Element position confidence acquisition method and device of high-precision map and vehicle

Abstract

The present application relates to the technical field of automatic driving, and particularly provides a high-precision map element position confidence acquisition method and device, medium and vehicle, aiming to solve the problem of how to effectively measure the accuracy of the position of the elements in the high-precision map obtained by crowdsourcing mapping. To this end, the present application is based on the first type of map element data obtained by the vehicle-mounted sensor, and based on the first type of map element data, the second type of map element data is calculated and obtained, based on the position data of the first type of map element data and the position data of the preset reference target, the distribution parameters of the first offset sample are obtained, and based on the distribution parameters of the first offset sample, the distribution parameters of the second offset sample are obtained, so as to obtain the position confidence of the first type of map element data and the second type of map element data based on the distribution parameters of the first offset sample and the distribution parameters of the second offset sample, which can make the position confidence more comprehensive and accurate.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, specifically providing a method, apparatus, medium, and vehicle for obtaining the confidence level of feature locations in a high-precision map. Background Technology

[0002] In the field of autonomous driving, high-precision maps, as service providers of prior environmental information, play a crucial role in high-precision positioning, assisted environmental perception, planning, and decision-making processes. Based on the requirements of autonomous driving systems for high-precision maps, and facing the rapidly changing real world, high-precision maps containing rich elements must provide autonomous driving systems with high-freshness, high-precision, and multi-dimensional road and additional information.

[0003] High-precision maps present new challenges to map service providers compared to traditional electronic maps in terms of production processes, feature quality, update cycles, attribute richness, and production costs. To reduce the cost of high-precision map production and increase update frequency, the use of crowdsourced vehicle-mounted sensors to replace traditional high-precision data collection vehicles has gradually gained traction among several manufacturers. Generally, this involves using the massive amounts of vectorized map feature data collected and analyzed from the sensing devices of multiple vehicles along the same road to improve and update existing maps or create new ones.

[0004] However, with the continuous expansion of mapping scope and the increasing number of connected vehicles, crowdsourced data is also constantly increasing. Furthermore, different vehicles and batches of onboard sensors may exhibit random errors in observing the location and attributes of individual map features. Additionally, in some cases, due to poor observation conditions or the vehicle's own positioning accuracy, systematic errors may also exist. How to effectively measure the accuracy of the geometric position of features in high-precision maps obtained through crowdsourced mapping is a problem that needs to be solved.

[0005] Accordingly, there is a need in this field for a new scheme to obtain the confidence level of map feature locations in order to solve the above problems. Summary of the Invention

[0006] To overcome the above-mentioned shortcomings, this invention is proposed to provide a solution, or at least a partial solution, to the problem of how to effectively measure the accuracy of the location of features in high-precision maps obtained through crowdsourced mapping.

[0007] In a first aspect, the present invention provides a method for obtaining the confidence level of feature locations in a high-precision map, the method comprising:

[0008] Acquire first-type map feature data, wherein the first-type map feature data is obtained by sensing through vehicle-mounted sensors;

[0009] Acquire second type of map feature data, wherein the second type of map feature data is calculated based on the first type of map feature data;

[0010] Based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target, the distribution parameters of the first offset sample of the first type of map feature data are obtained;

[0011] Based on the distribution parameters of the first offset sample, obtain the distribution parameters of the second offset sample of the second type of map feature data;

[0012] For the first type of map feature data and the second type of map feature data, the corresponding probability models are used according to their respective sample distribution parameters to obtain the position confidence level of the data falling into their respective preset confidence intervals, wherein the preset confidence intervals are related to the preset position accuracy.

[0013] In one technical solution of the above-mentioned method for obtaining the confidence level of the location of elements in a high-precision map, the second type of map element data is linear combination element data obtained by linearly combining multiple first type map element data that meet the combination conditions according to a linear relationship;

[0014] The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes:

[0015] Based on the linear relationship, the distribution parameters of the first offset samples of the multiple first-class map feature data are linearly combined to obtain the distribution parameters of the second offset samples of the second-class map feature data.

[0016] In one technical solution of the above-mentioned method for obtaining the confidence level of feature location in a high-precision map, the plurality of first-type map feature data includes first lane boundary line data and second lane boundary line data, and the second-type map feature data is lane centerline data;

[0017] The step of linearly combining the distribution parameters of the first offset samples of the plurality of first-type map feature data according to the linear relationship to obtain the distribution parameters of the second offset samples of the second-type map feature data includes:

[0018] The sample mean μ2 of the second offset sample of the lane centerline data is obtained: μ2 =

[0019] The sample standard deviation σ² of the second offset sample of the lane centerline data is:

[0020] Wherein, μ1 is the sample mean of the first offset sample of the first lane boundary line data, μ1' is the sample mean of the first offset sample of the second lane boundary line data, σ1 is the sample standard deviation of the first offset sample of the first lane boundary line data, and σ1' is the sample standard deviation of the first offset sample of the second lane boundary line data.

[0021] In one technical solution of the above-mentioned method for obtaining the confidence level of the location of elements in a high-precision map, the second type of map element data is the fused element data obtained by fusing multiple first type of map element data that meet the fusion conditions from the previously obtained first type of map element data and the currently obtained first type of map element data.

[0022] The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes:

[0023] Based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data, the distribution parameters of the second offset sample of the fused feature data are obtained.

