Networking joint positioning method and device for local seabed control point group

By constructing an overall positioning model containing relative observation data of sea surface and seabed and using the difference-resistant adaptive selection Kalman filtering algorithm, the problem of low positioning accuracy and efficiency of subsea acoustic transponder network is solved, and a more efficient subsea control point grouping is achieved position.

CN119064940APending Publication Date: 2024-12-03CHINESE PEOPLES LIBERATION ARMY KET FORCE SERGEANT SCHOOL
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411302342.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-18
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The prior art is limited in the network positioning of subsea acoustic transponders, and the efficiency is not high. It is mainly because it only relies on absolute observation data between the sea surface and a single subsea acoustic transponder, and the relative measurement between subsea acoustic transponders is not fully utilized. Distance mode.

Method used

By obtaining GNSS observation data, hull attitude observation data, water depth measurement data, sound velocity profile data and underwater acoustic absolute and relative observation data, we construct sea surface observation equations, sea surface-sea bottom absolute observation equations, and relative observation equations and depths between sea bottom points The overall positioning model of the constrained equations is used to filter and solve the filter using the anti-difference adaptive selection Kalman filtering algorithm, making full use of the relative observation data between seabed points.

Benefits of technology

It improves the positioning accuracy and efficiency of the subsea control point group, and can more effectively utilize the relative observation data between subsea points, reduces errors and improves calculation efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064940A_ABST
    Figure CN119064940A_ABST
Patent Text Reader

Abstract

The invention relates to a networking joint positioning method and device for a local submarine control point group, and belongs to the technical field of underwater acoustic positioning. According to the method, on the basis that data of a single seabed acoustic transponder is adopted in the prior art, a relative ranging mode between seabed acoustic transponders is fully considered, a relative observation equation set between seabed points is established through relative ranging data of the seabed acoustic transponders, and the relative observation equation set serves as a part of an overall positioning model; compared with a mode of independently solving each seabed acoustic transponder one by one, the method provided by the invention has the advantages that a networking observation model of the seabed control point group is established by utilizing a mutual measurement mode among the points, so that relative observation data among the seabed points can be fully utilized in a solving process, and the positioning efficiency and precision are further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a networking joint positioning method and device for a local seabed control point group, belonging to the technical field of underwater acoustic positioning. Background Art

[0002] Building seafloor geodetic benchmarks and forming a regional seafloor geodetic control network similar to the land observation network is an important basic task for countries to carry out marine activities. The more mature technical concept is to combine sea surface satellite positioning and underwater acoustic positioning to determine the seafloor control point position (i.e. GNSS-Acoustic, hereinafter abbreviated as GNSS-A). Through this sea surface-seafloor absolute measurement mode, the three-dimensional position benchmark of the sea surface (or land) can be transferred to the seafloor.

[0003] However, in practical applications, it is found that relying solely on the absolute observation mode is inefficient, takes a long time per time, and is prone to introducing sound velocity time domain ranging errors. In view of the drawbacks of a single observation mode, it is necessary to adopt a method of absolute measurement combined with relative measurement to complete the point determination of the local seabed control network. That is, using the GNSS-A absolute measurement mode, combined with the distance information measured between the seabed control points, the absolute geographic coordinates of each point in the local seabed control network of the measurement area are solved. When solving ocean control points, most existing studies only discuss the data processing of the sea surface-seabed absolute observation mode, or adopt the idea of ​​independent solution of underwater (seabed) observation information based on prior position information: the observation results of the sea surface platform (absolute observation solution results) are used as prior information, and the underwater (seabed) measurement process is independently solved based on this information.

