Vehicle location methods, vehicles, and storage media

Abstract

This invention discloses a vehicle localization method, a vehicle, and a storage medium. The method includes: acquiring first and second sensing data of the area where the vehicle is located; determining the correlation degree between target attributes of each first instance in the first sensing data and each second instance in the second sensing data, obtaining multiple correlation degrees; determining a one-to-one matching relationship between the multiple first instances and multiple second instances based on an association matrix constructed from the multiple correlation degrees, obtaining multiple matching groups, each matching group including first instances and second instances with matching relationships; determining position change parameters between the first and second sensing data based on multiple displacement estimates corresponding to the multiple matching groups; and determining the vehicle's localization within the area based on the position change parameters. This invention aims to improve the accuracy and robustness of mapping or localization.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and more particularly to vehicle positioning methods, vehicles, and storage media. Background Technology

[0002] Simultaneous Localization and Mapping (SLAM) is a key technology for autonomous driving, especially playing a crucial role in automated parking. The core of SLAM technology is mapping and localization. Both require using perception data detected by cameras, LiDAR, etc., to obtain pose changes between different frames, and then obtaining a fused localization or mapping result based on the determined positional changes.

[0003] However, perception data has perception defects such as false detection, missed detection, and attribute bias, which cause the perception data to not be ideally matched with each other and with the pre-stored map, resulting in pose estimation errors and affecting the accuracy of mapping or positioning. Summary of the Invention

[0004] The main objective of this invention is to provide a vehicle positioning method, a vehicle, and a storage medium, which aims to improve the accuracy and robustness of mapping or positioning.

[0005] To achieve the above objectives, the present invention provides a vehicle positioning method, the vehicle positioning method comprising the following steps:

[0006] Acquire first and second perception data of the area where the vehicle is located;

[0007] Determine the correlation degree between the target attributes of each first instance in the first perception data and each second instance in the second perception data to obtain multiple correlation degrees;

[0008] Based on the association matrix constructed from the multiple association degrees, a one-to-one matching relationship between multiple first instances and multiple second instances is determined, and multiple matching groups are obtained. Each matching group includes first instances and second instances that have a matching relationship.

[0009] The position change parameters between the first sensing data and the second sensing data are determined based on the multiple displacement estimates corresponding to the multiple matching groups, and the vehicle's location in the area is determined based on the position change parameters.

[0010] Optionally, the target attribute includes multiple sub-attributes, and the step of determining the correlation degree corresponding to the target attribute between each first instance in the first sensing data and each second instance in the second sensing data, and obtaining multiple correlation degrees, includes:

[0011] Combine all the first instances with all the second instances in pairs to obtain multiple instance groups;

[0012] The sub-association degree corresponding to the sub-attribute in each instance group is determined based on the first feature value and the second feature value corresponding to each sub-attribute in each instance group, wherein the first feature value is a feature of the first instance and the second feature value is a feature of the second instance;

[0013] The association degree of the corresponding instance group is determined based on the multiple sub-associations in each instance group and their corresponding first weight values, thereby obtaining the multiple association degrees.

[0014] Among them, the sub-associations corresponding to different sub-attributes have different first weight values.

[0015] Optionally, the target attribute includes multiple sub-attributes, including: category, color, area, centroid position, and principal component orientation.

[0016] Optionally, the step of determining the one-to-one matching relationship between the multiple first instances and the multiple second instances based on the association matrix constructed according to the multiple association degrees, and obtaining multiple matching groups, includes:

[0017] The correlation matrix is ​​constructed based on all correlation degrees greater than a preset value;

[0018] The correlation matrix is ​​solved according to the preset matching algorithm to obtain the solution result;

[0019] The multiple matching groups are determined based on the solution results.

[0020] Optionally, the step of determining the position change parameter between the first sensing data and the second sensing data based on the multiple displacement estimates corresponding to the multiple matching groups includes:

[0021] The global displacement is calculated based on the multiple displacement estimates and their corresponding second weight values; the second weight values ​​are determined based on the corresponding correlation degree.