[0024] In one technical solution of the above-mentioned method for obtaining the confidence level of feature location in a high-precision map, the plurality of first-class map feature data includes previously obtained third lane boundary line data and currently obtained fourth lane boundary line data, and the fused feature data is fused lane boundary line data;

[0025] The step of obtaining the distribution parameters of the second offset sample of the fused feature data based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data includes:

[0026] The sample mean μ3 of the second offset sample of the fused lane boundary line data is obtained:

[0027]

[0028] The sample standard deviation σ3 of the second offset sample of the fused lane boundary line data is obtained as follows:

[0029]

[0030] Where μ1 is the sample mean of the first offset sample of the third lane boundary line data, μ oth σ1 is the sample mean of the first offset sample of the fourth lane boundary line data, and σ2 is the sample standard deviation of the first offset sample of the third lane boundary line data. othThe sample standard deviation is the first offset sample of the fourth lane boundary line data.

[0031] In one technical solution of the above-mentioned method for obtaining the confidence level of feature location in a high-precision map, the step of obtaining the confidence level of the data falling into their respective preset confidence intervals by using corresponding probability models based on the sample distribution parameters of the first type of map feature data and the second type of map feature data includes:

[0032] The sample size of the first offset sample is obtained from the distribution parameters of the first offset sample of the first type of map feature data;

[0033] If the number of samples of the first offset sample is greater than or equal to the preset first quantity threshold, a Gaussian model is used to obtain the confidence level of the position where the data falls into the preset confidence interval;

[0034] If the number of samples in the first offset sample is less than a preset first quantity threshold, a t-distribution model is used to obtain the confidence level of the position where the data falls into a preset confidence interval; and / or

[0035] The sample size of the second offset sample is obtained from the distribution parameters of the second offset sample of the second type of map feature data;

[0036] If the number of samples of the second offset sample is greater than or equal to the preset second quantity threshold, a linear combination of Gaussian models is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0037] If the number of samples of the second offset sample is less than the preset second quantity threshold, a linear combination of the t-distribution model is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0038] The location confidence level is obtained by integrating the adopted probability model within the preset confidence interval.

[0039] In one technical solution of the above-mentioned method for obtaining the confidence level of feature location in a high-precision map, the step of obtaining the distribution parameters of the first offset sample of the first type of map feature data based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target includes:

[0040] Based on the geometric classification of the first type of map feature data, determine the correlation dimension related to the location accuracy of the first type of map feature data;

[0041] Based on the offset of the position data of the first type of map feature data in the correlation dimension relative to the position data of the preset reference target in the correlation dimension, the distribution parameters of the first offset sample of the first type of map feature data are obtained.

[0042] In one technical solution of the above-mentioned method for obtaining the confidence level of feature locations in high-precision maps, the method further includes:

[0043] The location confidence of the map elements is presented on the high-precision map.

[0044] In a second aspect, a control device is provided, comprising a processor and a storage device, the storage device being adapted to store a plurality of program codes, the program codes being adapted to be loaded and run by the processor to execute the high-precision map feature location confidence acquisition method described in any of the above-described technical solutions.

[0045] In a third aspect, a computer-readable storage medium is provided, wherein a plurality of program codes are stored therein, the program codes being adapted to be loaded and run by a processor to perform the method for obtaining the confidence level of the feature location of a high-precision map as described in any of the technical solutions of the above-described method for obtaining the confidence level of the feature location of a high-precision map.

[0046] In a fourth aspect, a vehicle is provided, the vehicle including the control device described in the above-mentioned control device technical solution.

[0047] The above-described technical solutions of the present invention have at least one or more of the following beneficial effects:

[0048] In implementing the technical solution of this invention, the invention can obtain a first type of map element data based on vehicle-mounted sensors, and calculate a second type of map element data based on the first type of map element data. Then, based on the offset of the position data of the first type of map element data relative to the position data of a preset reference target, the distribution parameters of a first offset sample of the first type of map element data are obtained. Based on the distribution parameters of the first offset sample, the distribution parameters of a second offset sample of the second type of map element data are obtained. Corresponding probability models are then used to obtain the position confidence levels of the first and second type of map element data within their respective preset confidence intervals, based on the distribution parameters of the first and second offset samples. Through this configuration, the invention can effectively obtain the position accuracy of the first and second type of map element data based on crowdsourced first type of map element data. Within a preset position accuracy range, the position accuracy of the first and second type of map element data is accurately predicted, providing better data support for high-precision map production, element iterative updates, and assisting autonomous vehicles in achieving safer path planning.

[0049] Solution 1. A method for obtaining the confidence level of feature location in a high-precision map, characterized in that the method includes:

[0050] Acquire first-type map feature data, wherein the first-type map feature data is obtained by sensing through vehicle-mounted sensors;

[0051] Acquire second type of map feature data, wherein the second type of map feature data is calculated based on the first type of map feature data;

[0052] Based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target, the distribution parameters of the first offset sample of the first type of map feature data are obtained;

[0053] Based on the distribution parameters of the first offset sample, obtain the distribution parameters of the second offset sample of the second type of map feature data;

[0054] For the first type of map feature data and the second type of map feature data, the corresponding probability models are used according to their respective sample distribution parameters to obtain the position confidence level of the data falling into their respective preset confidence intervals, wherein the preset confidence intervals are related to the preset position accuracy.