[0004] For example, the Chinese patent application document with application publication number CN111208520A discloses a method and device for positioning a submarine acoustic transponder. The solution uses GNSS observation data, hull attitude angle observation data, sound velocity profile data, underwater acoustic signal round-trip time and water depth data to establish a positioning model that includes a depth difference constraint between the ship bottom transducer and the submarine acoustic transponder, and uses a constrained Kalman filter algorithm to solve the positioning model to achieve the positioning of the submarine acoustic transponder. Although the solution can achieve the positioning of the submarine acoustic transponder, the positioning model established by the solution is aimed at a single submarine acoustic transponder. In the underwater solution process, only the time delay data between the sea surface and the single acoustic transponder on the seabed is used, and no consideration is given to how to use the relative propagation delay between transponders after the submarine acoustic transponders are networked, resulting in limited accuracy and low efficiency when the solution is applied to the positioning of submarine acoustic transponders. Summary of the invention

[0005] The purpose of the present invention is to provide a networking joint positioning method and device for a local seabed control point group, so as to solve the problem that the positioning accuracy is limited and the positioning efficiency is low when only absolute observation data between the sea surface and a single seabed acoustic transponder is used to perform positioning and solution of the local seabed control point group.

[0006] In order to solve the above technical problems, the present invention provides a method for joint positioning of a local seabed control point group, the positioning method comprising the following steps:

[0007] 1) Obtain GNSS observation data, ship attitude observation data, water depth measurement data, sound velocity profile data, underwater acoustic absolute and relative observation data;

[0008] 2) Using the acquired data, an overall positioning model is constructed, which includes the sea surface observation equation, the sea surface-sea bottom point absolute observation equation, the relative observation equation between sea bottom points and the depth constraint equation. The relative observation equation between sea bottom points is established based on the distance observation quantity converted from the relative time delay data between the transponders on the sea bottom.

[0009] 3) Filter and solve the established overall positioning model to obtain the position of each seabed transponder to achieve the positioning of the seabed control point group.

[0010] Furthermore, the relative observation equation between seabed points is:

[0011]

[0012] Among them SD * represents the mutual distance observation vector between each transponder; f(*) represents the distance vector between two coordinates; are the coordinates of each transponder; Represents the mutual measurement distance noise vector.

[0013] Furthermore, the depth constraint equation is:

[0014]

[0015] in, is the measured depth difference between each transducer and each transponder on the seabed; R′ represents the spatial reference conversion matrix.

[0016] Furthermore, in step 3), the Helmert variance component estimation method is used to adjust the random model of the overall positioning model during filtering and the variance component estimation between heterogeneous observations of different accuracies is iteratively updated to make the weight ratio relationship between the observations gradually reasonable.

[0017] Furthermore, in the step 3), a robust adaptive weighted Kalman filtering algorithm is used for filtering solution.

[0018] Furthermore, the adaptive weight selection strategy in the robust adaptive weighted Kalman filtering algorithm is:

[0019] The given prediction residual statistics or variance component ratio statistics are used to calculate the judgment evaluation standard, and the adaptive factors are constructed for each parameter by category, and then an adaptive factor matrix is ​​formed to participate in the filtering solution in the form of an adaptive factor matrix.

[0020] Furthermore, the filtering solution result is:

[0021]

[0022] in A is the prediction vector filtered at time k; k is the coefficient matrix, W k is the adaptive factor matrix; is the equivalent weight matrix, is the weight matrix corresponding to the prediction parameter vector; is the new information vector, and I is the unit matrix.

[0023] Furthermore, the data used in step 2) are the data preprocessed in step 1), and the preprocessing includes gross error elimination and time alignment.

[0024] Furthermore, the preprocessing also includes determining the initial grazing angle from the original underwater acoustic relative observation time delay data based on a strict equal-gradient ray tracking model, and then obtaining the corrected observation slant distance between the seabed points.

[0025] The present invention also provides a networking joint positioning device for a local seabed control point group, comprising a processor, wherein the processor is used to execute a computer program to implement the networking joint positioning method for the local seabed control point group of the present invention.

[0026] The beneficial effects of the present invention are: as an improved invention, the present invention fully considers the relative ranging mode between seabed acoustic transponders on the basis of the existing use of single seabed acoustic transponder data, and uses its relative ranging data to establish a relative observation equation group between seabed points, and uses the relative observation equation group as part of the overall positioning model. Compared with the method of independently solving each seabed acoustic transponder one by one, the present invention uses the mutual ranging mode between these points to establish a networked observation model of the seabed control point group, so that the relative observation data between the seabed points can be fully utilized in the solution process, further improving the efficiency and accuracy of positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the principle of the networking joint positioning method of the local seabed control point group of the present invention;

[0028] Figure 2 It is a flow chart of the networking joint positioning method of the local seabed control point group of the present invention. DETAILED DESCRIPTION

[0029] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings.

