Vehicle positioning method, apparatus, device, and storage medium

Abstract

Embodiments of the present disclosure disclose a vehicle positioning method and device, electronic equipment and storage medium, the method comprising: obtaining an absolute positioning pose of a target vehicle at a current time; determining whether to enable a relative positioning pose of the target vehicle at the current time according to the absolute positioning pose; if yes, determining an actual gap between the absolute positioning pose and the relative positioning pose, and in a case where the actual gap is greater than a preset threshold, obtaining a first score of each first sampling trajectory in an absolute positioning trajectory comprising at least m absolute positioning poses at historical time points and a second score of each second sampling trajectory in a relative positioning trajectory comprising at least m relative positioning poses according to a pre-trained time sequence trajectory evaluation model; and determining a positioning result of the target vehicle according to the first scores and the second scores. The present disclosure improves the positioning accuracy of the target vehicle and the safety of driving.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of automatic driving, and particularly relates to a vehicle positioning method and device, equipment and a storage medium. BACKGROUND

[0002] Positioning is a core component of an intelligent driving system. Current mainstream positioning methods can be divided into absolute positioning and relative positioning. Existing absolute positioning technologies generally use a global positioning map. The positioning data collected in real time is matched with the map obtained in advance to obtain a positioning result. However, the real environment is often changing, such as road construction and temporary parking on the roadside. The map needs a long period of time to update, and thus the map gradually becomes outdated, and the positioning accuracy is poor.

[0003] In addition, some existing technologies use a layered mapping method to distinguish static elements and dynamic targets in the map. The dynamic object part is filtered out through target detection during real-time positioning. This method may result in too little element data in the map and thus positioning failure. Meanwhile, the static elements also change, which results in poor positioning accuracy. SUMMARY

[0004] To solve the above technical problems or at least partially solve the above technical problems, the embodiments of the present disclosure provide a vehicle positioning method, device, equipment and storage medium, which improve the positioning accuracy of a target vehicle and the safety of driving.

[0005] In a first aspect, the embodiments of the present disclosure provide a vehicle positioning method, which comprises:

[0006] obtaining an absolute positioning pose of a target vehicle at a current time;

[0007] determining whether to enable a relative positioning pose of the target vehicle at the current time according to the absolute positioning pose;

[0008] if yes, determining an actual distance between the absolute positioning pose and the relative positioning pose, and in a case where the actual distance is greater than a preset threshold, obtaining a first score of each first sampling trajectory in an absolute positioning trajectory including absolute positioning poses of m historical times, and a second score of each second sampling trajectory in a relative positioning trajectory including relative positioning poses, according to a pre-trained time sequence trajectory evaluation model;

[0009] determining a positioning result of the target vehicle according to the first scores and the second scores.

[0010] In a second aspect, the embodiments of the present disclosure further provide a vehicle positioning device, comprising: a pose obtaining module configured to obtain an absolute positioning pose of a target vehicle at a current time point;

[0011] a judging module configured to judge whether to enable a relative positioning pose of the target vehicle at the current time point according to the absolute positioning pose;

[0012] an evaluation result determining module configured to, when the relative positioning pose of the target vehicle at the current time point is enabled, determine an actual gap between the absolute positioning pose and the relative positioning pose; and when the actual gap is greater than a preset threshold, input, according to absolute positioning trajectories of each historical time point associated with the current time point and including at least m absolute positioning poses and relative positioning trajectories of each historical time point associated with the current time point and including at least m relative positioning poses, into a pre-trained time sequence trajectory evaluation model to obtain a first score of each first sampling trajectory in the absolute positioning trajectories and a second score of each second sampling trajectory in the relative positioning trajectories;

[0013] a positioning result determining module configured to determine a positioning result of the target vehicle according to the first scores and the second scores.

[0014] In a third aspect, the embodiments of the present disclosure further provide an electronic device, comprising: one or more processors; a storage device configured to store one or more programs; and when the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle positioning method as described above.

[0015] In a fourth aspect, the embodiments of the present disclosure further provide a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vehicle positioning method as described above.

[0016] The vehicle positioning method provided by the embodiments of the present disclosure triggers relative positioning through a preset prior region or a first mean score, avoids low positioning accuracy caused by the prior art relying on absolute positioning pose alone, determines a first score of a first sampling trajectory of an absolute positioning trajectory and a second score of a second sampling trajectory of a relative positioning trajectory through a time sequence trajectory evaluation model, and determines a final result in combination with scores of absolute positioning and relative positioning, so as to solve the problem of poor positioning accuracy in the prior art. In addition, the time sequence trajectory evaluation model is used to predict and obtain a positioning result according to input trajectories containing multiple poses, so as to realize vehicle positioning in combination with multiple poses, and improve positioning accuracy of the target vehicle and safety of driving. BRIEF DESCRIPTION OF DRAWINGS

[0017] The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings. The same or similar components have the same or similar reference labels. It should be understood that the drawings are not necessarily to scale, with emphasis instead being placed upon illustrating the principles of the embodiments of the present disclosure.

[0018] Figure 1 A flowchart of a vehicle positioning method provided by an embodiment of the present disclosure.

[0019] Figure 2 A structural diagram of a timing trajectory evaluation model provided by the embodiment.

[0020] Figure 3 A structural diagram of obtaining a positioning result provided by the embodiment.

[0021] Figure 4 A flowchart of a vehicle positioning method provided by the embodiment.

[0022] Figure 5 A structural diagram of a vehicle positioning device provided by an embodiment of the present disclosure.