[0055] Option 2. The method according to Option 1, characterized in that,

[0056] The second type of map feature data is linear combination feature data obtained by linearly combining multiple first type map feature data that meet the combination conditions according to a linear relationship;

[0057] The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes:

[0058] Based on the linear relationship, the distribution parameters of the first offset samples of the multiple first-class map feature data are linearly combined to obtain the distribution parameters of the second offset samples of the second-class map feature data.

[0059] Option 3. The method according to Option 2, characterized in that,

[0060] The plurality of first-type map feature data includes first lane boundary line data and second lane boundary line data, and the second-type map feature data is lane centerline data;

[0061] The step of linearly combining the distribution parameters of the first offset samples of the plurality of first-type map feature data according to the linear relationship to obtain the distribution parameters of the second offset samples of the second-type map feature data includes:

[0062] The sample mean μ2 of the second offset sample of the lane centerline data is obtained as follows:

[0063] The sample standard deviation σ² of the second offset sample of the lane centerline data is:

[0064] Wherein, μ1 is the sample mean of the first offset sample of the first lane boundary line data, μ1' is the sample mean of the first offset sample of the second lane boundary line data, σ1 is the sample standard deviation of the first offset sample of the first lane boundary line data, and σ1' is the sample standard deviation of the first offset sample of the second lane boundary line data.

[0065] Option 4. The method according to Option 1, characterized in that,

[0066] The second type of map element data is the fused element data obtained by fusing multiple first type map element data that meet the fusion conditions from the previously obtained first type map element data and the currently obtained first type map element data;

[0067] The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes:

[0068] Based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data, the distribution parameters of the second offset sample of the fused feature data are obtained.

[0069] Option 5. The method according to Option 4, characterized in that,

[0070] The multiple first-class map feature data include the previously obtained third lane boundary line data and the currently obtained fourth lane boundary line data, and the fused feature data is the fused lane boundary line data;

[0071] The step of obtaining the distribution parameters of the second offset sample of the fused feature data based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data includes: The sample mean μ3 of the second offset sample of the fused lane boundary line data is obtained:

[0072] The sample standard deviation σ3 of the second offset sample of the fused lane boundary line data is obtained as follows:

[0073]

[0074] Where μ1 is the sample mean of the first offset sample of the third lane boundary line data, μ oth σ1 is the sample mean of the first offset sample of the fourth lane boundary line data, and σ2 is the sample standard deviation of the first offset sample of the third lane boundary line data. oth The sample standard deviation is the first offset sample of the fourth lane boundary line data.

[0075] Solution 6. The method according to any one of Solutions 1-5, characterized in that, for the first type of map feature data and the second type of map feature data, obtaining the position confidence degree of the data falling into their respective preset confidence intervals by using corresponding probability models according to their respective sample distribution parameters, includes:

[0076] The sample size of the first offset sample is obtained from the distribution parameters of the first offset sample of the first type of map feature data;

[0077] If the number of samples of the first offset sample is greater than or equal to the preset first quantity threshold, a Gaussian model is used to obtain the confidence level of the position where the data falls into the preset confidence interval;

[0078] If the number of samples in the first offset sample is less than a preset first quantity threshold, a t-distribution model is used to obtain the confidence level of the position where the data falls into a preset confidence interval; and / or

[0079] The sample size of the second offset sample is obtained from the distribution parameters of the second offset sample of the second type of map feature data;

[0080] If the number of samples of the second offset sample is greater than or equal to the preset second quantity threshold, a linear combination of Gaussian models is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0081] If the number of samples of the second offset sample is less than the preset second quantity threshold, a linear combination of the t-distribution model is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0082] The location confidence level is obtained by integrating the adopted probability model within the preset confidence interval.

[0083] Option 7. The method according to Option 1, characterized in that,

[0084] The step of obtaining the distribution parameters of the first offset sample of the first type of map feature data based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target includes:

[0085] Based on the geometric classification of the first type of map feature data, determine the correlation dimension related to the location accuracy of the first type of map feature data;

[0086] Based on the offset of the position data of the first type of map feature data in the correlation dimension relative to the position data of the preset reference target in the correlation dimension, the distribution parameters of the first offset sample of the first type of map feature data are obtained.

[0087] Solution 8. The method according to Solution 1, characterized in that the method further includes:

[0088] The location confidence of the map elements is presented on the high-precision map.

[0089] Scheme 9. A control device comprising a processor and a storage device, the storage device being adapted to store a plurality of program codes, characterized in that the program codes are adapted to be loaded and run by the processor to perform the feature location confidence acquisition method for a high-precision map as described in any one of Schemes 1 to 8.

[0090] Scheme 10. A computer-readable storage medium storing a plurality of program codes, characterized in that the program codes are adapted to be loaded and run by a processor to perform a method for obtaining feature location confidence of a high-precision map as described in any one of Schemes 1 to 8.

[0091] Option 11. A vehicle, characterized in that the vehicle includes the control device described in Option 9. Attached Figure Description

[0092] The disclosure of this invention will become more readily understood with reference to the accompanying drawings. It will be readily understood by those skilled in the art that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Wherein:

[0093] Figure 1 This is a flowchart illustrating the main steps of a method for obtaining the confidence level of feature locations in a high-precision map according to an embodiment of the present invention.

[0094] Figure 2 This is a schematic diagram illustrating the main process of obtaining fused element data according to one embodiment of the present invention.

[0095] Figure 3 This is a schematic diagram illustrating the main process of obtaining second-type map feature data based on first-type map feature data in a high-precision map according to an embodiment of the present invention.

