adaptive Kalman filtering is adopted to realize floating point solution, a mixed observation model is constructed, and plane parameters obtained in plane fitting are used as a new measurement value to be fused with GNSS double-difference observed quantity, and are output as the floating point solution of the vehicle to be positioned. And the final ambiguity resolution adopts a classical LAMBDA algorithm, and under the condition of ratio detection, the positioning system outputs a high-precision fixed solution, and the output result can be used as a historical positioning solution for the cooperative positioning of other vehicles. The specific implementation is as follows:

the first step is as follows: and constructing a road plane by utilizing the cooperative vehicle position information. As shown in fig. 2, the cooperative data plane based construction mainly includes two parts, namely cooperative vehicle positioning data and antenna height processing and target vehicle antenna ground distance height processing. The invention relates to two coordinate systems of data, one is an earth-centered earth-fixed coordinate system (ECEF), and the other is an EastNorth-east Earth coordinate system (ENU). The plane construction process and the positioning solution process collectively use the ECEF coordinate system, and specifically, the present invention uses the WGS84 coordinate system.

(1) Converting the vehicle positioning result into an ENU coordinate system

And defining a coordinate system transformation matrix from ECEF to ENU as S, wherein the origin of coordinates of the ENU is the phase center of the GNSS antenna of the reference station, and the coordinates of the reference station are accurately known. The calculation process of S is as follows:

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

and the included angle between the Greenwich mean square reference meridian plane and the meridian plane passing through the phase center of the antenna of the reference station is represented. Phi is the included angle between the equatorial plane and the normal line of the ellipsoid passing through the phase center of the reference station antenna, and can be obtained by iterative calculation:

wherein a and b represent the length of the major semi-axis and the minor semi-axis of the earth reference ellipsoid respectively, e is the ellipsoidal eccentricity of the earth, and N is the prime radius of curvature of the unitary fourth of twelve earthly branches of the reference ellipsoid. When calculating phi, the initial value is set to 0, and then calculated and updated according to the formula until convergence.

(2) Obtaining a vehicle data set of a fittable plane

Through the transformation matrix S, the coordinate of the phase center of the cooperative vehicle GNSS antenna under the ENU coordinate system can be calculated, and then the projection coordinate of the vehicle antenna on the road is obtained:

b′ _i,k ＝S·b _i,k -[0；0；h _i ]

where b is _i,k ＝x _i,k -x _r Representing a base line between the vehicle i and the reference station under an ECEF coordinate system, and recording the ECEF coordinate of the coordinated vehicle i in the k epoch as x _i,k The coordinate of the reference station under ECEF is x _r ，h _i The height of the GNSS antenna of vehicle i from the road surface. Through the formula, the projection position of the coordinated vehicle i on the road plane can be obtained, and then the ENU coordinate is converted into the ECEF coordinate by using a coordinate system conversion matrix S':

b″ _i,k ＝S′·b′ _i,k

here b ″) _i,k Representing the base line between the projected position of the vehicle on the road surface and the position of the base station in the ECEF coordinate system, and in the subsequent plane fitting and positioning process, the relative position (namely the base line) between the vehicle and the base station is adopted.

Assuming that the target vehicle can obtain the position information from the N cooperative vehicles in the k epoch, a Set of point sets for fitting a plane is obtained _fitting,k ＝{b″ _1,k ,...,b″ _N,k }. Generally, only one epoch of data is not enough to fit a plane, and the invention adopts cooperative vehicle positioning data of a plurality of continuous epochs to fit the plane, specifically, the data of 5 continuous epochs closest to the current moment is selected by default and Set is used _fitting To represent a set of points made up of the location data for these epochs, and by default all the location solutions within this set are used to fit a plane.

(3) Fitting a plane using known data

The invention adopts a classical least square method to fit a plane alpha on which the vehicle runs. Here, a number of point coordinates b "for fitting a plane are assumed _i,k ＝(x″ _i,k ,y″ _i,k ,z″ _i,k ) Assuming that the plane of the fit is

α:P _A x+P _B y+P _C z+P _D ＝0

Wherein P is _A 、P _B And P _C Is a plane normal vector. Given a constraint, normalizing the plane normal vector coordinates:

P _A ² +P _B ² +P _C ² ＝1

assume that the average coordinate of all positioning solutions used for plane fitting is

Then

Setting a matrix A as:

here, n represents the number of fitting points, X = [ P = _A ；P _B ；P _C ]The equation is equivalent to AX =0.

In practice, because a road is not a strict plane, and there is an error or even a large deviation in the positioning data, the above formula is not true, so the essence of the fitting is to find a plane, so that the sum of distances between all points for fitting and the plane is minimized, that is, a target function min | | AX | |, where | | X | =1 is set. The invention adopts a singular value decomposition method to search an optimal solution, and supposes that A can be subjected to singular value decomposition:

A＝UDV ^T

where U and V are unitary matrices and D belongs to a diagonal matrix, then | | | AX | = | | UDV ^T X||＝||DV ^T X | |, here V ^T X is a column matrix and satisfies | | | V ^T X||＝||X||＝1。

Since the diagonal elements of D are singular values, assume that its last diagonal element isCorresponding to the smallest singular value, then if and only if V ^T X＝[0；0；…；1]When the value of AX is minimum, the value is minimum, and then

Therefore, the optimal solution of min | | AX | | | is the feature vector corresponding to the minimum singular value, and this feature vector is the normal vector of the fitting plane:

next, P is calculated _D ：

(4) Constructing a plane beta parallel to the road surface and passing through the phase center of the GNSS antenna of the target vehicle

The plane of the target vehicle antenna can be represented as follows:

β:P _A x+P _B y+P _C z+P′ _D ＝0

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

here h is _target The antenna representing the target vehicle is high.