[0030] The current integrated PNT (Positioning, Navigation and Timing) system is becoming more and more flexible, and the processing of ocean spatiotemporal reference data has developed towards a multi-source observation information fusion mode. Conventional methods only build models based on absolute observation or relative observation modes, which is not complete enough and cannot be integrated into the flexible PNT system; and this error-independent approximate processing strategy has low computational efficiency, is not conducive to the efficient realization of accurate calibration of local seabed control network areas, and is prone to introduce additional errors. To this end, the present invention, based on the existing absolute observation data of a single acoustic transponder from the sea surface to the seabed, fully considers the relative ranging mode between seabed acoustic transponders, and uses its relative ranging data to establish a relative observation equation group between seabed points, and uses the relative observation equation group as part of the overall positioning model, so that the relative observation data between seabed points can be fully utilized in the solution process, thereby improving the efficiency and accuracy of positioning.

[0031] Embodiment of a networked joint positioning method for a local seafloor control point group

[0032] The positioning method of the submarine acoustic transponder of the present invention is achieved by combining GNSS positioning and acoustic positioning technology, such as Figure 1 As shown, the survey vessel is provided with a GNSS antenna for obtaining the positioning data of the survey vessel on the sea surface through GNSS satellites. The bottom of the survey vessel is equipped with an acoustic transducer, which can establish a connection with the seabed acoustic transponder deployed on the seabed through sound waves, that is, the acoustic transducer sends sound wave signals to the seabed and receives sound wave signals returned from the seabed acoustic transponder. The seabed transponders can also measure each other, and relative observation data can be obtained. The present invention solves the established overall positioning model including the joint observation equations and depth constraint equations of the sea surface, sea surface-seabed points, and seabed points through GNSS observation data, hull attitude angle observation data, sound velocity profile data, underwater acoustic absolute and relative observation data, and water depth data, thereby realizing the positioning of the seabed acoustic transponder. The specific implementation process is as follows. Figure 2 As shown, the following steps are included.

[0033] 1. Obtain GNSS observation data, ship attitude angle observation data, water depth measurement data, sound velocity profile data, underwater acoustic absolute and relative observation data.

[0034] Specifically, GNSS observation data is obtained through a GNSS antenna mounted on a sea surface platform. The GNSS antenna is used to receive GNSS satellite signals to determine the virtual reference space position of the sea surface platform at different times; the hull attitude observation data is obtained through a heading and attitude measurement system mounted on the sea surface platform. The system is mainly used to collect real-time attitude angle changes of the carrier, combined with the relative displacement from the center of the GNSS antenna to the center of the transducer (or arm vector), which can realize the conversion of the sea surface reference position from the center of the GNSS antenna to the center of the transducer array; the water depth measurement data is measured by a temperature and salinity depth meter and a water level pressure gauge, which can measure the depth information of the transducer and the transponder; the sound velocity profile data is measured by an acoustic measuring instrument; the underwater acoustic absolute and relative observation data are obtained by transmitting and receiving acoustic signals between the bottom transducer and the seabed transponder, and between the seabed transponders, to obtain the signal propagation delay observations.

[0035] After the above raw observation data are collected, the following preprocessing steps are required, including:

[0036] 1) Eliminate gross errors.

[0037] There are obvious large errors in the original observation information, which need to be eliminated. This embodiment eliminates the obvious large errors in the collected GNSS observation data, hull attitude data, sound speed profile data, underwater acoustic absolute and relative observation data.

[0038] 2) Combine, layer and appropriately simplify the sound velocity profile data based on the depth information.

[0039] Combined with the water depth information, adjacent depth intervals with unchanged sound velocity values ​​are merged. After drawing the sound velocity distribution curve, the characteristic points of the curve are selected as representative values ​​of the sound velocity distribution on the basic principle of ensuring equal gradients of sound velocity within the layer. This is used to form a streamlined sound velocity profile to achieve the purpose of simplifying the profile.

[0040] 3) Temporal alignment of measurement information from different types of sensors.