[0023] Figure 6 A structural diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] Embodiments of the present disclosure will be described in more detail by referring to the drawings. Although certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein, but rather these embodiments are provided so as to more thoroughly and completely understand the present disclosure. It is understood that the drawings and embodiments of the present disclosure are for exemplary purposes only and are not intended to limit the scope of protection of the present disclosure.

[0025] It should be noted that the terms “first”, “second”, and the like in the present disclosure are merely used to distinguish different devices, modules, or units, and do not imply the order or interdependence of the functions performed by these devices, modules, or units.

[0026] The names of messages or information exchanged between the devices in the embodiments of the present disclosure are merely used for illustrative purposes, and are not intended to limit the scope of the messages or information.

[0027] The embodiments of the present disclosure provide a vehicle positioning method, Figure 1 A flowchart of a vehicle positioning method provided by an embodiment of the present disclosure. The method can be performed by a vehicle positioning device, which can be implemented in software and / or hardware, and can be configured in an electronic device. As shown in FIG. 1, the method includes the following steps. Figure 1As shown, the method can specifically include the following steps:

[0028] S110, acquiring an absolute positioning pose of the target vehicle at the current time.

[0029] The absolute positioning pose of the target vehicle at the current time is mainly the absolute positioning pose collected by the sensor on the target vehicle at the current time, wherein the sensor can be a camera, a laser radar, an inertial measurement unit, a wheel speed meter, a GPS, etc.; and the absolute positioning pose can be a pose obtained by fusing one or more sensors such as a camera, a laser radar, an inertial measurement unit, a wheel speed meter, a GPS, etc.

[0030] S120, determining whether to enable the relative positioning pose of the target vehicle at the current time according to the absolute positioning pose.

[0031] Specifically, in the embodiment, the relative positioning pose of the target vehicle at the current time can be determined to be enabled when the positioning effect of the absolute positioning pose is poor.

[0032] In an embodiment, determining whether to enable the relative positioning pose of the target vehicle at the current time according to the absolute positioning pose includes:

[0033] If the absolute positioning pose is located in a preset prior region, the relative positioning pose of the target vehicle at the current time is enabled.

[0034] It can be understood that the preset prior region is a region determined by a person, for example, a region with poor absolute positioning pose effect obtained by prior statistics, such as a viaduct or a tunnel. Specifically, whether the target vehicle enters a preset prior region such as a viaduct or a tunnel can be determined according to the absolute positioning pose. By pre-setting the preset prior region, the region with inaccurate absolute positioning pose can be determined, and the relative positioning is triggered, so as to improve the accuracy of the vehicle positioning result by using the relative positioning pose.

[0035] In another embodiment, determining whether to enable the relative positioning pose of the target vehicle at the current time according to the absolute positioning pose includes:

[0036] If the absolute positioning pose is not located in the preset prior region, and the first mean score of the absolute positioning pose determined based on the time sequence trajectory evaluation model is less than a preset score threshold, the relative positioning pose of the target vehicle at the current time is enabled.

[0037] Specifically, when the absolute positioning pose is not located in the prior region such as a viaduct or a tunnel, the absolute positioning trajectory including at least m absolute positioning poses can be input into the time sequence trajectory evaluation model, so as to obtain the first score of each first sampling trajectory by the time sequence trajectory evaluation model.

[0038] It can be understood that the absolute positioning trajectory of the absolute positioning pose including at least m historical moments, for example, the absolute positioning trajectory includes P absolute positioning poses, P>m+g(g>6), the absolute positioning trajectory can be composed of the first to mth absolute positioning poses, the absolute positioning trajectory can be composed of the third to Pth absolute positioning poses, and the absolute positioning trajectory can be composed of the third to m+6th absolute positioning poses, as long as at least m absolute positioning poses constitute the absolute positioning trajectory, the number of positioning poses in the absolute positioning trajectory is not limited in the application.

[0039] For example, for the absolute positioning trajectory including at least m absolute positioning poses, assuming that the absolute positioning trajectory includes P absolute positioning poses, P=m+g (assuming g>6), the first first sampling trajectory can be composed of the first to mth absolute positioning poses, the second first sampling trajectory can be composed of the second to m+1th absolute positioning poses, the third first sampling trajectory can be composed of the third to m+2th absolute positioning poses, and the fourth first sampling trajectory can be composed of the fourth to m+3th absolute positioning poses. In this way, P-m+1 first sampling trajectories are obtained.

[0040] On the basis of the above, the absolute positioning trajectory including P absolute positioning poses is input into the pre-trained time sequence trajectory evaluation model to obtain the first scores of the first sampling trajectories in the absolute positioning trajectory. Specifically, the first score of the first first sampling trajectory, the first score of the second first sampling trajectory, the first score of the third first sampling trajectory, and the first score of the fourth first sampling trajectory can be obtained. In this way, the first scores of P-m+1 first sampling trajectories are obtained. When the number of first scores is not less than the number threshold K of first scores, K is not less than 3, at this time, the first scores of all first sampling trajectories can be counted by using the continuous frame score counter, the mean value of all first scores is determined, and the mean value of all first scores is taken as the first mean score. The number threshold K of first scores is determined according to actual conditions.

[0041] Further, it is judged whether the first mean score is less than a preset score threshold. When it is determined that the first mean score is less than the preset score threshold, the relative positioning pose of the target vehicle at the current moment is enabled. In this embodiment, the first scores are obtained by using the time sequence trajectory evaluation model, the first mean score is calculated based on the first scores, and when it is confirmed that the first mean score is less than the preset score threshold, the relative positioning is triggered, the relative positioning pose and the absolute positioning pose are used to determine the final positioning result, and the positioning accuracy of the vehicle is improved.