The second step is that: detecting usability of a plane using an average residual of a plane fit

The specific calculation method is as follows:

wherein the content of the first and second substances,

where N denotes the number of cooperating vehicles, k _start And k _end Respectively representing the beginning and ending epochs of the positioning solution used in the fitting, d _i,k Represents a point b ″) _i,k ＝(x″ _i,k ,y″ _i,k ,z″ _i,k ) From the fitting plane, if all the cooperative vehicles travel in one plane and their positioning solutions are accurate enough, the average residual value of the fitting plane will be very small, otherwise the residual will become large. Thus, whether a fitted plane is available can be determined by comparing the fitted residual with a set threshold:

wherein T is _fitting Represents a threshold for plane detection, the size of which depends on the distribution of the plane fit residuals and the false alarm rate.

The third step: designing a floating point estimator for RTK

(1) Determining a state vector to be estimated for Kalman filtering

The Kalman filtering state vector comprises a three-dimensional position, a speed and an acceleration of a vehicle to be estimated, and ambiguity of a double-difference carrier phase:

the first three terms represent the baseline, velocity and acceleration in the ECEF coordinate system, respectively, and the last term is a double-difference ambiguity vector whose dimensionality depends on the number of double-difference carrier phase measurements.

(2) Prediction

The prediction process for floating point estimation is expressed as follows:

X _k,k-1 ＝FX _k-1 +ω _k ,ω _k ～N(0,Q)

P _k,k-1 ＝FP _k-1 F ^T +Q

wherein, X _k,k-1 Is the state vector of the current epoch predicted from the previous epoch, X _k-1 Is the state vector of the last epoch estimate and F is the state transition matrix. Omega _k And Q represents the noise of the prediction process and the covariance matrix of the process noise, P _k,k-1 Covariance matrix, P, representing prediction error _k-1 The estimated error matrix for the previous epoch.

The state transition matrix for the floating point estimate may be expanded as follows:

F＝diag(F ₁ ,F ₂ )

F ₂ ＝diag(I _m )

where τ represents the update period of Kalman filtering, I _m Is an identity matrix of dimension m.

(3) Measurement update

The observation model of kalman filtering can be expressed as:

Z _k ＝HX _k +v _k ,v _k ～N(0,R)

wherein Z is _k An observation vector representing the current epoch, H represents an observation matrix, v _k And R represents the measurement noise and the covariance matrix of the measurement noise, respectively. According to the GNSS double-difference observation model and the plane constraint model derived in the previous subsection, the observation matrix can be represented as:

wherein the content of the first and second substances,

H _G ＝diag([...,λ,...] _1×m )

H _P ＝[P _A P _B P _C ]

u in the above formula _i Representing a unit direction vector, u, pointed at the satellite i by the vehicle _p The lower subscript p in (1) refers to the double-differenced selected primary satellite. m is the number of double-differenced measurements for the current epoch, assuming that the number of double-differenced pseudorange measurements corresponds to the number of double-differenced carrier-phase measurements, H ₀ And H _G Are all observation sub-matrices related to double difference measurements, and H _P Is an observation submatrix related to plane constraints, H _P Element P in (1) _A 、P _B And P _C Is the plane normal vector parameter calculated in the previous section.

Thus, the Kalman filter gain may be expressed as:

the measurements are updated as follows:

X _k ＝X _k,k-1 +K _k (Z _k -H _k X _k,k-1 )

P _k ＝(I-K _k H _k )P _k,k-1

wherein, X _k Estimated state vector for current epoch, P _k The covariance matrix of the estimated error for the current epoch. The method realizes the estimation of the floating solution, obtains the floating state vector and the floating ambiguity, and fixes the integer ambiguity and corrects the floating solution.

The fourth step: noise matrix selection

The present invention employs information-based adaptive kalman filtering to update the measurement noise matrix R, and it is noted that if the process noise matrix Q and the measurement noise matrix R are estimated simultaneously, the estimator may be misled by the correlation between them and diverge, in order to avoid this problem, only one of the noise matrices, i.e., the R matrix, is estimated and updated in the example, while the other Q matrix is still an unusual matrix.

Innovation is defined as the difference between the observed and predicted values, and a specific innovation vector can be expressed as:

e _k ＝Z _k -H _k X _k,k-1

the covariance matrix of the innovation sequence can be expressed as:

in practical use, the matrix is calculated as follows:

where k represents the current epoch and L is the window length of the innovation sequence. If the number of equations used to estimate the adaptive noise matrix is less than the number of unknown parameters in the noise matrix, the estimation process is at risk of divergence. Thus, the window length cannot be too small, and should be larger than the number of measurements when updating the measurement noise matrix R, and larger than the filter state vector dimension when updating the process noise matrix Q.