[0096] Figure 4 It is based on Figure 3A schematic diagram showing the location confidence results of the first and second types of map feature data;

[0097] Figure 5 This is a schematic flowchart of the main steps of a method for obtaining the confidence level of feature location in a high-precision map according to an embodiment of the present invention.

[0098] Figure 6 This is a schematic diagram illustrating the relationship between location confidence and confidence interval in an example according to an embodiment of the present invention;

[0099] Figure 7 This is a schematic diagram comparing the distribution of linear combination feature data and first type map feature data in an example according to an embodiment of the present invention.

[0100] Figure 8 This is a schematic diagram comparing the distribution of fused element data and first-class map element data in an example according to an embodiment of the present invention. Detailed Implementation

[0101] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0102] In the description of this invention, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and may also include software components, such as program code, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Non-transitory computer-readable storage media includes any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.

[0103] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of a method for obtaining the confidence level of feature locations in a high-precision map according to an embodiment of the present invention. Figure 1As shown, the method for obtaining the confidence level of feature location in a high-precision map in this embodiment of the invention mainly includes the following steps S101-S105.

[0104] Step S101: Obtain the first type of map feature data, wherein the first type of map feature data is obtained by sensing through vehicle-mounted sensors.

[0105] In this embodiment, the first type of map feature data can be lane boundaries, road boundaries, traffic signs, etc., which can be obtained by measuring feature data through vehicle-mounted sensors.

[0106] In one implementation, the vehicle sensor can be an onboard sensor currently installed on the vehicle, such as an onboard camera or onboard LiDAR.

[0107] In one implementation, the vehicle-mounted sensor can be a vehicle-mounted sensor installed on multiple vehicles, that is, the first type of map feature data collected by the vehicle-mounted sensors on multiple vehicles is obtained through crowdsourcing.

[0108] Step S102: Obtain the second type of map feature data, which is calculated based on the first type of map feature data.

[0109] In this embodiment, the second type of map feature data can be map feature data calculated based on the first type of map feature data. The calculation can be linear combination, semantic calculation, or fusion calculation, etc.

[0110] In one implementation, the second type of map feature data can be lane centerlines, lane speed limits, etc., obtained by linear combination or semantic calculation of the first type of map feature data.

[0111] In one implementation, the second type of map feature data can be fused feature data such as lane boundaries and road boundaries obtained by fusion calculation from the first type of map feature data or the second type of map feature data obtained through prior calculation.

[0112] Step S103: Based on the offset of the location data of the first type of map feature data relative to the location data of the preset reference target, obtain the distribution parameters of the first offset sample of the first type of map feature data.

[0113] In this embodiment, a preset reference target can be pre-defined and used as a lane reference line to calculate the offset of the first type of map feature data relative to the preset reference target, thereby obtaining the distribution parameters of the first offset sample of the first type of map feature data. The preset reference target can be a reference line obtained by aggregating the positions of previously obtained first type of map feature data (e.g., lane boundaries), a reference line obtained by aggregating multiple traffic trajectories, or a road data point in a regular map.

[0114] In one embodiment, the first map feature data can be mapped to the reference base coordinate system where the preset reference target is located, and the offset of the first map feature data relative to the preset reference target in the three dimensions of horizontal, vertical and elevation can be calculated as the first offset sample, and the distribution parameters of the first offset sample can be obtained based on the first offset sample.

[0115] In one implementation, a correlation dimension related to the positional accuracy of the first type of map feature data can be determined based on the geometric classification of the first type of map feature data. Then, based on the offset of the positional data of the first type of map feature data in the correlation dimension relative to the positional data of a preset reference target in the correlation dimension, the distribution parameters of the first offset sample of the first type of map feature data can be obtained. The geometric classification can include linear features and point features. For example, lane boundaries and road boundaries are classified as linear features; traffic signs and poles are classified as point features. The positional accuracy of linear features is highly correlated with lateral and elevation offsets, while the positional accuracy of point features is highly correlated with lateral, longitudinal, and elevation offsets.

[0116] Step S104: Based on the distribution parameters of the first offset sample, obtain the distribution parameters of the second offset sample of the second type of map feature data.

[0117] In this embodiment, the distribution parameters of the second offset sample of the second type of map feature data can be obtained based on the distribution parameters of the first offset sample.

[0118] Step S105: For the first type of map feature data and the second type of map feature data, the corresponding probability models are used according to their respective sample distribution parameters to obtain the location confidence of the data falling into their respective preset confidence intervals, wherein the preset confidence intervals are related to the preset location accuracy.

[0119] In this embodiment, a probabilistic model can be applied to calculate the location confidence level of each of the first and second types of map feature data based on the sample distribution parameters of the first and second types of map feature data, respectively, using preset confidence intervals. The preset confidence intervals can be determined by a preset location accuracy.

[0120] In one implementation, the location confidence levels of the first type of map feature data and the second type of map feature data can be presented on a high-precision map to provide data support for high-precision map production, feature iterative updates, and assisting autonomous vehicles.