[0022] The position change parameters are determined based on the global displacement.

[0023] Optionally, the step of determining the position change parameter based on the global displacement includes:

[0024] The position change parameters are determined based on the global displacement and its corresponding third weight value, and the predicted displacement and its corresponding fourth weight value. The predicted displacement is determined based on the mechanical odometer detection parameters of the vehicle.

[0025] Optionally, before the step of determining the position change parameter based on the global displacement and its corresponding third weight value and the predicted displacement and its corresponding fourth weight value, the method further includes:

[0026] Determine the matching error between the first sensing data and the second sensing data;

[0027] The third weight value is determined based on the matching error.

[0028] Optionally, the step of determining the matching error between the first sensed data and the second sensed data includes:

[0029] The first sensing data is translated according to the global displacement to obtain reference sensing data;

[0030] The displacement deviation between each second instance and its corresponding third instance is determined to obtain multiple displacement deviations; the third instance is the instance in the reference sensing data after the first instance matched with the corresponding second instance has been translated.

[0031] The matching error is determined based on the plurality of displacement deviations.

[0032] Optionally, the first sensing data and the second sensing data are two sensing data collected by the vehicle at adjacent time points;

[0033] Alternatively, the first sensing data may be pre-stored map data of the area, and the second sensing data may be sensing data currently collected by the vehicle.

[0034] In addition, to achieve the above objectives, this application also proposes a vehicle comprising: a memory, a processor, and a vehicle positioning program stored in the memory and executable on the processor, wherein the vehicle positioning program, when executed by the processor, implements the steps of the vehicle positioning method as described in any of the preceding claims.

[0035] In addition, to achieve the above objectives, this application also proposes a storage medium storing a vehicle positioning program, which, when executed by a processor, implements the steps of the vehicle positioning method as described in any of the preceding claims.

[0036] This invention proposes a vehicle positioning method. The method acquires different sensing data corresponding to the vehicle's location area, determines the correlation degree between all instances of the different sensing data to obtain multiple correlation degrees, and uses an association matrix constructed based on these multiple correlation degrees to perform one-to-one matching between two sets of sensing data to obtain multiple matching groups. Based on multiple displacement estimates corresponding to the multiple matching groups, the method determines the positional change parameters between the two sets of sensing data, and performs vehicle-related mapping or positioning based on these positional change parameters. The association matrix accurately reflects the similarity between instances of the two sets of sensing data. Instances of sensing data are matched based on similarity, rather than directly, thus avoiding the impact of sensing defects in the sensing data on the accuracy and robustness of vehicle-related mapping or positioning, effectively improving the accuracy and robustness of mapping or positioning. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the hardware structure involved in the operation of a vehicle according to an embodiment of the present invention;

[0038] Figure 2 This is a flowchart illustrating an embodiment of the vehicle positioning method of the present invention;

[0039] Figure 3 This is a flowchart illustrating another embodiment of the vehicle positioning method of the present invention;

[0040] Figure 4 This is a flowchart illustrating another embodiment of the vehicle positioning method of the present invention;

[0041] Figure 5 This is a flowchart illustrating another embodiment of the vehicle positioning method of the present invention.

[0042] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0043] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0044] The main solution of this invention is as follows: acquiring first and second sensing data of the area where the vehicle is located; determining the correlation degree between each first instance in the first sensing data and each second instance in the second sensing data, obtaining multiple correlation degrees; determining a one-to-one matching relationship between multiple first instances and multiple second instances based on the correlation matrix constructed by the multiple correlation degrees, obtaining multiple matching groups, each matching group including first instances and second instances with matching relationships; determining position change parameters between the first sensing data and the second sensing data based on multiple displacement estimates corresponding to the multiple matching groups, and determining the positioning of the vehicle in the area based on the position change parameters.

