GNSS positioning method and device based on high-precision map

Abstract

The application provides a GNSS positioning method and device based on a high-precision map, the method comprising: obtaining a target distance from a vehicle to a shielding element based on pre-acquired real-time positioning information of the vehicle and in combination with the high-precision map; performing error prediction according to the target distance to obtain a predicted error; adjusting GNSS covariance of a filter according to the predicted error to obtain fused covariance; and performing positioning by using the filter according to the fused covariance to obtain a positioning result. The GNSS covariance of the filter is adjusted by using the predicted error, and then positioning is performed according to the fused covariance obtained after the adjustment, so as to reduce the influence of position drift or jumping caused by GNSS calculation errors due to large-area shielding on overall positioning of the vehicle, and improve the positioning accuracy of the vehicle.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a GNSS positioning method and apparatus based on high-precision maps. Background Technology

[0002] Autonomous driving positioning systems rely on multi-sensor fusion to obtain high-precision positioning output. The filtering techniques used require evaluating the measurement errors of each sensor, typically represented by covariance and comatrix, to achieve positioning. Sensors include Global Navigation Satellite System (GNSS), Inertial Measurement Unit (IMU), wheel speed sensors, lidar, and cameras. Among these, GNSS provides global positioning information and is a crucial component for vehicle multi-sensor fusion positioning.

[0003] GNSS offers high accuracy in open terrain. However, in scenarios with significant obstruction, such as toll booths, gantries, and tunnels, the GNSS module may fail to accurately assess the calculation error (covariance) if the calculation is not timely, leading to errors and drift in the output positioning results. Furthermore, if the actual positioning error is large, the covariance provided by the GNSS module may be smaller than the actual value. This can cause the positioning system to over-rely on the GNSS information, resulting in positioning deviations for autonomous vehicles. Summary of the Invention

[0004] This invention provides a GNSS positioning method and device based on high-precision maps to solve the defects in the prior art caused by large-area occlusion scenes, which leads to positioning deviations, improves the calculation accuracy, avoids positioning result drift, and improves positioning accuracy.

[0005] This invention provides a GNSS positioning method based on a high-precision map, comprising: obtaining the target distance from the vehicle to an occluding element based on pre-acquired real-time vehicle positioning information and combined with a high-precision map; performing error prediction based on the target distance to obtain a prediction error; adjusting the GNSS covariance of the Global Navigation Satellite System pre-calculated based on GNSS equipment based on the prediction error to obtain a fused covariance; and performing positioning based on the fused covariance and using a filter to obtain a positioning result.

[0006] According to a GNSS positioning method based on a high-precision map provided by the present invention, the step of adjusting the GNSS covariance pre-calculated based on the GNSS device according to the prediction error to obtain a fused covariance includes: adjusting the GNSS covariance pre-calculated based on the GNSS device according to a preset weighting system range of the GNSS error covariance; and obtaining the fused covariance based on the adjusted GNSS covariance, combined with the preset weighting system range of the GNSS error covariance, a preset distance covariance conversion coefficient, and the prediction error.

[0007] According to a GNSS positioning method based on a high-precision map provided by the present invention, the step of performing error prediction based on the target distance to obtain the prediction error includes: inputting the target distance into an error prediction model to obtain the prediction error output by the error prediction model; wherein, the error prediction model is trained based on pre-acquired GNSS position training data, the real position corresponding to the GNSS position training data, and the training distance data from the real position to the nearest occluded element.

[0008] According to a GNSS positioning method based on a high-precision map provided by the present invention, training the error prediction model includes: acquiring GNSS position training data and the real positions corresponding to the GNSS position training data; obtaining GNSS training error based on the GNSS position training data and the real positions corresponding to the GNSS position training data; obtaining training distance data from the vehicle to the nearest occluded element based on the real positions corresponding to the GNSS position training data; using the training distance data as input data for the model to be trained, and using the GNSS training error as the label of the model to be trained, training the model to obtain the error prediction model.

[0009] According to the present invention, a GNSS positioning method based on a high-precision map is provided, wherein obtaining GNSS position training data and the real position corresponding to the GNSS position training data includes: obtaining GNSS position training data; and performing calculations based on the GNSS position data using graph optimization or RTS smoothing to obtain the real data corresponding to the GNSS position data.