[0121] Based on steps S101-S105 above, this embodiment of the invention can obtain a first type of map element data based on vehicle-mounted sensors, and calculate a second type of map element data based on the first type of map element data. Then, based on the offset of the position data of the first type of map element data relative to the position data of a preset reference target, it obtains the distribution parameters of a first offset sample of the first type of map element data. Based on the distribution parameters of the first offset sample, it obtains the distribution parameters of a second offset sample of the second type of map element data. Using corresponding probability models based on the distribution parameters of the first and second offset samples, it obtains the position confidence levels of the first and second type of map element data within their respective preset confidence intervals. Through this configuration, this embodiment of the invention can effectively obtain the position accuracy of the first and second type of map element data based on crowdsourced first type of map element data. Within a preset position accuracy range, it accurately predicts the position accuracy of the first and second type of map element data, providing better data support for high-precision map production, element iterative updates, and assisting autonomous vehicles in achieving safer path planning.

[0122] The following provides a further explanation of steps S104 and S105.

[0123] In one embodiment of the present invention, the second type of map feature data is linear combination feature data obtained by linearly combining multiple first type map feature data that meet the combination conditions according to a linear relationship. Step S104 may include the following step S1041:

[0124] Step S1041: Based on the linear relationship, the distribution parameters of the first offset samples of multiple first-class map feature data are linearly combined to obtain the distribution parameters of the second offset samples of second-class map feature data.

[0125] In this embodiment, the multiple first-type map feature data may include first lane boundary line data and second lane boundary line data, and the second-type map feature data may be lane centerline data. Step S1041 may further include the following steps S10411 to S10412:

[0126] Step S10411: Obtain the sample mean μ2 of the second offset sample of the lane centerline data according to the following formula (1):

[0127]

[0128] Step S10412: Obtain the sample standard deviation σ2 of the second offset sample of the lane centerline data according to the following formula (2):

[0129]

[0130] Where μ1 is the sample mean of the first offset sample of the first lane boundary line data, μ1' is the sample mean of the first offset sample of the second lane boundary line data, σ1 is the sample standard deviation of the first offset sample of the first lane boundary line data, and σ1' is the sample standard deviation of the first offset sample of the second lane boundary line data.

[0131] In one embodiment of the present invention, the second type of map feature data can be fused feature data obtained by fusing multiple first type map feature data that meet the fusion conditions from previously obtained first type map feature data and currently obtained first type map feature data. Step S104 may include the following step S1042:

[0132] Step S1042: Based on the distribution parameters of the first offset sample of the first type of map feature data obtained previously and the distribution parameters of the first offset sample of the currently obtained first type of map feature data, obtain the distribution parameters of the second offset sample of the fused feature data.

[0133] In this embodiment, the fusion conditions may include: the lateral offset of the two elements to be fused is less than a preset offset threshold; the angle between the directions of the two elements to be fused is less than a preset angle threshold; and the two elements to be fused have the same attributes, such as both being lane boundaries. Those skilled in the art can select the above fusion conditions or set new fusion conditions according to the needs of actual applications.

[0134] In one embodiment, the multiple first-type map feature data may include previously obtained third lane boundary line data and currently obtained fourth lane boundary line data, and the fused feature data is the fused lane boundary line data. Step S1042 may further include the following steps S10421 and S10422:

[0135] Step S10421: Obtain the sample mean μ3 of the second offset sample of the fused lane boundary line data according to the following formula (3):

[0136]

[0137] Step S10422: Obtain the sample standard deviation σ3 of the second offset sample of the fused lane boundary line data according to the following formula (4):

[0138]

[0139] Where μ1 is the sample mean of the first offset sample of the third lane boundary line data, μ oth σ1 is the sample mean of the first offset sample of the fourth lane boundary line data, and σ2 is the sample standard deviation of the first offset sample of the third lane boundary line data. oth The sample standard deviation of the first offset sample of the fourth lane boundary line data.

[0140] In one implementation, see Appendix Figure 2 , Figure 2 This is a schematic diagram illustrating the main process of obtaining fused element data according to one embodiment of the present invention. Figure 2 As shown, A1, A2, A3, C1, C2, and C3 represent the previously obtained third lane boundary line data, B1, B2, and B3 represent the currently obtained fourth lane boundary line data, and D1 represents the preset reference target. By fusing A1, A2, A3, C1, C2, C3, B1, B2, and B3 accordingly, the fused lane boundary line data can be obtained.

[0141] In one embodiment of the present invention, step S105 may include steps S1051 to S1053:

[0142] Step S1051: Obtain the number of samples of the first offset sample from the distribution parameters of the first offset sample of the first type of map feature data.

[0143] Step S1052: If the number of samples of the first offset sample is greater than or equal to the preset first quantity threshold, a Gaussian model is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0144] Step S1053: If the number of samples of the first offset sample is less than the preset first quantity threshold, the t-distribution model is used to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0145] In this embodiment, for the first type of map feature data collected by crowdsourcing, if the potential system error caused by hardware devices is ignored, then if the number of samples of the first offset sample of the first type of map feature data is greater than or equal to the first quantity threshold, it can be approximately considered that the first offset sample data presents a normal distribution. A Gaussian model can be used to obtain the location confidence of the first map feature data. The first probability density function of the Gaussian model can be obtained by the following formula (5).

[0146]

[0147] Where f1(x) is the first probability density function, x is the offset sample, σ is the standard deviation of the offset sample, and μ is the mean of the offset sample.