[0045] Because existing technologies suffer from perceptual defects such as false detections, missed detections, and attribute biases in the perceived data, the perceived data cannot be ideally matched with each other or with the pre-stored map, resulting in pose estimation errors and affecting the accuracy of mapping or positioning.

[0046] The present invention provides the above-mentioned solution, which aims to improve the accuracy and robustness of mapping or positioning.

[0047] This invention provides a vehicle. Specifically, the vehicle can be any type of vehicle, such as a car or a truck.

[0048] In this embodiment, the vehicle includes an onboard sensor 1, a vehicle positioning device 2, and a mechanical odometer 3. Both the onboard sensor 1 and the mechanical odometer 3 are connected to the vehicle positioning device 2. The onboard sensor 1 is used to collect perception data of the area where the vehicle is located; the onboard sensor 1 may include a camera, lidar, etc. The mechanical odometer 3 is used to detect the vehicle's mileage data; the mechanical odometer 3 may include a wheel speedometer or an inertial navigation odometer, etc.

[0049] In this embodiment of the invention, reference is made to Figure 1 The vehicle positioning device includes a processor 1001 (e.g., CPU), a memory 1002, etc. The components in the control device are connected via a communication bus. The memory 1002 can be a high-speed RAM or a stable, non-volatile memory, such as a disk drive. Optionally, the memory 1002 can also be a storage device independent of the aforementioned processor 1001.

[0050] Those skilled in the art will understand that Figure 1 The device structure shown does not constitute a limitation on the device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0051] like Figure 1 As shown, the memory 1002, which serves as a storage medium, may include a vehicle positioning program. Figure 1 In the device shown, the processor 1001 can be used to call the vehicle positioning program stored in the memory 1002 and execute the relevant steps of the vehicle positioning method in the following embodiments.

[0052] This invention also provides a vehicle positioning method, applied to the aforementioned vehicle.

[0053] Reference Figure 2 This application proposes an embodiment of the vehicle positioning method. In this embodiment, the vehicle positioning method includes:

[0054] Step S10: Obtain first and second perception data of the area where the vehicle is located;

[0055] In this embodiment, both the first perception data and the second perception data are bird's-eye views of the area where the vehicle is located.

[0056] Both the first and second perception data are detected by onboard sensors.

[0057] The first and second sensing data can both be sensing data collected in real time at two adjacent moments, or both can be pre-stored data, or one can be real-time data and the other can be pre-stored data.

[0058] Step S20: Determine the correlation degree between the target attributes of each first instance in the first perception data and each second instance in the second perception data, and obtain multiple correlation degrees;

[0059] All first instances in the first perception data are paired with all second instances in the second perception data to obtain multiple instance groups, and the correlation degree of the corresponding instance group is determined according to the image feature value corresponding to the target attribute in each instance in each instance group.

[0060] Specifically, in this embodiment, both the first and second sensing data are image data. The first sensing data is segmented into multiple first instances according to an image segmentation algorithm. The attributes of each first instance include, but are not limited to, type, color, area, centroid position, and principal component direction. The second sensing data is segmented into multiple second instances according to an image segmentation algorithm. The attributes of each second instance include, but are not limited to, type, color, area, centroid position, and principal component direction.

[0061] The target attribute here can be a pre-set fixed attribute or an attribute determined based on the region type of the area where the vehicle is located. There can be one or more target attributes, and the corresponding number of first instance feature values ​​and second instance feature values ​​can also be more than one. One or more of the following attributes can be used as target attributes: category, color, area, centroid location, and principal component direction, etc.

[0062] Specifically, each instance group includes a first instance and a second instance, and correspondingly, each instance group has at least two instance feature values. Within each instance group, the instance feature value of the target attribute in the first instance is determined to obtain the first feature value, and the instance feature value of the target attribute in the second instance is determined to obtain the second feature value. The correlation degree of the corresponding instance group can be determined based on the first feature value and the second feature value.