[0010] According to the present invention, a GNSS positioning method based on a high-precision map is provided. The step of obtaining the target distance from the vehicle to the occluding element based on the pre-acquired real-time vehicle positioning information and combined with the high-precision map includes: querying the high-precision map based on the pre-acquired real-time vehicle positioning information to determine the occluding element that is currently closest to the vehicle; and obtaining the target distance based on the real-time vehicle positioning information and the occluding element that is currently closest to the vehicle.

[0011] According to a GNSS positioning method based on a high-precision map provided by the present invention, the step of obtaining the target distance based on the real-time positioning information of the vehicle and the nearest occluding element currently close to the vehicle includes: obtaining the kinematic center point of the vehicle based on the real-time positioning information of the vehicle; and obtaining the perpendicular distance between the kinematic center point of the vehicle and the nearest edge of the nearest occluding element currently close to the vehicle as the target distance.

[0012] The present invention also provides a GNSS positioning device based on a high-precision map, comprising: a target distance acquisition module, which obtains the target distance from the vehicle to the occluding element based on pre-acquired real-time vehicle positioning information and combined with a high-precision map; an error prediction module, which performs error prediction based on the target distance to obtain a prediction error; a fusion module, which adjusts the GNSS covariance of the Global Navigation Satellite System pre-calculated based on the GNSS device according to the prediction error to obtain a fused covariance; and a positioning module, which performs positioning based on the fused covariance and uses a filter to obtain a positioning result.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the GNSS positioning method based on high-precision maps as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the GNSS positioning method based on high-precision maps as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described GNSS positioning methods based on high-precision maps.

[0016] The GNSS positioning method and apparatus based on high-precision maps provided by this invention obtains the target distance from the vehicle to the occluded element through real-time vehicle positioning information. Based on the target distance from the vehicle to the occluded element, error prediction is performed to obtain the prediction error. The prediction error is then used to adjust the GNSS covariance of the filter, which facilitates positioning based on the adjusted fusion covariance. This reduces the impact of position drift or fluctuation caused by GNSS calculation errors due to large-area occlusion on the overall vehicle positioning and improves the vehicle positioning accuracy. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is one of the flowcharts illustrating the GNSS positioning method based on high-precision maps provided by this invention;

[0019] Figure 2 This is the second flowchart of the GNSS positioning method based on high-precision maps provided by the present invention;

[0020] Figure 3 This is a schematic diagram of obtaining training distance data provided by the present invention;

[0021] Figure 4 This is a schematic diagram of the structure of the GNSS positioning device based on high-precision maps provided by the present invention;

[0022] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0024] Figure 1 This invention illustrates a flowchart of a GNSS positioning method based on a high-precision map, comprising:

[0025] S11, based on the pre-acquired real-time vehicle location information and combined with a high-precision map, obtains the target distance from the vehicle to the occluded element;

[0026] S12, Based on the target distance, perform error prediction to obtain the prediction error;

[0027] S13. Based on the prediction error, the GNSS covariance of the filter is adjusted to obtain the fused covariance.

[0028] S14. Based on the fused covariance and using a filter for localization, the localization result is obtained.

[0029] It should be noted that S1N in this manual does not represent the order of GNSS positioning methods based on high-precision maps. The following will explain this in detail. Figures 2-3 The present invention describes a GNSS positioning method based on high-precision maps.

[0030] Step S11: Based on the pre-acquired real-time vehicle positioning information and combined with the high-precision map, the target distance from the vehicle to the occluded element is obtained.

[0031] In this embodiment, the target distance from the vehicle to the occluding element is obtained based on pre-acquired real-time vehicle positioning information and combined with a high-precision map. This includes: querying the high-precision map based on the pre-acquired real-time vehicle positioning information to determine the nearest occluding element to the vehicle; and obtaining the target distance based on the real-time vehicle positioning information and the nearest occluding element. It should be noted that the real-time vehicle positioning information can be determined based on the real-time Global Navigation Satellite System (GNSS) position obtained by the GNSS module.