[0041] Since there are many types of sensors used in the present invention, time alignment is required. In this embodiment, GPS time or UTC time is selected as a unified time reference, and the time when the selected sensor starts stable measurement is used as the basis. The measurement data of other sensors before this time is eliminated, and the missing data of the epoch to be calculated is interpolated through a fixed number of existing node data to ensure time consistency.

[0042] 4) Benchmark offset compensation.

[0043] The measurement information of the temperature-salinity-depth instrument provides high-precision shipborne transducer depth information. There is a reference offset between the coordinates of the transducer in the water depth direction in the self-built local ocean coordinate system. In order to avoid the influence of systematic errors on the results, it is necessary to compensate the offset of the transducer coordinates and the approximate transponder coordinates in the local coordinate system in the depth direction.

[0044] 5) Conversion and correction of slant distance observations between seabed points.

[0045] The original relative observation information between seabed transponders is the propagation delay of the acoustic signal. Using the equal-gradient ray tracking algorithm, according to the relative position relationship between the transponders, the initial grazing angle is iteratively refined by the observed time delay data, and then the corrected ray trajectory is obtained, and the inter-point slant range observation required for subsequent solution is obtained.

[0046] 2. Use the acquired data to construct an overall positioning model that includes the sea surface observation equation, the absolute observation equation of the sea surface-seabed points, the relative observation equation between seabed points and the depth constraint equation.

[0047] The sea surface virtual reference is the spatial position of the sea surface collection or receiving point at the time of measurement. The spatial position at this moment can be used as an observation alone, can form a spatial baseline with a shore-based reference station, or can form a virtual baseline with the spatial position at another moment. The overall positioning model constructed by the present invention includes sea surface observation equations, sea surface-seabed point absolute observation equations, relative observation equations between seabed points, and depth constraint equations. The construction of each equation is explained below.

[0048] The sea surface GNSS observation equation adopted in the present invention is a general form, which can represent the precise point positioning (PPP) mode of the sea surface platform that is not limited by the operating distance, and can also represent the differential positioning mode of establishing a reference station along the coast of the nearshore waters and performing GNSS observation synchronously with the survey ship.

[0049]

[0050] Among them, the subscript sat * 、r * They represent the observation satellites and shipborne GNSS antennas respectively; P and Φ represent the pseudorange and carrier phase observations respectively, which are expressed as ionosphere-free combination in non-difference mode and as double difference in differential mode; sat * represents the measured distance between the transducer and the transponder calculated from the acoustic delay measurement value; f(*) represents the distance vector between two coordinates; Represents the receiver clock difference vector, and the corresponding element is 0 in differential mode; T ZWD Represents the delayed wet component of the troposphere, expressed in double difference form in differential mode; ε P , εΦ is the observation noise and other unmodeled error vector; Respectively represent the integrated ambiguity parameter vectors in each solution mode.

[0051] The GNSS antenna center on the sea surface platform actually receives the satellite signal, so determining the position of the sea surface platform is essentially to obtain the instantaneous position of the GNSS antenna center on the sea surface platform. However, it is not the GNSS antenna on the sea surface platform that sends and receives signals with the target point underwater, but the transducer usually installed at the bottom of the platform, so a reference conversion is also required, that is, to complete the conversion of the sea surface reference position from the GNSS antenna center to the transducer array center.

[0052]

[0053] Among them, X td =(x,y,z) td is the transducer coordinate, is the coordinate of each transponder; R represents the coordinate axis rotation matrix including the attitude angle, l is the relative vector from the GNSS antenna center to the transceiver center of the bottom transducer in the hull coordinate system, this vector can also be called the arm vector, that is, the projection of the position of the antenna phase center relative to the transceiver center in the hull coordinate system. In some hull structural designs, the two types of sensors are connected by a fixed connecting rod that does not deform, and can be calibrated separately before operation.

[0054] By transmitting signals between the instantaneous positions of the sea surface platform benchmark and the underwater target positions, multiple distance quantities are obtained to achieve spatial intersection, and then the coordinate position of the underwater target is solved. This is the sea surface-seabed acoustic absolute observation mode, and the constructed sea surface-seabed point absolute observation equation is:

[0055]

[0056] Where ρ represents the measured distance from the transducer to the transponder calculated from the acoustic delay measurement value; f(*) represents the distance vector between two coordinates; X td =(x,y,z) td is the transducer coordinate, are the coordinates of each transponder; Represents the error vector of the ranging system with varying sound speed; represents other acoustic positioning random error vectors.