[0063] Specifically, the correspondence between at least two image feature values ​​and their correlation can be preset, and can be a calculation formula or mapping relationship, etc. Based on this correspondence, the correlation of the instance group corresponding to at least two image feature values ​​can be determined. In this embodiment, the correlation of the corresponding instance group can be calculated using the first feature value and the second feature value. In other embodiments, the correlation of the corresponding instance group can also be obtained by looking up a table using the first feature value and the second feature value.

[0064] Each instance group corresponds to a correlation score, which represents the degree of similarity between two instances of different perceptual data within the same instance group. The higher the correlation score, the more similar the corresponding instances are.

[0065] Furthermore, before step S20, after segmenting and obtaining multiple first instances and multiple second instances, mileage data detected by the mechanical odometer can be obtained. Motion compensation is performed on the data in the first sensing data and / or the second sensing data based on the mileage data, and the correlation degree is calculated according to the first instance and the second instance obtained after compensation.

[0066] Step S30: Determine the one-to-one matching relationship between multiple first instances and multiple second instances based on the association matrix constructed by the multiple association degrees, and obtain multiple matching groups, each of the matching groups including the first instance and the second instance with matching relationship;

[0067] Specifically, a one-to-one matching relationship means that one first instance corresponds to one second instance, and each matching group includes one first instance and one second instance.

[0068] The correlation matrix includes all correlation degrees between the first and second perceived data. Specifically, a pre-set optimization objective can be obtained, the correlation matrix can be solved according to the optimization objective to obtain the optimal solution, and the one-to-one matching relationship can be determined based on the optimal solution obtained from the correlation matrix.

[0069] The optimization objectives here include minimizing the association score between the first and second instances in a matching group, minimizing the sum of association scores for all matching groups, or maximizing the association data.

[0070] Step S40: Determine the position change parameter between the first sensing data and the second sensing data based on the multiple displacement estimates corresponding to the multiple matching groups, and determine the vehicle's location in the area based on the position change parameter.

[0071] Each matching group can correspond to a displacement estimate. The displacement estimate corresponding to the matching group can be determined according to a preset displacement estimation algorithm (such as the ICP algorithm).

[0072] Position change parameters between two sensing data points can be determined based on multiple displacement estimates. Based on these parameters, the vehicle can be located, and the first and second sensing data can be stitched together to obtain a global map of the vehicle's location. Alternatively, the vehicle's position change within its area can be determined based on the determined position change parameters.

[0073] This invention proposes a vehicle positioning method. The method acquires different sensing data corresponding to the area where the vehicle is located, determines the correlation degree between all instances of the different sensing data to obtain multiple correlation degrees, and uses an association matrix constructed based on these multiple correlation degrees to perform one-to-one matching of two sensing data sets to obtain multiple matching groups. Based on multiple displacement estimates corresponding to the multiple matching groups, the method determines the positional change parameters between the two sensing data sets, and performs vehicle-related mapping or positioning based on the positional change parameters. The association matrix accurately reflects the similarity between instances of the two sensing data sets. Instances of the sensing data are matched based on similarity, rather than directly, thereby avoiding the impact of sensing defects in the sensing data on the accuracy and robustness of vehicle-related mapping or positioning, and effectively improving the accuracy and robustness of mapping or positioning.

[0074] Furthermore, the first perception data and the second perception data are two perception data collected by the vehicle at adjacent times; or, the first perception data is pre-stored map data of the area, and the second perception data is the perception data currently collected by the vehicle.

[0075] Thus, this technology can be used to locate a vehicle in an environment without a map, using first and second perception data. Then, based on the location, the vehicle's position information and the perception data obtained by the onboard sensors can be determined to construct a map of the vehicle's journey.

[0076] On the other hand, when a vehicle is in an environment with an existing map, the vehicle's predicted location is obtained based on the onboard system such as GPS, inertial navigation, wheel speed sensor, etc. Based on this location, the predicted position of the vehicle in the map and the map data corresponding to the predicted position are obtained as the first perception data. Then, the map data and the second perception data currently collected by the onboard sensors are matched to complete the vehicle's positioning in the map, thereby improving the positioning accuracy.