[0032] Furthermore, based on the vehicle's real-time location information and the nearest occluding element to the vehicle, the target distance is obtained, including: obtaining the vehicle's kinematic center point based on the vehicle's real-time location information; and obtaining the perpendicular distance between the vehicle's kinematic center point and the nearest edge of the nearest occluding element to the vehicle, as the target distance.

[0033] It should be noted that, due to compliance requirements, occluded elements are typically represented as a planar polygon on the map, and the vehicle's position is abstracted as the vehicle's kinematic center, which can be represented by a point, such as the center point of the vehicle's rear wheel. Therefore, the target distance is equivalent to the perpendicular distance from the point to the nearest edge of the polygon. Additionally, occluded elements include at least one of toll booths, gantries, tunnels, and overlapping roads (elevated roads). The map can be a high-precision map with higher requirements for detail or richness than a standard SD map, such as ADAS maps used in driver assistance systems.

[0034] Step S12: Based on the target distance, perform error prediction to obtain the prediction error.

[0035] In this embodiment, reference Figure 2 Based on the target distance, error prediction is performed to obtain the prediction error, including: inputting the target distance into the error prediction model to obtain the prediction error output by the error prediction model; wherein, the error prediction model is trained based on pre-acquired GNSS position training data, the real position corresponding to the GNSS position training data, and the training distance data from the real position to the nearest occluded element.

[0036] It should be noted that the prediction error is expressed as:

[0037] e GNSS =c0+c1×d OCCLUSION

[0038] Among them, e GNSS c represents the prediction error; c0 and c1 represent the fitting coefficients; d OCCLUSION This indicates the target distance to the nearest occluding element to the vehicle.

[0039] In an alternative embodiment, continue to refer to Figure 2 Before inputting the target distance into the error prediction model, the process includes: training the error prediction model. Specifically, training the error prediction model includes:

[0040] S121, Obtain GNSS position training data and the corresponding real positions of GNSS position training data.

[0041] It should be noted that obtaining GNSS position training data and the corresponding real positions includes: obtaining GNSS position training data; and using graph optimization or RTS smoothing to solve for the GNSS position data to obtain the corresponding real positions.

[0042] S122, GNSS training error is obtained based on GNSS position training data and the actual positions corresponding to the GNSS position training data.

[0043] S123: Based on the actual location corresponding to the GNSS position training data, obtain the training distance data from the vehicle to the nearest occluded element.

[0044] To go further, refer to Figure 3 Based on the real-world locations corresponding to the GNSS location training data, training distance data from the vehicle to the nearest occluding element is obtained. This includes: finding occluding elements in the corresponding map based on the real-world locations corresponding to the GNSS location training data, and selecting the occluding element closest to the vehicle; and obtaining the training distance data based on the selected closest occluding element and its real-world location. It should be noted that the training distance data can refer to the target distance described above, and will not be repeated here.

[0045] S124: Use the training distance data as the input data of the model to be trained, and use the GNSS training error as the label of the model to be trained to train the model and obtain the error prediction model.

[0046] It should be noted that the model to be trained can be an existing network built into the training device, which typically includes a network structure, or it can be another network specified by the user. In this embodiment, training the model to be trained includes: inputting training distance data into the model to be trained to obtain the predicted training error output by the model; constructing a loss function based on the predicted training error and the GNSS training error; and terminating the training based on the convergence of the loss function. Furthermore, the training error prediction model can be an offline process and does not affect the subsequent online process of using the trained error prediction model to predict errors based on the target distance. It should be noted that the loss function can be constructed based on the least squares method, or other commonly used deep learning loss functions can be used, such as mean squared error loss, mean absolute error loss, and cross-entropy loss, etc., without further limitation here.

[0047] Step S13: Adjust the GNSS covariance calculated in advance based on the GNSS equipment according to the prediction error to obtain the fused covariance.

[0048] In this embodiment, the GNSS covariance calculated in advance based on the GNSS device is adjusted according to the prediction error to obtain the fused covariance. This includes: adjusting the GNSS covariance calculated in advance based on the GNSS device according to the weighting range of the preset GNSS error covariance; and obtaining the fused covariance based on the adjusted GNSS covariance, combined with the weighting range of the preset GNSS error covariance, the preset distance covariance conversion coefficient, and the prediction error.