[0042] S130, if yes, determining an actual gap between the absolute positioning pose and the relative positioning pose, in a case that the actual gap is greater than a preset threshold, obtaining a first score of each first sampling trajectory in the absolute positioning trajectory and a second score of each second sampling trajectory in the relative positioning trajectory according to the absolute positioning trajectory including at least m absolute positioning poses, the relative positioning trajectory including at least m relative positioning poses, and a pre-trained time sequence trajectory evaluation model.

[0043] It can be understood that if yes indicates enabling the relative positioning pose of the target vehicle at the current time, in the case of starting the relative positioning pose, when the poses of the absolute positioning trajectory and the relative positioning trajectory exceed m, the absolute positioning trajectory and the relative positioning trajectory are input into the pre-trained time sequence trajectory evaluation model to obtain the first score of each first sampling trajectory in the absolute positioning trajectory and the second score of each second sampling trajectory in the relative positioning trajectory, when the poses of the absolute positioning trajectory or the relative positioning trajectory are less than m, the absolute positioning trajectory or the relative positioning trajectory is not input into the pre-trained time sequence trajectory evaluation model, and the first score or the second score is not needed to be obtained.

[0044] In an embodiment, the determining the actual gap between the absolute positioning pose and the relative positioning pose comprises:

[0045] determining a pose alignment matrix according to the absolute positioning pose and the relative positioning pose, and converting the relative positioning pose into a first absolute positioning pose according to the pose alignment matrix;

[0046] determining an actual gap between the first absolute positioning pose and the absolute positioning pose.

[0047] The pose alignment matrix can be a matrix used for converting the relative positioning pose into the first absolute positioning pose, and the first absolute positioning pose is a pose of the relative positioning pose converted to a coordinate of the absolute positioning pose.

[0048] Specifically, the pose alignment matrix can be calculated in advance through the absolute positioning poses and the relative positioning poses at each historical time, further, the relative positioning pose is converted into the first absolute positioning pose through the pose alignment matrix, and then the actual gap between the first absolute positioning pose and the absolute positioning pose is calculated, wherein the actual gap is described by the Euclidean distance difference between the absolute positioning pose and the relative positioning pose.

[0049] In the above embodiment, the relative positioning pose is converted into the first absolute positioning pose through the pose alignment matrix, and then the positioning result is evaluated according to the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose, thereby improving the accuracy of the vehicle positioning result.

[0050] In an optional embodiment, the determining the pose alignment matrix according to the absolute positioning pose and the relative positioning pose comprises:

[0051] Taking the absolute positioning pose and the relative positioning pose of the same frame as a positioning pose pair, when the positioning pose pair is not less than N, determining an initial pose alignment matrix according to the positioning pose pair;

[0052] Optimizing the initial pose alignment matrix according to the absolute positioning pose and the relative positioning pose until the Euclidean distance difference between the absolute positioning pose and the relative positioning pose of the current frame is less than a first threshold value, and taking the optimized initial pose alignment matrix of the current frame as the pose alignment matrix.

[0053] For example, based on the absolute positioning pose and the relative positioning pose of the same frame collected by one or more sensors such as a camera, a laser radar, an inertial measurement unit, a wheel speed meter, a GPS, etc., the collected absolute positioning pose and the relative positioning pose are taken as a positioning pose pair, when the positioning pose pair is not less than N, the positioning pose pair is optimized by using the least square method to obtain an initial pose alignment matrix; the relative positioning pose is converted according to the initial pose alignment matrix to obtain a first absolute positioning pose; it is judged whether the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is less than a first threshold value, if not, the initial pose alignment matrix is continuously optimized according to the positioning pose pair until the Euclidean distance difference between the absolute positioning pose and the first absolute positioning pose of the current frame is less than the first threshold value, and the optimized initial pose alignment matrix of the current frame is taken as the pose alignment matrix.

[0054] If the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is less than the first threshold value, the initial pose alignment matrix is taken as the pose alignment matrix. The first threshold value is set artificially according to actual requirements, and the first threshold value range is preferably between 0.1 and 0.9, for example, the first threshold value is set to 0.173 (equal to ), which means that the difference between the first absolute positioning pose and the absolute positioning pose in the x, y, and z directions is 0.1 m.

[0055] After the pose alignment matrix is determined, a new relative positioning pose is obtained, the new relative positioning pose is converted into a first absolute positioning pose according to the pose alignment matrix, the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose, i.e. the actual difference, is determined, and it is judged whether the actual difference is greater than a preset threshold value.

[0056] In the above embodiment, the relative positioning pose is converted into the first absolute positioning pose by the pose alignment matrix, and then the positioning result is evaluated according to the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose, thereby improving the accuracy of the vehicle positioning result.

[0057] It should be noted that the process of determining each second sampling trajectory and the second score according to the relative positioning trajectory including at least m relative positioning poses can refer to the process of determining each first sampling trajectory and the first score according to the absolute positioning trajectory, which will not be described herein.

[0058] In an embodiment, when the actual gap is greater than the preset threshold, the first score of each first sampling trajectory in the absolute positioning trajectory and the second score of each second sampling trajectory in the relative positioning trajectory are obtained according to the absolute positioning trajectory including at least m absolute positioning poses, the relative positioning trajectory including at least m relative positioning poses, and the pre-trained time sequence trajectory evaluation model, including:

[0059] The absolute positioning trajectory and the relative positioning trajectory are input into the time sequence trajectory evaluation model, the first sampling trajectory and the second sampling trajectory are determined by using a sliding window in the pre-processing module in the time sequence trajectory evaluation model, and the first sampling trajectory and the second sampling trajectory are normalized by the pre-processing module.