[0077] Furthermore, based on the above embodiments, another embodiment of the vehicle positioning method of this application is proposed. In this embodiment, the target attribute includes multiple sub-attributes, referred to... Figure 3 Step S20 includes:

[0078] Step S21: Combine all the first instances with all the second instances in pairs to obtain multiple instance groups;

[0079] Based on the principle of forming an instance group with one first instance and one second instance, we can obtain multiple instance groups by pairwise combining multiple first instances and multiple second instances. There are N first instances, M second instances, and N*M instance groups. For example, if multiple first instances include A1, A2, and A3, and multiple second instances include B1, B2, and B3, then the resulting instance groups are A1B1, A2B2, A3B3, A1B2, A1B3, A2B1, A2B3, A3B1, and A3B2.

[0080] Step S22: Determine the sub-association degree corresponding to the sub-attribute in the corresponding instance group based on the first feature value and the second feature value corresponding to each sub-attribute in each instance group, wherein the first feature value is a feature of the first instance and the second feature value is a feature of the second instance;

[0081] Specifically, the deviation between the first feature value and the second feature value corresponding to each sub-attribute is determined, and the sub-association degree corresponding to each sub-attribute is determined based on the deviation. Since each instance group contains feature values ​​corresponding to multiple sub-attributes, multiple sub-associations corresponding to multiple sub-attributes can be determined for each instance group.

[0082] Step S23: Determine the correlation degree of the corresponding instance group based on the multiple sub-correlation degrees and their corresponding first weight values ​​in each instance group, and obtain the multiple correlation degrees; wherein, the sub-correlation degrees corresponding to different sub-attributes have different first weight values.

[0083] The first weight value corresponding to the sub-attribute can be a pre-set fixed value, or it can be a value determined according to the actual driving status of the vehicle (such as vehicle speed) or the environment of the area (such as urban roads, highways, parking lots, etc.).

[0084] The association degree corresponding to each instance group can be obtained by weighting and averaging multiple sub-associations based on multiple first weight values. Each instance group corresponds to one association degree, and multiple instance groups can have multiple association degrees.

[0085] In this embodiment, the multiple sub-attributes include: category, color, area, centroid position, and principal component direction, where the principal component direction is the direction in which the principal component exists in the corresponding instance. Taking the multiple sub-attributes mentioned here as an example, F1 is defined as the sub-association degree corresponding to the category, F2 as the sub-association degree corresponding to the color, F3 as the sub-association degree corresponding to the area, F4 as the sub-association degree corresponding to the centroid position, and F5 as the sub-association degree corresponding to the principal component direction. The process of determining the association degree involved in this embodiment is explained as follows:

[0086] The specific details for F1 can be found in the following table:

[0087]

[0088] Where S1 is the category of the first instance and S2 is the category of the second instance, the value obtained by querying the table above through the categories of the first instance and the second instance can be used as the sub-association degree F1 corresponding to the category in the corresponding instance group.

[0089] F2 can be calculated using the following formula 1:

[0090] F2=exp[-1*((u1-u2)^2+(v1-v2)^2)]——Formula 1;

[0091] Where u1 and v1 are the average values ​​of each pixel corresponding to the U and V channels in the first instance, and u2 and v2 are the average values ​​of each pixel corresponding to the U and V channels in the second instance.

[0092] F3 can be calculated using the following formula 2:

[0093] F3 = exp[-1*(p1-p2)^2] — Formula 2;

[0094] Where p1 is the number of pixels in the first instance and p2 is the number of pixels in the second instance.

[0095] F4 can be calculated using the following formula 3:

[0096] F4=exp[-1*((x1-x 2)^2+(y1-y 2)^2)]——Formula 3;

[0097] Where x1 and y1 are the centroid coordinates of the first instance, and x2 and y2 are the centroid coordinates of the second instance.