[0049] It should be added that the fused covariance is expressed as:

[0050]

[0051] Among them, Cov fusion The fusion covariance is represented by α; α represents the weighting range of the preset GNSS error covariance, typically (0,1); β represents the preset distance covariance transformation coefficient; Cov gnss This represents the GNSS covariance calculated in advance based on GNSS equipment; e GNSS This indicates the prediction error.

[0052] It should be noted that the larger α is, the more confident one is. The smaller α is, the more reliable the GNSS covariance calculated based on GNSS equipment is.

[0053] Step S14: Based on the fusion covariance, the localization result is obtained using a filter. It should be noted that after obtaining the fusion covariance, it is input into the filter for localization to obtain a more robust localization result and improve localization accuracy.

[0054] In one optional embodiment, localization is performed based on the fusion covariance and using a filter, including: updating the filter using the fusion covariance; and using the updated filter, combined with the localization result from the previous time step, to predict the current position and obtain the localization result. It should be noted that the specific method of using the filter for localization can be determined based on the actual filter selected; the filter can be a Kalman filter, etc., and no further restrictions are imposed here.

[0055] In summary, this embodiment of the invention obtains the target distance from the vehicle to the occluded element using real-time vehicle positioning information. Based on this target distance, error prediction is performed to obtain the prediction error. This prediction error is then used to adjust the GNSS covariance of the filter, facilitating positioning based on the adjusted fused covariance. This reduces the impact of position drift or fluctuation caused by GNSS calculation errors due to large-area occlusion on the overall vehicle positioning, thereby improving the vehicle's positioning accuracy.

[0056] The GNSS positioning device based on high-precision maps provided by the present invention will be described below. The GNSS positioning device based on high-precision maps described below can be referred to in correspondence with the GNSS positioning method based on high-precision maps described above.

[0057] Figure 4 A schematic diagram of a GNSS positioning device based on a high-precision map is shown. The device includes:

[0058] The target distance acquisition module 41 obtains the target distance from the vehicle to the occluded element based on the pre-acquired real-time vehicle positioning information and combined with the high-precision map.

[0059] Error prediction module 42 performs error prediction based on the target distance to obtain the prediction error;

[0060] The fusion module 43 adjusts the GNSS covariance of the Global Navigation Satellite System (GNSS) pre-calculated based on GNSS equipment according to the prediction error to obtain the fused covariance.

[0061] The positioning module 44 performs positioning based on the fused covariance and uses a filter to obtain the positioning result.

[0062] In this embodiment, the target distance acquisition module 41 includes: an occlusion element determination unit, which queries a high-precision map based on pre-acquired real-time vehicle positioning information to determine the occlusion element that is currently closest to the vehicle; and a target distance acquisition unit, which obtains the target distance based on the real-time vehicle positioning information and the occlusion element that is currently closest to the vehicle.

[0063] Furthermore, the target distance acquisition unit includes: a center point acquisition subunit, which obtains the vehicle's kinematic center point based on the vehicle's real-time positioning information; and a target distance acquisition subunit, which obtains the perpendicular distance between the vehicle's kinematic center point and the nearest edge of the nearest occluding element currently closest to the vehicle, as the target distance.

[0064] The error prediction module 42 includes: a data input unit for inputting the target distance into the error prediction model; an error prediction unit for using the error prediction model to predict the error of the input target distance and obtain the prediction error; and a data input unit for outputting the prediction error obtained by the error prediction model. The error prediction model is trained based on pre-acquired GNSS position training data, the real position corresponding to the GNSS position training data, and the training distance data from the real position to the nearest occluded element.

[0065] In an optional embodiment, the device further includes: a training module for training the error prediction model before inputting the target distance into the error prediction model. Specifically, the training module includes: a training data acquisition unit for acquiring GNSS position training data and the corresponding real positions; a training error acquisition unit for obtaining GNSS training error based on the GNSS position training data and the corresponding real positions; a training distance acquisition unit for obtaining training distance data from the vehicle to the nearest occluded element based on the corresponding real positions; and a training unit for using the training distance data as input data to the model to be trained and the GNSS training error as a label for the model to be trained, thereby training the model to obtain the error prediction model.