[0148] In another scenario, considering the significant differences in the number of crowdsourced data collections across different regions, the first offset samples may exhibit an uneven distribution, meaning some regions have a large number of samples while others have a small number. Therefore, for the first offset samples with a sample size less than the first threshold, a t-distribution model can be used to obtain the location confidence of the first map feature data. The t-distribution is used to estimate the mean of a normally distributed population with unknown variance based on a small sample size. The second probability density function of the t-distribution model can be obtained using the following formula (6):

[0149]

[0150] Where f2(t) is the second probability density function, n is the number of samples, v is the degrees of freedom, v = n-1, and t is the transformation variable of the offset sample, which is obtained according to the following formula (7).

[0151]

[0152] S n Let v be the predicted variance of the offset sample, and Gam be the gamma function. When v is even, the following formula (8) is satisfied:

[0153]

[0154] When v is odd, the following formula (9) is satisfied:

[0155]

[0156] In one embodiment of the present invention, step S105 may further include steps S1054 to S1056:

[0157] Step S1054: Obtain the sample number of the second offset sample from the distribution parameters of the second offset sample of the second type of map feature data.

[0158] Step S1055: If the number of samples of the second offset sample is greater than or equal to the preset second quantity threshold, use a linear combination of Gaussian models to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0159] Step S1056: If the number of samples of the second offset sample is less than the preset second quantity threshold, use a linear combination of the t-distribution model to obtain the confidence level of the position where the data falls into the preset confidence interval.

[0160] In this embodiment, the second type of map element data can also determine the corresponding probability model based on the number of samples. That is, in the case of small samples, the t-distribution model (formula (6)) is used to obtain the location confidence; in the case of sufficient samples, the Gaussian model (formula (5)) is used to obtain the location confidence.

[0161] In one implementation, when the second type of map feature data is fused feature data obtained by fusing multiple first type map feature data that meet the fusion conditions from the previously obtained first type map feature data and the currently obtained first type map feature data, it can be regarded as the fusion of the distribution parameters between the previously obtained first type map feature data and the currently obtained first type map feature data. When the number of samples of the second offset sample of the fused feature data is greater than or equal to the second quantity threshold, the third probability density function of the fused feature data can be obtained according to the following formula (10):

[0162]

[0163] Here, f3(x) is the third probability density function.

[0164] When the number of samples of the second offset sample of the fused feature data is less than the second quantity threshold, the fourth probability density function of the fused feature data can be obtained according to the following formula (11):

[0165]

[0166] Here, f4(t) is the fourth probability density function.

[0167] In one implementation, the first quantity threshold and the second quantity threshold can be 30.

[0168] In one embodiment of the present invention, step S105 may further include the following step S1057:

[0169] Step S1057: Integrate the adopted probability model within a preset confidence interval to obtain the location confidence level.

[0170] In this embodiment, the location confidence level can be obtained by integrating the adopted probability model over a preset confidence interval.

[0171] If a Gaussian model is used, the location confidence can be obtained according to the following formula (12):

[0172]

[0173] Where F(x) can be either f1(x) or f3(x), and δ is the position accuracy.

[0174] If the t-distribution model is used, the location confidence can be obtained according to the following formula (13):

[0175]

[0176] F(t) can be either f2(t) or f4(t).

[0177] For an example, please refer to the appendix. Figures 3 to 8 , Figure 3 This is a schematic diagram illustrating the main process of obtaining second-type map element data based on first-type map element data in a high-precision map according to an embodiment of the present invention. Figure 4 It is based on Figure 3 A schematic diagram showing the location confidence results of the first and second types of map feature data; Figure 5 This is a schematic flowchart of the main steps of a method for obtaining the confidence level of feature location in a high-precision map according to an embodiment of the present invention. Figure 6 This is a schematic diagram illustrating the relationship between location confidence and confidence interval in an example according to an embodiment of the present invention; Figure 7 This is a schematic diagram comparing the distribution of linear combination feature data and first type map feature data in an example according to an embodiment of the present invention. Figure 8 This is a schematic diagram comparing the distribution of fused element data and first-class map element data in an example according to an embodiment of the present invention.

[0178] like Figure 3 As shown, lane boundaries are used as the first type of map feature data, vehicle centerlines are used as linear composite feature data, and lane markings are used as fused feature data. Among them, A-1 is the first type of map feature data obtained earlier in the map, and its first offset sample follows the distribution N(1.8, 0.3^2). The map records that the mean lateral offset of A-1 is 1.8 and the standard deviation of the lateral offset is 0.3. The frame 1 represents the distribution of the offset of boundary A-1.

[0179] Boundary A and boundary B were obtained based on crowdsourced data, with corresponding samples and distributions of X_A~N(2.3, 0.6^2) and X_B~N(-2.1, 0.4^2), respectively. The mean lateral offset of the reported data for boundary A is 2.3, and the standard deviation of the lateral offset is 0.6. The mean lateral offset of the reported data for boundary B is -2.1, and the standard deviation of the lateral offset is 0.4. Figures 2 and 3 illustrate the distribution of the reported element offsets for boundaries A and B, respectively.

[0180] Through judgment, it is obtained that boundary A and the existing boundary A-1 of the base map meet the fusion condition, and boundary A-2 can be obtained after fusion. Calculated by formula (10), the sample X_A2 follows the distribution N(1.9, 0.27^2), wherein the mean value of the lateral offset of boundary A-2 is 1.9, and the standard deviation of the lateral offset is 0.27. Frame 4 represents the distribution diagram of the offset of boundary A-2.