[0098] F5 can be calculated using the following formulas 4 and 5:

[0099] d=min((abs(a1-a2), abs(a1-a2+pi), abs(a1-a2-pi)))——Formula 4;

[0100] F5 = exp[-1*d^2)] — Formula 5;

[0101] Where a1 is the direction of the principal component obtained by principal component analysis of the first instance, and a2 is the direction of the principal component obtained by principal component analysis of the second instance.

[0102] After obtaining F1, F2, F3, F4, and F5, the correlation degree of the corresponding instance group can be calculated according to the following formula 6:

[0103] F pair =w1*F1+w2*F2+w3*F3+w4*F4+w5*F5——Formula 6;

[0104] Among them, w1, w2, w3, w4, and w5 are the first weight values ​​corresponding to F1, F2, F3, F4, and F5, respectively.

[0105] It should be noted that, in addition to the sub-attributes mentioned above, multiple sub-attributes may also include other sub-attributes, such as the types of other instances adjacent to this instance, the distance from the reference position, etc.

[0106] In this embodiment, the correlation degree between any two instances of two sensing data is determined by weighting the sub-correlation degrees corresponding to multiple sub-attributes. This ensures that the determined correlation degree can accurately reflect the similarity between two instances in the corresponding instance group, which is beneficial to improving the accuracy of subsequent matching results and further improving the accuracy of mapping or positioning related to vehicle operation.

[0107] Furthermore, based on any of the above embodiments, another embodiment of the vehicle positioning method of this application is proposed. In this embodiment, reference is made to... Figure 4 Step S30 includes:

[0108] Step S31: Construct the correlation matrix based on all correlation degrees greater than a preset value;

[0109] The preset value can be a fixed value set in advance, or it can be a value determined according to the environmental condition parameters (such as temperature, humidity and / or weather) of the area where the vehicle is located.

[0110] The association matrix can include all association degrees greater than a preset value.

[0111] Step S32: Solve the correlation matrix according to the preset matching algorithm to obtain the solution result;

[0112] The preset matching algorithm may include the Hungarian algorithm or the JPDA algorithm, etc. Specifically, based on the above-mentioned optimization objective, the preset matching algorithm is used to find the optimal solution in the correlation matrix that satisfies the optimization objective, and the solution result includes multiple sets of instances that satisfy the optimization objective.

[0113] Step S33: Determine the multiple matching groups based on the solution results.

[0114] The solution results include multiple instance groups that satisfy the optimization objective, and these multiple instance groups can serve as multiple matching groups here.

[0115] In this embodiment, an association matrix is ​​constructed based on all association degrees greater than a preset value. Only instances with sufficient similarity are used to calculate the positional change relationship between two sensing data. This can further avoid matching errors caused by sensing defects and further improve the accuracy and robustness of vehicle-related mapping or positioning.

[0116] Furthermore, based on any of the above embodiments, another embodiment of the vehicle positioning method of this application is proposed. In this embodiment, reference is made to... Figure 5 The step of determining the position change parameter between the first sensing data and the second sensing data based on the multiple displacement estimates corresponding to the multiple matching groups includes:

[0117] Step S41: Determine the global displacement based on the multiple displacement estimates and their corresponding second weight values; the second weight values ​​are determined based on the corresponding correlation degree.

[0118] Specifically, the second weight value corresponding to the displacement estimate is determined based on the correlation degree of the matching group corresponding to the displacement estimate.

[0119] Different degrees of correlation correspond to different second weight values. The correlation degree and the second weight value are positively correlated. Specifically, the second weight value is calculated from the correlation degree using the Softmax function.

[0120] Based on this, the global displacement can be obtained by weighting multiple displacement estimates using multiple second weight values. Alternatively, the weighted average can be corrected according to preset values ​​and then used as the global displacement.

[0121] Step S42: Determine the position change parameters based on the global displacement.

[0122] Different global displacements correspond to different position change parameters.