[0066] Furthermore, the training data acquisition unit includes: a data acquisition subunit, which acquires GNSS location training data; and a solution subunit, which performs solution calculations based on the GNSS location data using graph optimization or RTS smoothing to obtain the actual data corresponding to the GNSS location data.

[0067] The training distance acquisition unit includes: an element finding subunit, which searches for occluding elements in the corresponding map based on the real location corresponding to the GNSS position training data, and selects the occluding element closest to the vehicle; and a training data acquisition subunit, which obtains training distance data based on the selected occluding element closest to the vehicle and the real location.

[0068] In addition, the training unit includes: an error acquisition subunit, which inputs training distance data into the model to be trained to obtain the predicted training error output by the model; a loss function construction subunit, which constructs a loss function based on the predicted training error and the GNSS training error; and a training subunit, which terminates training based on the convergence of the loss function.

[0069] The fusion module 43 includes: an adjustment unit, which adjusts the GNSS covariance calculated in advance based on the GNSS equipment according to the preset weight system range of the GNSS error covariance; and a fusion unit, which obtains the fused covariance based on the adjusted GNSS covariance and in combination with the preset weight system range of the GNSS error covariance, the preset distance covariance conversion coefficient, and the prediction error.

[0070] In an optional embodiment, the positioning module 44 includes: an update unit that updates the filter using fusion covariance; and a positioning unit that uses the updated filter and combines it with the positioning result of the previous time step to predict the current position and obtain a positioning result.

[0071] In summary, this embodiment of the invention uses a target distance acquisition module to obtain the target distance from the vehicle to the occluded element based on the vehicle's real-time positioning information. An error prediction module then uses this target distance to predict the error in real time, obtaining a prediction error. The fusion module then uses this prediction error to adjust the GNSS covariance of the filter, facilitating positioning by the positioning module based on the adjusted fusion covariance. This reduces the impact of position drift or fluctuations caused by GNSS calculation errors due to large-area occlusion on the overall vehicle positioning, thereby improving the vehicle's positioning accuracy.

[0072] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include a processor 51, a communication interface 52, a memory 53, and a communication bus 54. The processor 51, communication interface 52, and memory 53 communicate with each other via the communication bus 54. The processor 51 can call logical instructions in the memory 53 to execute a GNSS positioning method based on a high-precision map. This method includes: obtaining the target distance from the vehicle to occluded elements based on pre-acquired real-time vehicle positioning information and combined with a high-precision map; performing error prediction based on the target distance to obtain a prediction error; adjusting the GNSS covariance of the filter based on the prediction error to obtain a fused covariance; and performing positioning based on the fused covariance and using the filter to obtain the positioning result.

[0073] Furthermore, the logical instructions in the aforementioned memory 53 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0074] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the GNSS positioning method based on a high-precision map provided by the above methods. The method includes: obtaining the target distance from the vehicle to the occluding element based on pre-acquired real-time vehicle positioning information and combined with a high-precision map; performing error prediction based on the target distance to obtain a prediction error; adjusting the GNSS covariance of the filter based on the prediction error to obtain a fusion covariance; and performing positioning based on the fusion covariance and using the filter to obtain a positioning result.

[0075] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the GNSS positioning method based on a high-precision map provided by the above methods. The method includes: obtaining the target distance from the vehicle to the occluding element based on pre-acquired real-time vehicle positioning information and combined with a high-precision map; performing error prediction based on the target distance to obtain a prediction error; adjusting the GNSS covariance of the filter based on the prediction error to obtain a fusion covariance; and performing positioning based on the fusion covariance and using the filter to obtain a positioning result.

[0076] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0077] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A GNSS positioning method based on high-precision maps, characterized in that, include: Based on the pre-acquired real-time vehicle location information and combined with high-precision maps, the target distance from the vehicle to the occluded element is obtained. Based on the target distance, error prediction is performed to obtain the prediction error; Based on the prediction error, the GNSS covariance of the Global Navigation Satellite System, which was previously calculated based on GNSS equipment, is adjusted to obtain the fused covariance; Based on the fused covariance, and using a filter for localization, the localization result is obtained; The prediction error is expressed as: ； in, Indicates prediction error; and Represents the fitting coefficient; Indicates the target distance of the nearest occluding element to the vehicle; The fusion covariance is expressed as: ； in, Indicates the fusion covariance; This indicates the range of the weighting system for the preset GNSS error covariance, typically (0,1). Indicates the preset distance covariance transformation coefficient; This represents the GNSS covariance calculated in advance based on GNSS equipment; This indicates the prediction error.