[0181] Through judgment, it is obtained that boundary A-2 and boundary B can form the left and right boundaries of the same lane, and the linear combination element - lane center line C can be calculated. According to formulas (3) to (5), X_C follows the distribution N(-0.1, 0.24^2). The mean value of the geometric lateral offset of boundary C is -0.1, and the standard deviation of the lateral offset is 0.24. Frame 5 represents the distribution diagram of the offset of boundary C. The finally output elements include the first type of map element data - lane boundary B, the fusion element data - lane boundary A-2, and the linear combination element data - lane center line C.

[0182] obtaining Figure 3 the schematic diagram of the main steps and process of the position confidence of the elements shown, as shown in Figure 5 , respectively obtaining the sample distribution parameters of the previously obtained first type of map element data, the first type of map element data reported by crowdsourcing, the fusion element data and the linear combination element data. For the distribution characteristics of different types of elements, the corresponding Gaussian model is obtained according to the distributions of formula (5) and formula (10), or the corresponding t-distribution model is obtained according to formula (6) and formula (11), and the position confidence of different element data is calculated based on the corresponding probability model.

[0183] As shown in Figure 6 , when it is assumed that the element data follows a t-distribution (the number of samples is n=10, then the degree of freedom of T is n-1=9), and P(-A<T<A)=0.8 is satisfied, it represents that the 80% confidence interval of μ is obtained. When it is assumed that T follows a normal distribution and P(-A<T<A)=0.8 is satisfied, it represents that the 80% confidence interval of μ is Figure 6 among the 15 groups of samples (each horizontal line is one group), there are 12 groups of values (dark horizontal lines) whose corresponding mean values and confidence intervals contain the true value μ (black dotted line), with an accuracy rate of about 80%, which is consistent with the previous probability integral value.

[0184] In this example, if the position accuracy is assumed to be 0.2m, and the number of samples of boundary A, boundary B and boundary A1 is 17, then in Figure 3 the distribution situation corresponding to the high-precision map scene elements shown, calculated by integrating formula (12) and formula (13), the results as shown in Figure 4The confidence levels for the locations of each element are shown. The confidence level for the location of boundary A is 77%, for boundary B it is 93%, and for boundary A1 it is 98.5%. When merging boundaries A and A1, their locations are relatively close, and the confidence level for the merged boundary locations is increased to 99%. When calculating the lane centerline C using lane boundaries B and A2, the confidence level for the calculated centerline C location is further increased to 99.6% due to the concentrated boundary locations.

[0185] like Figure 7 As shown in the figure, tile 1 represents the offset distribution X of element B of the lane. B ~N(μ1, σ1) 2 )=N(-2.1,0.4 2 Plot 2 represents the offset distribution X_A2~N(μ2,σ2) of element A2. 2 )=N(1.9,0.27 2 Figure 3 represents the offset distribution X of lane centerline C. C ~N((μ1+μ2) / 2,(σ1) 2 +σ2 2 ) / 4)=N(-0.1,0.24 2 The subdivision information interval corresponding to block 3 is located between the left and right boundaries, which is in line with the model design expectations.

[0186] like Figure 8 As shown, tile 1 represents the offset distribution X_A1~N(μ1,σ1) of feature A1 in the high-precision map. 2 ) = N(1.8, 0.3 2 Plot 2 represents the offset distribution X_A~N(μ2,σ2) of the boundary A corresponding to the crowdsourced report. 2 ) = N(2.3, 0.6 2 Plot 3 represents the offset distribution X after fusion. A2 ~N((μ2*σ1) 2 +μ1*σ2 2 ) / (σ1 2 +σ2 2 ),σ1 2 *σ2 2 / (σ1 2 +σ2 2 ))=N(1.9,0.27 2 When the location μ1 of an existing feature in a high-precision map is very close to the location μ2 reported by the crowdsourced platform and the variance is very small, the probability of the fused boundary distribution is more concentrated, the probability of the true location increases, which meets the design expectations.

[0187] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of the present invention.

[0188] Those skilled in the art will understand that all or part of the processes in the method of the above embodiment of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0189] Furthermore, the present invention also provides a control device. In one embodiment of the control device according to the present invention, the control device includes a processor and a storage device. The storage device can be configured to store a program for executing the feature location confidence acquisition method of the high-precision map described in the above-described method embodiments. The processor can be configured to execute the program in the storage device, which includes, but is not limited to, a program for executing the feature location confidence acquisition method of the high-precision map described in the above-described method embodiments. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. This control device can be a control device device comprising various electronic devices.

[0190] Furthermore, the present invention also provides a computer-readable storage medium. In one embodiment of the computer-readable storage medium according to the present invention, the computer-readable storage medium can be configured to store a program for performing the feature location confidence acquisition method of the high-precision map described in the above-described method embodiments. This program can be loaded and run by a processor to implement the feature location confidence acquisition method of the high-precision map described above. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The computer-readable storage medium can be a storage device comprising various electronic devices. Optionally, in the embodiments of the present invention, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0191] Furthermore, the present invention also provides a vehicle. In one embodiment of the vehicle according to the present invention, the vehicle includes the control device described in the control device embodiments above.

[0192] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of software and hardware. Therefore, the number of modules shown in the figures is merely illustrative.

[0193] Those skilled in the art will understand that the various modules in the device can be adaptively split or combined. Such splitting or combining of specific modules will not cause the technical solution to deviate from the principles of the present invention; therefore, the technical solutions after splitting or combining will fall within the protection scope of the present invention.