[0123] In this embodiment, the position change parameter can be determined based on the global displacement and the predicted displacement, with the predicted displacement determined according to the mechanical odometer detection parameters of the vehicle. In other embodiments, the global displacement can also be directly used as the position change parameter. In other embodiments, a fixed correction value can be preset, and the position change parameter is obtained after correcting the global displacement according to the correction value.

[0124] Furthermore, in this embodiment, step S42 includes: determining the position change parameter based on the global displacement and its corresponding third weight value and the predicted displacement and its corresponding fourth weight value, wherein the predicted displacement is determined based on the mechanical odometer detection parameters of the vehicle.

[0125] The third and fourth weight values ​​can be preset fixed values, or they can be values ​​determined based on the actual situation of the first and second perception data.

[0126] Mechanical odometers include wheel speed meters and inertial navigation odometers. The predicted displacement can be calculated using the mileage data determined by the mechanical odometer.

[0127] The position change parameter can be obtained by weighting the global displacement and the predicted displacement according to the third and fourth weight values. Specifically, the position change parameter S can be calculated according to the following formula: S = p*S1 + q*S2, where S1 is the global displacement, S2 is the predicted displacement, p is the third weight value, and q is the fourth weight value.

[0128] In this embodiment, the global displacement is obtained by weighted averaging the displacement estimates corresponding to each matching group based on the correlation degree corresponding to different matching groups. The recognition error caused by perception defects in the position change parameters determined by the global displacement can be further reduced. The fusion of the first and second perception data based on the position change parameters in the subsequent mapping and localization process can be more accurate, thereby further improving the accuracy and robustness of vehicle mapping and localization. In particular, combining the global displacement and the predicted displacement corresponding to the mechanical odometer to determine the position change parameters is beneficial for using the displacement determined by the mechanical odometer to perform motion compensation on the position of feature instances, which is conducive to further improving the accuracy and robustness of mapping or localization.

[0129] Furthermore, in this embodiment, before step S41, the method further includes: determining the matching error between the first sensing data and the second sensing data; and determining the third weight value based on the matching error. Specifically, the matching error can be determined based on the difference between the first instance and the second instance in the matching group. Different matching errors correspond to different third weight values. The matching error and the third weight value can be negatively correlated; that is, the larger the error, the smaller the third weight value. Based on this, the third weight value can quantify the confidence level of the matching result. It can ensure that when a sensing error causes a large matching error between two sensing data points, the displacement detected by the mechanical odometer can be used more to compensate for the error in the position conversion parameters caused by the sensing defect, thus ensuring the accuracy of the position conversion parameters and further improving the accuracy and robustness of vehicle mapping and positioning.

[0130] Furthermore, in this embodiment, the process of determining the matching error between the first sensing data and the second sensing data is as follows: the first sensing data is translated according to the global displacement to obtain reference sensing data; the displacement deviation between each second instance and the corresponding third instance is determined to obtain multiple displacement deviations; the third instance is the instance in the reference sensing data after the first instance matched with the corresponding second instance is translated; the matching error is determined according to the multiple displacement deviations.

[0131] The reference sensing data here can be understood as a predicted virtual sensing data.

[0132] Specifically, the average of multiple displacement deviations can be determined as the matching error. Alternatively, a fifth weight value can be determined for each displacement deviation based on the correlation degree corresponding to each matching group. The weighted average of multiple displacement deviations according to these fifth weight values ​​can then be used as the matching error.

[0133] This method can accurately assess the matching error between two sensing data points, thereby ensuring that subsequent mapping or positioning can be accurately performed based on the matching error, and further improving the accuracy of vehicle-related mapping or positioning.

[0134] Furthermore, this embodiment of the invention also proposes a storage medium storing a vehicle positioning program, which, when executed by a processor, implements the relevant steps of any of the above embodiments of the vehicle positioning method.