Theoretically, the innovation covariance matrix should satisfy the following relationship:

the Q and R matrices can be calculated as follows:

although both Q and R estimation methods are presented here, the invention updates only the R matrix and the Q matrix is set to the constant matrix when adaptive filtering is used. It is worth noting that if the Q matrix is negative or the R matrix is semi-negative, the estimator is at risk of divergence, which happens occasionally in severe areas of multipath. When the above situation occurs, the present invention still adopts the conventional kalman filter noise matrix setting method.

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. 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 (10)

1. A method for restraining RTK positioning by using a road plane is characterized by comprising the following steps: a road plane is constructed by utilizing a positioning solution of the cooperative vehicle, so that an RTK positioning result is restrained, and the positioning results of the vehicle in the horizontal and vertical directions are restrained by introducing road surface information. The method comprises the following steps:

the method comprises the following steps: performing plane fitting based on singular value decomposition by using the cooperative vehicle position information;

step two: performing plane detection based on the fitting residual error by using the cooperative vehicle position information;

step three: according to the fitted road plane obtained in the first step and the second step, a high-precision positioning method using plane constraint is provided, a floating point estimator based on plane constraint is designed, and the RTK floating point estimation process is constrained by the plane.

2. The method for RTK positioning using road plane constraint according to claim 1, wherein the cooperative positioning is achieved by using only the historical position data of the cooperative vehicle without additional sensors, without the GNSS raw data of the cooperative vehicle and the workshop measurement data requiring time synchronization such as the wireless ranging value of the workshop; the specific process of the step one is as follows:

s21, because the heights of the vehicles are different, the relative heights of the GNSS antenna and the road plane are different, and in order to ensure that the fitting result correctly reflects the road plane on which the vehicles run, the positioning data is preprocessed, and the plane is fitted after the influence of the heights of the antennae is eliminated.

S22, calculating a base line b' between the projection position of the vehicle on the road surface and the position of the base station _i,k 。

And S23, selecting proper vehicle data. The baseline position information calculated in S12 generates a Set of point sets for fitting a plane _fitting,k ＝{b″ _1,k ,...,b″ _N,k }。

In the invention, the cooperative vehicle positioning data of a plurality of continuous epochs are adopted to fit a plane, specifically, the data of 5 continuous epochs closest to the current time is selected by default and Set is used _fitting To represent a set of points made up of the location data of these epochs;

and S24, fitting a plane alpha where the vehicle runs. The invention adopts a classical least square method to fit the driving plane of the vehicle.

And S25, constructing a plane beta which is parallel to the road surface and passes through the phase center of the GNSS antenna of the target vehicle.

3. Base line b "between base station positions according to S12 in claim 2 _i,k And k represents the kth epoch, and i represents the position information of the ith cooperative vehicle, which represents the base line between the projected position of the vehicle on the road surface and the position of the base station in the ECEF coordinate system.

4. The Set of plane points of claim 2S 23 _fitting,k ＝{b″ _1,k ,...,b″ _N,k And the method is characterized in that collaborative vehicle positioning data of a plurality of continuous epochs are adopted to fit a plane, and data of 5 continuous epochs closest to the current time are selected.

5. The plane α along which the vehicle travels of claim 2, wherein a plane is fitted based on the set of points formed by the positioning data, the sum of the distances between all the data and the plane is minimized, and the optimal solution is found by using singular value decomposition.

6. The plane β for establishing a parallel road surface and passing through the GNSS antenna phase center of the target vehicle of claim 2 wherein the antenna height of the target vehicle is introduced into the road plane equation.

7. The fitting residual-based performing plane detection as claimed in claim 1, wherein a suitable threshold is set according to the average residual distribution and the false alarm rate of the plane fitting, and the set threshold is compared with the fitting residual to determine whether the fitting plane is usable:

8. the method of claim 1, wherein the positioning process of the target vehicle is constrained by using plane information, and the plane information obtained by the cooperation of the workshop is introduced into a floating point solution process in the design of the RTK floating point estimator, wherein the specific process is as follows:

and S81, setting a state vector and an observation vector to be estimated by Kalman filtering in consideration of plane constraint information.

And S82, setting relevant vectors and matrixes in the process based on plane constraints aiming at the prediction and measurement update of the floating point estimation.

9. The method of S81 of claim 8 wherein the kalman filtered observation vector comprises plane equation parameters in addition to GNSS pseudorange double difference measurements and carrier phase double difference measurements, and is specified by the following equations:

10. the method for setting correlation vectors and matrices in prediction and measurement update as claimed in S82 of claim 8, wherein the observation matrix H incorporates plane constraint information, which can be expressed as:

wherein the content of the first and second substances,

H _G ＝diag([...,λ,...] _1×m )

H _P ＝[P _A P _B P _C ]

H _P is an observation submatrix related to plane constraints, H _P Element P in (1) _A 、P _B And P _C Is the plane normal vector parameter calculated in the previous section.

Images