2. The GNSS positioning method based on high-precision maps according to claim 1, characterized in that, The step of adjusting the GNSS covariance pre-calculated based on the GNSS equipment according to the prediction error to obtain the fused covariance includes: The GNSS covariance calculated in advance based on the GNSS equipment is adjusted according to the preset weight system range of the GNSS error covariance. Based on the adjusted GNSS covariance, and combined with the preset weighting range of the GNSS error covariance, the preset distance covariance conversion coefficient, and the prediction error, the fused covariance is obtained.

3. The GNSS positioning method based on high-precision maps according to claim 1, characterized in that, The step of performing error prediction based on the target distance to obtain the prediction error includes: The target distance is input into the error prediction model to obtain the prediction error output by the error prediction model; The error prediction model is trained based on pre-acquired GNSS location training data, the actual location corresponding to the GNSS location training data, and the training distance data from the actual location to the nearest occluded element.

4. The GNSS positioning method based on high-precision maps according to claim 3, characterized in that, Training the error prediction model includes: Obtain GNSS location training data and the actual locations corresponding to the GNSS location training data; The GNSS training error is obtained based on the GNSS position training data and the actual position corresponding to the GNSS position training data. Based on the actual location corresponding to the GNSS location training data, the training distance data from the vehicle to the nearest occluded element is obtained; The training distance data is used as the input data for the model to be trained, and the GNSS training error is used as the label for the model to be trained. The model is then trained to obtain an error prediction model.

5. The GNSS positioning method based on high-precision maps according to claim 4, characterized in that, The acquisition of GNSS position training data and the corresponding real positions includes: Acquire GNSS location training data; Based on the GNSS location training data, graph optimization or RTS smoothing is used to solve the problem and obtain the real data corresponding to the GNSS location training data.

6. The GNSS positioning method based on high-precision maps according to claim 1, characterized in that, The process of obtaining the target distance from the vehicle to the occluded element based on pre-acquired real-time vehicle location information and combined with a high-precision map includes: Based on the pre-acquired real-time vehicle location information, query the high-precision map to determine the nearest occluding element to the vehicle; The target distance is obtained based on the vehicle's real-time location information and the nearest occluding element to the vehicle.

7. The GNSS positioning method based on high-precision maps according to claim 6, characterized in that, The step of obtaining the target distance based on the vehicle's real-time positioning information and the nearest occlusion element to the vehicle includes: Based on the real-time vehicle positioning information, the vehicle's kinematic center point is obtained; Obtain the vertical distance between the vehicle's kinematic center point and the nearest edge of the occluding element that is currently closest to the vehicle, and use this distance as the target distance.

8. A GNSS positioning device based on a high-precision map, characterized in that, include: The target distance acquisition module obtains the target distance from the vehicle to the occluded element based on the pre-acquired real-time vehicle positioning information and combined with the high-precision map. The error prediction module performs error prediction based on the target distance to obtain the prediction error; The fusion module adjusts the GNSS covariance of the Global Navigation Satellite System (GNSS) pre-calculated based on GNSS equipment according to the prediction error to obtain the fused covariance. The localization module performs localization based on the fused covariance and uses a filter to obtain the localization result; The prediction error is expressed as: ； in, Indicates prediction error; and Represents the fitting coefficient; Indicates the target distance of the nearest occluding element to the vehicle; The fusion covariance is expressed as: ； in, Indicates the fusion covariance; This indicates the range of the weighting system for the preset GNSS error covariance, typically (0,1). Indicates the preset distance covariance transformation coefficient; This represents the GNSS covariance calculated in advance based on GNSS equipment; This indicates the prediction error.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the GNSS positioning method based on high-precision maps as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the GNSS positioning method based on high-precision maps as described in any one of claims 1 to 7.

Images