[0057] The original observation data obtained by beacons transmitting and receiving acoustic signals between seabed transponders is the signal propagation delay. Taking into account the propagation trajectory of the sound line, it can be converted into the side length observation data between each point. The relative observation equation between seabed points is constructed as follows:

[0058]

[0059] Among them, SD * represents the mutual distance observation vector between each transponder; f(*) represents the distance vector between two coordinates; are the coordinates of each transponder; Represents the mutual measurement distance noise vector.

[0060] Furthermore, in order to improve the vertical solution accuracy, it is necessary to use the water level gauge pressure depth value with relatively higher observation accuracy to constrain the observation model. The depth constraint equation constructed in the present invention is:

[0061]

[0062] in, is the measured depth difference between each transducer and each transponder on the seabed; R′ represents the spatial reference conversion matrix. By unifying the observation information at the spatiotemporal level and integrating the above observation equations, the GNSS-A overall positioning model proposed in this invention can be obtained:

[0063]

[0064] The above model is linearized and rewritten into the form of error equation:

[0065]

[0066] Where V = [V P V Φ V ρ V SD ] T Represents the combined residual term, which is composed of the residual vectors corresponding to the pseudorange observation equation, the carrier phase observation equation, the sea surface-seafloor absolute observation equation, and the seafloor mutual ranging equation; A = [A P A Φ A ρ A SD ] T , l=[l P l Φ l ρ l SD ] T represent the combined coefficient matrix and free terms respectively.

[0067] 3. Filter and solve the established overall positioning model to obtain the position of each seabed transponder to achieve the positioning of the seabed control point group.

[0068] Considering that the integrated observation model has many parameters to be estimated and is updated frequently, the model can be pre-estimated based on Kalman filtering. This embodiment uses the Helmert variance component estimation method to adjust the weight ratio relationship between various observations, and uses the robust adaptive weighted filtering algorithm to perform iterative update calculations to achieve the positioning of the seabed control point group.

[0069] The acquisition of the position of the sea surface platform needs to give play to the advantages of multi-system observation data fusion positioning, but its pre-test random model is often difficult to adapt to the changes in the observed geometric structure in real time; in underwater acoustic signal ranging, the change mechanism of the sound speed in the sea surface-seabed absolute observation information and the seabed relative observation information is mainly due to the change of temperature and static pressure. Therefore, the weight ratio relationship between heterogeneous observations is difficult to give in advance. After the above process is unified into the joint observation model, whether the established random model can accurately reflect the relative weight relationship between different types of observations and balance the contribution of observations of different precision to the filtering results will directly affect the final solution result. In response to this problem, the present invention introduces the Helmert variance component estimation method to adjust the random model.

[0070] According to the empirical model, the variance and covariance of the same observation can be set so that the k-time filter prediction vector δX k As a pseudo-observation, the above error equation can be expressed as:

[0071]

[0072] Under the assumption that the observations are independent of each other, the initial weight matrix P k The covariance matrix Σ of each observation can be k and the inaccurate unit-weight variance component calculate:

[0073]

[0074] According to the theoretical derivation results, the relationship between the variance components of different observation unit weights in this model can be obtained:

[0075]

[0076] Where V i and P i are the residual vector and weight matrix corresponding to the i-th observation respectively; m i Indicates the dimension of the observation vector; tr(*) indicates the trace operation. The above formula can be written for each observation, and then the Helmert rigorous calculation formula is obtained, which is expressed in matrix form after merging:

[0077]

[0078] In the formula, is the coefficient matrix of the combined equation, is the free term of the equation. Solving the above equation can obtain the variance component estimation vector of the joint observation model in this iteration at time k.