[0060] The normalized results are sequentially input into the backbone network and the specific operator in the time sequence trajectory evaluation model to obtain the first score of each first sampling trajectory and the second score of each second sampling trajectory.

[0061] Specifically, the absolute positioning trajectory and the relative positioning trajectory are input into the pre-processing module, the first sampling trajectory including m absolute positioning poses and the second sampling trajectory including m relative positioning poses are determined by using a sliding window in the pre-processing module, each sampling trajectory is subjected to timestamp normalization processing and pose relativization processing to obtain a normalized result in the form of a positioning trajectory matrix, and the normalized result is input into the backbone network and the specific operator in the time sequence trajectory evaluation model to obtain the first score of each first sampling trajectory and the second score of each second sampling trajectory.

[0062] In the above embodiment, the timestamp normalization processing and the pose relativization processing of each sampling trajectory by the pre-processing module are to enable the network to learn better model parameters.

[0063] In an embodiment, the normalization processing of each first sampling trajectory and each second sampling trajectory by the pre-processing module includes:

[0064] For each first sampling trajectory and each second sampling trajectory, the timestamps of other positioning poses in the sampling trajectory are normalized based on the timestamp of the first positioning pose in the sampling trajectory.

[0065] Specifically, the timestamps of other positioning poses in the first sampling trajectory and the second sampling trajectory are converted into new relative timestamps through normalization processing, based on the timestamp of the first absolute positioning pose in the first sampling trajectory and the timestamp of the first relative positioning pose in the second sampling trajectory.

[0066] The embodiment normalizes the timestamps of the sampling trajectories, so that the network learns better model parameters.

[0067] In an embodiment, the normalization processing of the first sampling trajectories and the second sampling trajectories by the preprocessing module includes:

[0068] For each first sampling trajectory and each second sampling trajectory, the other positioning poses in the sampling trajectory are normalized based on the first positioning pose in the sampling trajectory.

[0069] Specifically, the positions (x, y, z) and attitudes (roll, pitch, yaw) of other positioning poses in the first sampling trajectory and the second sampling trajectory are respectively processed, based on the positions (x, y, z) and attitudes (roll, pitch, yaw) of the first absolute positioning pose in the first sampling trajectory and the first relative positioning pose in the second sampling trajectory.

[0070] The embodiment relativizes the positions and attitudes of the sampling trajectories, so that the network learns better model parameters.

[0071] Exemplarily, Figure 2 A structural schematic diagram of a time sequence trajectory evaluation model provided by the embodiment is shown in the figure, Figure 2 As shown in the figure, the time sequence trajectory evaluation model includes a preprocessing module, a backbone network, and a specific operator, wherein the specific operator includes an AvgPool operator and a Sigmoid operator.

[0072] It can be understood that the embodiment provides a trained time sequence trajectory evaluation model; the training process of the time sequence trajectory evaluation model is as follows: first, data is collected, including absolute positioning trajectories, relative positioning trajectories (converted into first absolute positioning trajectories through a pose alignment matrix), and true value trajectories (such as GPS); a labeling personnel manually judges the pros and cons of the absolute positioning trajectories and the relative positioning trajectories, the judgment criteria include the smoothness of the trajectories, the degree of coincidence with the vehicle kinematics model, and the degree of coincidence with the true value trajectories, etc., and a score between 0 and 1 (0 is the worst) is given to the pros and cons of each trajectory; when the manual score and the score of the time sequence trajectory evaluation model are basically consistent, or less than a certain threshold, it is considered that the time sequence trajectory evaluation model has been trained.

[0073] The process of inputting the absolute positioning trajectory and the relative positioning trajectory into the trained time sequence trajectory evaluation model respectively is:

[0074] The absolute positioning trajectory according to the absolute positioning pose including m historical time points and the relative positioning trajectory including m relative positioning poses are input into a pre-processing module, the pre-processing module determines each first sampling trajectory and each second sampling trajectory by using a sliding window, and the process of normalizing each first sampling trajectory and each second sampling trajectory by the pre-processing module is:

[0075] Each first sampling trajectory and each second sampling trajectory composed of m positioning poses are determined by using a sliding window capable of storing m positioning poses, wherein the positioning pose includes an absolute positioning pose or a relative positioning pose, and the positioning pose includes information such as a timestamp, a position (x, y, z), a roll angle, a pitch angle, and a yaw angle.

[0076] The time (T n ) of other positioning poses in the first sampling trajectory and the second sampling trajectory is converted into a new relative time representation (denoted as T n ′) based on the timestamp of the first absolute positioning pose in the first sampling trajectory and the timestamp of the first relative positioning pose in the second sampling trajectory.

[0077]

[0078]

[0079]

[0080] The position (x, y, z) of the first absolute positioning pose in the first sampling trajectory and the position (x, y, z) of the first relative positioning pose in the second sampling trajectory, also known as the translation vector (denoted as t1, dimension 3x1), are used as the reference to relatively process the positions of other positioning poses in the first sampling trajectory and the second sampling trajectory, and the relatively processed positions are obtained, and the position expression after relative processing is:

[0081]

[0082] The attitude (roll, pitch, yaw) of the first absolute positioning pose in the first sampling trajectory and the attitude (roll, pitch, yaw) of the first relative positioning pose in the second sampling trajectory, also referred to as a rotation vector (denoted as q1, with a dimension of 3x1), are taken as the reference, and the attitudes of other positioning poses in the first sampling trajectory and the second sampling trajectory are subjected to relative processing to obtain the relative processed attitude, and the expression of the relative processed attitude is:

[0083]

[0084] wherein, R n denotes q n The corresponding rotation matrix, and Q(·) denotes the transformation from the rotation matrix to the Euler angle.