[0194] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for obtaining the confidence level of feature locations in a high-precision map, characterized in that, The method includes: Acquire first-type map feature data, wherein the first-type map feature data is obtained by sensing through vehicle-mounted sensors; Acquire second type of map feature data, wherein the second type of map feature data is calculated based on the first type of map feature data; Based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target, the distribution parameters of the first offset sample of the first type of map feature data are obtained; Based on the distribution parameters of the first offset sample, obtain the distribution parameters of the second offset sample of the second type of map feature data; For the first type of map feature data and the second type of map feature data, the corresponding probability models are used according to their respective sample distribution parameters to obtain the position confidence level of the data falling into their respective preset confidence intervals, wherein the preset confidence intervals are related to the preset position accuracy.

2. The method according to claim 1, characterized in that, The second type of map feature data is linear combination feature data obtained by linearly combining multiple first type map feature data that meet the combination conditions according to a linear relationship; The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes: Based on the linear relationship, the distribution parameters of the first offset samples of the multiple first-class map feature data are linearly combined to obtain the distribution parameters of the second offset samples of the second-class map feature data.

3. The method according to claim 2, characterized in that, The plurality of first-type map feature data includes first lane boundary line data and second lane boundary line data, and the second-type map feature data is lane centerline data; The step of linearly combining the distribution parameters of the first offset samples of the plurality of first-type map feature data according to the linear relationship to obtain the distribution parameters of the second offset samples of the second-type map feature data includes: The sample mean of the second offset sample of the lane centerline data. : ; The sample standard deviation of the second offset sample of lane centerline data. : ; in, The first offset sample of the first lane boundary line data is the sample mean. This is the sample mean of the first offset sample of the second lane boundary line data. The standard deviation of the first offset sample of the first lane boundary line data is given. The sample standard deviation of the first offset sample of the second lane boundary line data.

4. The method according to claim 1, characterized in that, The second type of map element data is the fused element data obtained by fusing multiple first type map element data that meet the fusion conditions from the previously obtained first type map element data and the currently obtained first type map element data; The step of obtaining the distribution parameters of the second offset sample of the second type of map feature data based on the distribution parameters of the first offset sample includes: Based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data, the distribution parameters of the second offset sample of the fused feature data are obtained.

5. The method according to claim 4, characterized in that, The multiple first-class map feature data include the previously obtained third lane boundary line data and the currently obtained fourth lane boundary line data, and the fused feature data is the fused lane boundary line data; The step of obtaining the distribution parameters of the second offset sample of the fused feature data based on the distribution parameters of the first offset sample of the previously obtained first type of map feature data and the distribution parameters of the first offset sample of the currently obtained first type of map feature data includes: The sample mean of the second offset sample of the fused lane boundary line data is obtained. : ； The sample standard deviation of the second offset sample of the fused lane boundary line data is obtained. : in, The first offset sample of the third lane boundary line data is the sample mean. The first offset sample of the fourth lane boundary line data is the sample mean. The standard deviation of the first offset sample of the third lane boundary line data is given. The sample standard deviation is the first offset sample of the fourth lane boundary line data.

6. The method according to any one of claims 1-5, characterized in that, For both the first and second types of map feature data, the corresponding probability models are used based on their respective sample distribution parameters to obtain the location confidence level of the data falling within their respective preset confidence intervals, including: The sample size of the first offset sample is obtained from the distribution parameters of the first offset sample of the first type of map feature data; If the number of samples of the first offset sample is greater than or equal to the preset first quantity threshold, a Gaussian model is used to obtain the confidence level of the position where the data falls into the preset confidence interval; If the number of samples in the first offset sample is less than a preset first quantity threshold, a t-distribution model is used to obtain the confidence level of the position where the data falls into a preset confidence interval; and / or The sample size of the second offset sample is obtained from the distribution parameters of the second offset sample of the second type of map feature data; If the number of samples of the second offset sample is greater than or equal to the preset second quantity threshold, a linear combination of Gaussian models is used to obtain the confidence level of the position where the data falls into the preset confidence interval. If the number of samples of the second offset sample is less than the preset second quantity threshold, a linear combination of the t-distribution model is used to obtain the confidence level of the position where the data falls into the preset confidence interval. The location confidence level is obtained by integrating the adopted probability model within the preset confidence interval.

7. The method according to claim 1, characterized in that, The step of obtaining the distribution parameters of the first offset sample of the first type of map feature data based on the offset of the location data of the first type of map feature data relative to the location data of a preset reference target includes: Based on the geometric classification of the first type of map feature data, a correlation dimension related to the location accuracy of the first type of map feature data is determined; wherein, the geometric classification includes linear features and point features; the correlation dimension of the linear features includes lateral offset and elevation offset; the correlation dimension of the point features includes lateral offset, vertical offset, and elevation offset. Based on the offset of the position data of the first type of map feature data in the correlation dimension relative to the position data of the preset reference target in the correlation dimension, the distribution parameters of the first offset sample of the first type of map feature data are obtained.

8. The method according to claim 1, characterized in that, The method further includes: The location confidence of the map elements is presented on the high-precision map.

9. A control device, comprising a processor and a storage device, said storage device being adapted to store a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by the processor to perform the feature location confidence acquisition method for a high-precision map according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by a processor to perform the method for obtaining the feature location confidence of a high-precision map according to any one of claims 1 to 8.

11. A vehicle, characterized in that, The vehicle includes the control device as described in claim 9.

Images