[0079] The abnormal pollution brought to the original observation data by the complex measurement environment of the ocean is not only reflected in the construction of the above-mentioned random model. If the conventional Kalman filter is directly used to estimate the parameters of the joint positioning model, the abnormal information will greatly affect the model estimation results due to the poor anti-interference ability of the algorithm itself; on the other hand, the dynamic model anomaly caused by the carrier disturbance needs to be considered during data processing. This phenomenon mainly occurs because the movement process of the sea surface platform changes frequently, and the preset dynamic model is often fixed, which is inconsistent with the real movement state of the carrier in real-time change. The obvious difference between the two will bring a large error to the result. The conventional Kalman filter algorithm cannot perform good self-adjustment, resulting in filter divergence or parameter estimation distortion.

[0080] Based on the above analysis, the present invention introduces a reasonable adaptive adjustment strategy in the parameter estimation stage. At the same time, combined with the robust estimation principle, the robust adaptive weighted Kalman filter estimation method is used to identify abnormal model interference in real time and make corresponding adjustments to obtain parameter estimates after weakening abnormal observation information.

[0081] Under the robust estimation criterion, the concept of equivalent weight is introduced, with the equivalent weight element Instead of the corresponding elements of the original weight matrix, the objective function should be written as:

[0082]

[0083] Based on this objective function, the robust Kalman filter solution of the joint observation model can be obtained by filtering and solving:

[0084]

[0085] By designing an adaptive weight selection strategy and integrating it into the parameter estimation of the joint observation model, the impact of abnormal dynamic information can be effectively reduced. The basic idea is to use the close connection between the forecast parameter vector and the dynamic model error to construct a reasonable discriminant statistic and evaluation standard, based on which the model error reflected by a certain parameter in the current epoch is given, and then the adaptive factor corresponding to the abnormal parameter is constructed in combination with the corresponding adaptive factor function. All adaptive factors constitute the adaptive factor matrix The weight matrix corresponding to the original forecast parameter vector can be adjusted in real time to weaken the model error and obtain a more accurate estimation result for the current epoch:

[0086]

[0087] Accordingly, the filtering solution based on the adaptive weight selection strategy is:

[0088]

[0089] The key links mentioned above include how to judge the anomaly of the forecast state parameters and how to construct the adaptive factor matrix. The difference between the forecast information and the actual motion state can be reflected by a reasonable discriminant statistic. In the GNSS-A joint observation model, the prediction residual statistic or the variance component ratio statistic can be used. The construction method of the adaptive factors of each parameter is slightly different. The adaptive factor corresponding to the receiver clock error parameter with poor stability should be set to 0; the adaptive factor corresponding to the relatively stable and reliable tropospheric delay wet component and ambiguity parameter can be set to 1; and for the two coordinates or the increased sound speed error parameters, the anomaly judgment and adaptive weight selection should be performed epoch by epoch.

[0090] Considering the combined effects of dynamic model anomalies and observation errors, the above two filtering solutions can be obtained:

[0091]

[0092] The above formula is the parameter estimation solution and covariance matrix form of the GNSS-A joint positioning model using the robust adaptive weighted Kalman filter. This filter solution is a better result obtained after the influence of model anomalies and observation gross error interference are controlled to a certain extent. If iterative solution is required, the parameter estimation vector obtained this time is re-substituted into the error equation to obtain the residual vector of this solution, and then the residual vector is used to re-estimate the variance component, update the variance component estimation, and once again use the robust adaptive weighted filter to update the parameter estimation vector to achieve iterative calculation.

[0093] In the above iterative solution process, it is necessary to judge the posterior information of various observations. If the set loop termination condition is not met, a new round of iterative solution is started. If the termination requirement is met, the parameter estimation process is completed to achieve the positioning of the seabed control point group. The loop termination condition is: after each iterative update, the ratio of the variance component estimation vectors corresponding to various observations is less than the corresponding set threshold.

[0094] The positioning method adopted by the present invention can more effectively control the model error and balance the contribution of various observation quantities to the positioning result.

[0095] Embodiment of a networked joint positioning device for a local seafloor control point group

[0096] The present invention also provides a networking joint positioning device for a local seabed control point group, including a processor, the processor is used to execute a computer program to implement the networking joint positioning method based on the local seabed control point group of the present invention. The specific implementation process of the method has been described in detail in the embodiment of the networking joint positioning method based on the local seabed control point group, and will not be repeated here.