[0085] The positioning trajectory matrix Mx7x1xm obtained after the pre-processing is in the form of a positioning trajectory matrix, and the normalization result is expressed in the form of a positioning trajectory matrix. The normalization result includes an absolute positioning trajectory matrix and a relative positioning trajectory matrix. M=1 indicates that only the absolute positioning trajectory is included, M=2 indicates that the absolute positioning trajectory and the relative positioning trajectory are included, 7 is the timestamp, the position (x, y, z), the roll angle roll, the pitch angle pitch, and the yaw angle yaw, and m is the number of positioning poses in the sliding window.

[0086] The backbone network is formed by one-dimensional convolution, and the normalized result (the absolute positioning trajectory matrix and the relative positioning trajectory matrix) is input into the backbone network, which can output a tensor of Mx1x1xm.

[0087] The AvgPool operator is used to average the features of each channel, and the tensor of Mx1x1xm is input into the AvgPool operator, and the output is a tensor of Mx1x1x1.

[0088] Then, the tensor of Mx1x1x1 is input into the Sigmoid operator, and the output is the score of M positioning trajectories in the interval of 0-1. Specifically, when the relative positioning pose is not enabled, that is, M=1, the first score of each first sampling trajectory in the absolute positioning trajectory is output; when the relative positioning pose is enabled, that is, M=2, the first score of each first sampling trajectory in the absolute positioning trajectory and the second score of each second sampling trajectory in the relative positioning trajectory are output.

[0089] The timing trajectory evaluation model provided by the embodiment of the present disclosure improves the robustness of the intelligent driving positioning system through the evaluation of the relative positioning trajectory and the absolute positioning trajectory. The relative positioning pose and the absolute positioning pose are decoupled and do not affect each other, and the intelligent driving positioning system is more flexible.

[0090] S140, determining the positioning result of the target vehicle according to each first score and each second score.

[0091] The first score is used to describe the track rationality of the corresponding first sampling track, and the second score is used to describe the track rationality of the corresponding second sampling track. Specifically, the accuracy of the absolute positioning track can be judged in combination with each first score, and the accuracy of the relative positioning track can be judged in combination with each second score.

[0092] For example, the maximum value or the average value in each first score and the maximum value or the average value in each second score can be determined. If the maximum value or the average value in each first score is greater than the maximum value or the average value in each second score, the positioning result is determined according to the absolute positioning track. If the maximum value or the average value in each second score is less than the maximum value or the average value in each second score, the positioning result is determined according to the relative positioning track.

[0093] In an embodiment, the positioning result of the target vehicle is determined according to each first score and each second score, comprising:

[0094] The number of first scores determined based on the absolute positioning track and the number of second scores determined based on the relative positioning track are determined by a continuous frame score calculator, and it is judged whether the number of first scores or second scores is greater than K.

[0095] In the case that the number of first scores is less than K or the number of second scores is less than K, the last frame absolute positioning pose in the absolute positioning track is taken as the positioning result of the target vehicle.

[0096] For example, Figure 3 A structure diagram for obtaining a positioning result is provided for the embodiment. As Figure 3 shown, on the basis of the above embodiment, the absolute positioning track and the relative positioning track are input into a time sequence track evaluation model, each first score and each second score output by the time sequence track evaluation model are counted by a continuous frame score calculator, and in the case that the number of first scores is less than K or the number of second scores is less than K, the last frame absolute positioning pose in the absolute positioning track is taken as the positioning result of the target vehicle.

[0097] Specifically, the quantity threshold K of the first scores is determined according to actual conditions, K is preferably not less than 3, the absolute positioning trajectory including at least q absolute positioning poses and the relative positioning trajectory including at least q relative positioning poses are input to the time sequence trajectory evaluation model, wherein q < m+2, when q is an integer, q = m+1, the first sampling trajectory can be composed of the first to m absolute positioning poses, the second first sampling trajectory can be composed of the second to m+1 absolute positioning poses, and 2 first sampling trajectories are inferred, the time sequence trajectory evaluation model outputs 2 first scores and 2 second scores through the continuous frame score calculator, when q < m+2 and K is not less than 4, if the number of first scores is less than 4, the qth frame in the absolute positioning trajectory is taken as the positioning result of the target vehicle. When q < m+2 and K is not less than 4, if the number of second scores is less than 4, the qth frame in the absolute positioning trajectory is taken as the positioning result of the target vehicle.

[0098] In an embodiment, in the case that the number of first scores is not less than K and the number of second scores is not less than K, the first mean score of each first score and the second mean score of each second score are determined, and the last frame positioning pose in the positioning trajectory with the highest mean score between the first mean score and the second mean score is taken as the positioning result of the target vehicle, wherein the positioning trajectory includes the absolute positioning trajectory or the relative positioning trajectory.

[0099] Continuing to refer to Figure 3 As shown in the above embodiment, the number of first sampling trajectories is determined according to the absolute positioning trajectory, when the absolute positioning trajectory includes P absolute positioning poses, P > m+g (g > 6), the first first sampling trajectory can be composed of the first to m absolute positioning poses, the second first sampling trajectory can be composed of the second to m+1 absolute positioning poses, the third first sampling trajectory can be composed of the third to m+3 absolute positioning poses, the fourth first sampling trajectory can be composed of the fourth to m+4 absolute positioning poses, and so on, and P-m+1 first sampling trajectories are obtained. After obtaining P-m+1 first sampling trajectories, the first scores of P-m+1 first sampling trajectories can be obtained, when the number of first scores is not less than the quantity threshold K of first scores, K is not less than 3, at this time, the first scores of all first sampling trajectories can be counted by using the continuous frame score calculator, the mean value of all first scores is determined, and the mean value of all first scores is taken as the first mean score.