[0135] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0136] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0137] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, 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 is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, vehicle, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0138] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A vehicle positioning method, characterized in that, The vehicle positioning method includes the following steps: Acquire first and second perception data of the area where the vehicle is located; Determine the correlation degree between the target attributes of each first instance in the first perception data and each second instance in the second perception data to obtain multiple correlation degrees; Based on the association matrix constructed from the multiple association degrees, a one-to-one matching relationship between multiple first instances and multiple second instances is determined, and multiple matching groups are obtained. Each matching group includes a first instance and a second instance that have a matching relationship. The position change parameters between the first sensing data and the second sensing data are determined based on the multiple displacement estimates corresponding to the multiple matching groups, and the vehicle's location in the area is determined based on the position change parameters. The target attribute includes multiple sub-attributes. The step of determining the correlation degree between the target attribute and each first instance in the first sensing data and each second instance in the second sensing data, and obtaining multiple correlation degrees, includes: Combine all the first instances with all the second instances in pairs to obtain multiple instance groups; The sub-association degree corresponding to the sub-attribute in each instance group is determined based on the first feature value and the second feature value corresponding to each sub-attribute in each instance group, wherein the first feature value is a feature of the first instance and the second feature value is a feature of the second instance; The association degree of the corresponding instance group is determined based on the multiple sub-associations in each instance group and their corresponding first weight values, thereby obtaining the multiple association degrees. Among them, the sub-associations corresponding to different sub-attributes have different first weight values.

2. The vehicle positioning method as described in claim 1, characterized in that, The target attribute includes multiple sub-attributes, including: category, color, area, centroid position, and principal component orientation.

3. The vehicle positioning method as described in claim 1, characterized in that, The step of determining the one-to-one matching relationship between multiple first instances and multiple second instances based on the association matrix constructed according to the multiple association degrees, and obtaining multiple matching groups, includes: The correlation matrix is ​​constructed based on all correlation degrees greater than a preset value; The correlation matrix is ​​solved according to a preset matching algorithm to obtain the solution result; The multiple matching groups are determined based on the solution results.

4. The vehicle positioning method as described in claim 1, characterized in that, The first sensing data and the second sensing data are two sensing data collected by the vehicle at adjacent time points; Alternatively, the first sensing data may be pre-stored map data of the area, and the second sensing data may be sensing data currently collected by the vehicle.

5. The vehicle positioning method as described in any one of claims 1 to 4, characterized in that, The step of determining the position change parameter between the first sensing data and the second sensing data based on the multiple displacement estimates corresponding to the multiple matching groups includes: The global displacement is calculated based on the multiple displacement estimates and their corresponding second weight values; the second weight values ​​are determined based on the corresponding correlation degree. The position change parameters are determined based on the global displacement.

6. The vehicle positioning method as described in claim 5, characterized in that, The step of determining the position change parameter based on the global displacement includes: The position change parameters are determined based on the global displacement and its corresponding third weight value, and the predicted displacement and its corresponding fourth weight value. The predicted displacement is determined based on the mechanical odometer detection parameters of the vehicle.

7. The vehicle positioning method as described in claim 6, characterized in that, Before the step of determining the position change parameter based on the global displacement and its corresponding third weight value and the predicted displacement and its corresponding fourth weight value, the method further includes: Determine the matching error between the first sensing data and the second sensing data; The third weight value is determined based on the matching error.

8. The vehicle positioning method as described in claim 7, characterized in that, The step of determining the matching error between the first sensed data and the second sensed data includes: The first sensing data is translated according to the global displacement to obtain reference sensing data; The displacement deviation between each second instance and its corresponding third instance is determined to obtain multiple displacement deviations; the third instance is the instance in the reference sensing data after the first instance matched with the corresponding second instance has been translated. The matching error is determined based on the plurality of displacement deviations.

9. A vehicle, characterized in that, The vehicle includes: a memory, a processor, and a vehicle positioning program stored in the memory and executable on the processor, wherein the vehicle positioning program, when executed by the processor, implements the steps of the vehicle positioning method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium stores a vehicle positioning program, which, when executed by a processor, implements the steps of the vehicle positioning method as described in any one of claims 1 to 8.

Images