[0097] The present invention fully considers the relative ranging mode between seabed acoustic transponders, and uses its relative ranging data to establish a relative observation equation group between seabed points, and uses the relative observation equation group as part of the overall positioning model, so that the relative observation data between seabed points can be fully utilized in the solution process, thereby improving the efficiency and accuracy of positioning. In the positioning solution process, the idea of ​​variance component estimation is used to rebalance the weight ratio relationship between various types of observations, and the random model of joint positioning is refined. In view of the many error effects brought to the joint model by the dynamics and complexity of the ocean measurement environment, a reasonable adaptive weight selection and abnormal information de-weighting strategy is introduced. By constructing an anti-error adaptive weight selection filter solution for the joint observation model, the influence of abnormal dynamic information and observation information on the solution results is effectively weakened, thereby ensuring the accuracy and reliability of the integrated processing results of the seabed control network.

Claims

1. A networked joint positioning method for a local seafloor control point group, It is characterized in that The positioning method comprises the following steps: 1) Obtain GNSS observation data, ship attitude observation data, water depth measurement data, sound velocity profile data, underwater acoustic absolute and relative observation data; 2) Using the acquired data, an overall positioning model is constructed, which includes the sea surface observation equation, the sea surface-sea bottom point absolute observation equation, the relative observation equation between sea bottom points and the depth constraint equation. The relative observation equation between sea bottom points is established based on the distance observation quantity converted from the relative time delay data between the transponders on the sea bottom. 3) Filter and solve the established overall positioning model to obtain the position of each seabed transponder to achieve the positioning of the seabed control point group.

2. The networked joint positioning method of the local seabed control point group according to claim 1, It is characterized in that The relative observation equation between the seabed points is: Among them SD * represents the mutual distance observation vector between each transponder; f(*) represents the distance vector between two coordinates; are the coordinates of each transponder; Represents the mutual measurement distance noise vector.

3. The networked joint positioning method of the local seabed control point group according to claim 1, It is characterized in that The depth constraint equation is: in, is the measured depth difference between each transducer and each transponder on the seabed; R′ represents the spatial reference conversion matrix.

4. The method for joint positioning of a local seafloor control point group according to claim 1, It is characterized in that In the step 3), the Helmert variance component estimation method is used to adjust the random model of the overall positioning model during the filtering solution, and the variance component estimation between heterogeneous observations of different accuracies is iteratively updated to make the weight ratio relationship between the observations gradually reasonable.

5. The method for joint positioning of a local seabed control point group according to claim 1, It is characterized in that In the step 3), a robust adaptive weighted Kalman filtering algorithm is used for filtering solution.

6. The method for joint positioning of a local seabed control point group according to claim 5, It is characterized in that The adaptive weight selection strategy in the robust adaptive weighted Kalman filtering algorithm is: The given prediction residual statistics or variance component ratio statistics are used to calculate the judgment evaluation standard, and the adaptive factors are constructed for each parameter by category, and then an adaptive factor matrix is ​​formed to participate in the filtering solution in the form of an adaptive factor matrix.

7. The method for joint positioning of a local seabed control point group according to claim 6, It is characterized in that The filtering solution result is: in A is the prediction vector filtered at time k; k is the coefficient matrix, W k is the adaptive factor matrix; is the equivalent weight matrix, is the weight matrix corresponding to the prediction parameter vector; is the new information vector, and I is the unit matrix.

8. The method for joint positioning of a local seabed control point group according to claim 1, It is characterized in that The data used in step 2) are the data preprocessed in step 1), and the preprocessing includes gross error elimination and time alignment.

9. The method for joint positioning of a local seabed control point group according to claim 8, It is characterized in that The preprocessing also includes determining the initial grazing angle from the original underwater acoustic relative observation time delay data based on a strict equal-gradient sound line tracking model, and then obtaining the corrected observation slant distance between the seabed points.

10. A networked joint positioning device for a local seabed control point group, comprising a processor, It is characterized in that The processor is used to execute a computer program to implement the networking joint positioning method of a local seabed control point group according to any one of claims 1 to 9.

Citation Information

Patent Citations

  • Method and device for positioning seabed acoustic transponder

    CN111208520A