[0100] Specifically, the absolute positioning trajectory including at least P absolute positioning poses and the relative positioning trajectory including at least P relative positioning poses are input into the time sequence trajectory evaluation model, P-m+1 first scores and P-m+1 second scores output by the time sequence trajectory evaluation model are counted by the continuous frame score counter, when P-m+1>3 and K is not less than 3, the number of the first scores is not less than 3 and the number of the second scores is not less than 3, the first mean score of the P-m+1 first scores and the second mean score of the P-m+1 second scores are determined, and the last frame positioning pose in the positioning trajectory with the highest mean score between the first mean score and the second mean score is determined as the positioning result of the target vehicle, wherein the positioning trajectory includes the absolute positioning trajectory or the relative positioning trajectory.

[0101] It should be noted that the process of obtaining the first mean score is consistent with the process of obtaining the first mean score of the absolute positioning pose when the first mean score is less than the preset score threshold. The second score is obtained in the same way as the first score, and the second mean score is obtained in the same way as the first mean score, which will not be described in detail herein.

[0102] The vehicle positioning method provided by the embodiments of the present disclosure triggers relative positioning through a preset prior region or a first mean score, thereby improving the accuracy of the positioning pose; the relative positioning trajectory and the absolute positioning trajectory are input into the time sequence trajectory evaluation model, the scores of the relative positioning and the absolute positioning are compared, the last frame positioning pose of the positioning trajectory with the higher score is determined as the positioning result of the target vehicle, thereby improving the positioning accuracy and the safety of vehicle driving.

[0103] In an embodiment, when the actual gap is greater than the preset threshold, if the number of absolute positioning poses in the absolute positioning trajectory is less than m, or if the number of relative positioning poses in the relative positioning trajectory is less than m, the last frame absolute positioning pose in the absolute positioning trajectory is determined as the positioning result of the target vehicle.

[0104] It can be understood that, on the basis of the above-mentioned embodiments, when the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is greater than the preset threshold, when the number of absolute positioning poses in the absolute positioning trajectory is less than m, or when the number of relative positioning poses in the relative positioning trajectory is less than m, the time sequence trajectory evaluation model is not entered, and the last frame absolute positioning pose in the absolute positioning trajectory is directly determined as the positioning result of the target vehicle.

[0105] It should be noted that when the absolute positioning track includes less than m absolute positioning poses, the absolute positioning pose of the last frame in the absolute positioning track is taken as the positioning result of the target vehicle; for example, the absolute positioning track includes t absolute positioning poses, m>t, at this time, the absolute positioning pose of the tth frame in the absolute positioning track is taken as the positioning result of the target vehicle.

[0106] In the case where the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is greater than the preset threshold, when the relative positioning track includes less than m relative positioning poses, the absolute positioning pose of the last frame in the absolute positioning track is taken as the positioning result of the target vehicle; for example, the relative positioning track includes t relative positioning poses, m>t, at this time, the absolute positioning pose of the last frame in the absolute positioning track is taken as the positioning result of the target vehicle.

[0107] In an embodiment, in the case where the actual difference is not greater than the preset threshold, the absolute positioning pose of the last frame in the absolute positioning track is taken as the positioning result of the target vehicle.

[0108] It can be understood that, on the basis of the above-mentioned embodiments, in the case where the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is not greater than the preset threshold, the absolute positioning pose of the last frame in the absolute positioning track is directly taken as the positioning result of the target vehicle.

[0109] It should be noted that in the case where the Euclidean distance difference between the first absolute positioning pose and the absolute positioning pose is not greater than the preset threshold, when the absolute positioning track includes more than m absolute positioning poses, the absolute positioning pose of the last frame in the absolute positioning track is taken as the positioning result of the target vehicle; for example, the absolute positioning track includes s absolute positioning poses, s>m, at this time, the absolute positioning pose of the s th frame in the absolute positioning track is taken as the positioning result of the target vehicle.

[0110] In summary, on the basis of the technical solutions of the above-mentioned embodiments, Figure 4 A vehicle positioning method flowchart is provided for the embodiment, as shown in Figure 4As shown: acquire absolute positioning pose, judge whether to enter the preset prior region, if yes, enable relative positioning pose; if no, judge whether the number of absolute positioning poses is greater than m, if the number of absolute positioning poses is less than or equal to m, the absolute positioning pose of the last frame is taken as the positioning result; if the number of absolute positioning poses is greater than m, the absolute positioning trajectory including at least m absolute positioning poses is input into the time sequence trajectory evaluation model, and then it is judged whether the number of first scores is greater than K, if not, the absolute positioning pose of the last frame is taken as the positioning result; if greater than K, it is judged whether the first average score is less than the preset score threshold, if the first average score is greater than or equal to the preset score threshold, the absolute positioning pose of the last frame is taken as the positioning result, if the first average score is less than the preset score threshold, the relative positioning pose is enabled. For the case of starting the relative positioning pose, the absolute positioning pose and the relative positioning pose are input into the pose alignment matrix, and then it is judged whether the actual gap between the relative positioning pose converted into the first absolute positioning pose by the pose alignment matrix and the absolute positioning pose is greater than the preset threshold, if not, the absolute positioning pose of the last frame is taken as the positioning result; if greater than the preset threshold, it is further judged whether the number of relative positioning poses and the number of absolute positioning poses are greater than m, if the number of absolute positioning poses is not greater than m or the number of relative positioning poses is not greater than m, the absolute positioning pose of the last frame is taken as the positioning result; if the number of relative positioning poses is greater than m and the number of relative positioning poses is also greater than m, the absolute positioning trajectory and the relative positioning trajectory are simultaneously input into the time sequence trajectory evaluation model, the first score and the second score are output, and it is further judged whether the number of first scores and the number of second scores is not less than K, if the number of first scores is not less than K and the number of second scores is not less than K, the last frame positioning pose in the positioning trajectory with the highest average score between the first average score and the second average score is taken as the positioning result of the target vehicle, wherein the positioning trajectory includes the absolute positioning trajectory or the relative positioning trajectory; if the number of first scores is less than K or the number of second scores is less than K, the last frame absolute positioning pose in the absolute positioning trajectory is taken as the positioning result of the target vehicle.

[0111] Figure 5 A structural schematic diagram of a vehicle positioning device in an embodiment of the present disclosure is shown. As shown in the figure, the device includes a pose acquisition module 210, a judgment module 220, an evaluation result determination module 230, and a positioning result determination module 240. Figure 5

[0112] The pose acquisition module 210 is configured to acquire an absolute positioning pose of a target vehicle at a current time.

[0113] The judgment module 220 is configured to judge whether to enable a relative positioning pose of the target vehicle at the current time according to the absolute positioning pose.​

[0114] The evaluation result determination module 230 is configured to determine an actual difference between the absolute positioning pose and the relative positioning pose when enabling the relative positioning pose of the target vehicle at the current time, and determine, when the actual difference is greater than a preset threshold, a first score of each first sampling trajectory in the absolute positioning trajectory and a second score of each second sampling trajectory in the relative positioning trajectory by inputting the absolute positioning trajectory and the relative positioning trajectory associated with each historical time at the current time to a pre-trained time sequence trajectory evaluation model.

[0115] The positioning result determination module 240 is configured to determine a positioning result of the target vehicle according to the first scores and the second scores.

[0116] In an embodiment, the positioning result determination module 240 is further configured to determine, by a continuous frame score calculator, a number of the first scores determined based on the absolute positioning trajectory and a number of the second scores determined based on the relative positioning trajectory, and determine whether the number of the first scores or the number of the second scores is greater than K.

[0117] In a case where the number of the first scores is less than K or the number of the second scores is less than K, the last frame absolute positioning pose in the absolute positioning trajectory is determined as the positioning result of the target vehicle.

[0118] In an embodiment, the positioning result determination module 240 is further configured to, in a case where the number of the first scores is not less than K and the number of the second scores is not less than K, determine a first average score of the first scores and a second average score of the second scores, and determine, as the positioning result of the target vehicle, a last frame positioning pose in a positioning trajectory with the highest average score between the first average score and the second average score, wherein the positioning trajectory includes the absolute positioning trajectory or the relative positioning trajectory.

[0119] In an embodiment, the positioning result determination module 240 is further configured to, in a case where the actual difference is not greater than the preset threshold, determine the last frame absolute positioning pose in the absolute positioning trajectory as the positioning result of the target vehicle.

[0120] In an embodiment, the positioning result determination module 240 is further configured to, in a case where the actual difference is greater than the preset threshold, determine, as the positioning result of the target vehicle, the last frame absolute positioning pose in the absolute positioning trajectory, if the number of the absolute positioning poses in the absolute positioning trajectory is less than m, or if the number of the relative positioning poses in the relative positioning trajectory is less than m.

[0121] In an embodiment, the evaluation result determination module 230 is further configured to input the absolute positioning trajectory and the relative positioning trajectory into the timing trajectory evaluation model, determine each first sampling trajectory and each second sampling trajectory by using a sliding window through a pre-processing module in the timing trajectory evaluation model, and perform normalization processing on the first sampling trajectory and the second sampling trajectory through the pre-processing module.

[0122] The normalized results are sequentially input into a backbone network and a specific operator in the timing trajectory evaluation model to obtain a first score of each first sampling trajectory and a second score of each second sampling trajectory.

[0123] In an embodiment, the evaluation result determination module 230 comprises a normalization processing module configured to, for each first sampling trajectory and each second sampling trajectory, perform normalization processing on timestamps of other positioning poses in the sampling trajectory based on a timestamp of a first positioning pose in the sampling trajectory.

[0124] In an embodiment, the normalization processing module is further configured to, for each first sampling trajectory and each second sampling trajectory, perform normalization processing on other positioning poses in the sampling trajectory based on a first positioning pose in the sampling trajectory.

[0125] In an embodiment, the judgment module 220 comprises a first judgment module configured to, if the absolute positioning pose is located in a preset prior region, enable the relative positioning pose of the target vehicle at the current time.

[0126] In an embodiment, the judgment module 220 comprises a second judgment module configured to, if the absolute positioning pose is not located in the preset prior region and a first average score of the absolute positioning pose determined based on the timing trajectory evaluation model is less than a preset score threshold, enable the relative positioning pose of the target vehicle at the current time.

[0127] In an embodiment, the evaluation result determination module 230 comprises an actual gap determination module configured to determine a pose alignment matrix according to the absolute positioning pose and the relative positioning pose, convert the relative positioning pose into a first absolute positioning pose according to the pose alignment matrix, and determine an actual gap between the first absolute positioning pose and the absolute positioning pose.

[0128] In an embodiment, the actual gap determination module comprises an alignment matrix determination module configured to take the absolute positioning pose and the relative positioning pose of the same frame as a positioning pose pair, when the positioning pose pair is not less than N, determine an initial pose alignment matrix according to the positioning pose pair.

[0129] According to the absolute positioning pose and the relative positioning pose, the initial pose alignment matrix is optimized until the Euclidean distance difference of the absolute positioning pose of the current frame and the relative positioning pose is less than a first threshold, and the optimized initial pose alignment matrix of the current frame is taken as the pose alignment matrix.

[0130] The vehicle positioning apparatus provided by the embodiments of the present disclosure can perform the steps in the vehicle positioning method provided by the embodiments of the present disclosure, and the same beneficial effects can be obtained, which will not be described here.

[0131] Figure 6 FIG. 1 is a structural schematic diagram of an electronic device according to an embodiment of the present disclosure. The following specifically refers to Figure 6 which shows a structural schematic diagram of an electronic device 500 suitable for implementing the electronic device according to an embodiment of the present disclosure. Figure 6 The electronic device shown is merely an example, and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0132] As shown in Figure 6 , the electronic device 500 can include a processing device (such as a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to the programs stored in a read-only memory (ROM) 502 or loaded from a storage device 508 into a random access memory (RAM) 503 to implement the method of the embodiments as described in the present disclosure. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0133] In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, the embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for executing the method shown in the flowcharts, thereby implementing the vehicle control method as described above. In such embodiments, the computer program can be downloaded and installed from a network through a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiments of the present disclosure are performed.

[0134] It should be noted that the computer readable medium described above in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium may, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination of the above. More specific examples of the computer readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus. In the present disclosure, the computer readable signal medium can include a data signal carried in a baseband or as a part of a carrier wave, which carries computer readable program code. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal or any suitable combination of the above. The computer readable signal medium can also be any computer readable medium other than the computer readable storage medium, which can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or apparatus. The program code contained in the computer readable medium can be transmitted by any suitable medium, including but not limited to a wire, a cable, an RF (radio frequency) or the like, or any suitable combination of the above.

[0135] The computer readable medium described above can be contained in the electronic device described above; or can exist separately and not be assembled into the electronic device. The computer readable medium described above carries one or more programs, which, when executed by the electronic device, cause the electronic device to: obtain an absolute positioning pose of a target vehicle at a current time; determine whether to enable a relative positioning pose of the target vehicle at the current time according to the absolute positioning pose; if so, determine an actual gap between the absolute positioning pose and the relative positioning pose, and in a case where the actual gap is greater than a preset threshold, obtain a first score of each first sampling trajectory in an absolute positioning trajectory including absolute positioning poses of m historical time points and a second score of each second sampling trajectory in a relative positioning trajectory including relative positioning poses of m historical time points according to a pre-trained time sequence trajectory evaluation model; and determine a positioning result of the target vehicle according to the first scores and the second scores.

[0136] Optionally, when the one or more programs are executed by the electronic device, the electronic device can further perform other steps as described in the above embodiments.

[0137] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more of: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0138] The above description is merely illustrative of the exemplary embodiments of this disclosure and the principles of the technology involved. It is understood that modifications and variations of the disclosed embodiments are possible, and also fall within the scope of the disclosure, which is defined by the appended claims. For example, the disclosed features can be combined in any combination, and are not limited to the combinations explicitly disclosed herein.

Claims

1. A vehicle positioning method characterized by comprising: The method comprises: acquiring an absolute positioning pose of a target vehicle at a current time point; determining whether to enable a relative positioning pose of the target vehicle at the current time point according to the absolute positioning pose; if yes, determining an actual distance between the absolute positioning pose and the relative positioning pose, and in a case where the actual distance is greater than a preset threshold, inputting an absolute positioning trajectory comprising at least m absolute positioning poses of historical time points and a relative positioning trajectory comprising at least m relative positioning poses into a time sequence trajectory evaluation model, determining each first sampling trajectory of the absolute positioning trajectory and each second sampling trajectory of the relative positioning trajectory by using a sliding window through a pre-processing module in the time sequence trajectory evaluation model, and performing normalization processing on the each first sampling trajectory and the each second sampling trajectory through the pre-processing module; inputting the normalization results into a backbone network and a specific operator of the time sequence trajectory evaluation model in sequence to obtain a first score of each first sampling trajectory and a second score of each second sampling trajectory; determining a positioning result of the target vehicle according to the first score and the second score.

2. The method of claim 1, wherein, The method further comprises: determining a number of first scores based on the absolute positioning trajectory and a number of second scores based on the relative positioning trajectory through a continuous frame score determinator, and determining whether the number of first scores or the number of second scores is greater than K; in a case where the number of first scores is less than K or the number of second scores is less than K, taking a last frame absolute positioning pose in the absolute positioning trajectory as the positioning result of the target vehicle.

3. The method of claim 2, wherein, The method further comprises: in a case where the number of first scores is not less than K and the number of second scores is not less than K, determining a first average score of the first scores and a second average score of the second scores, and taking a last frame positioning pose in a positioning trajectory with the highest average score between the first average score and the second average score as the positioning result of the target vehicle, wherein the positioning trajectory comprises the absolute positioning trajectory or the relative positioning trajectory.

4. The method of claim 1, wherein, The method further comprises: in a case where the actual distance is not greater than the preset threshold, taking a last frame absolute positioning pose in the absolute positioning trajectory as the positioning result of the target vehicle.

5. The method of claim 1, wherein, The method further comprises: in a case where the actual distance is greater than the preset threshold, if a number of absolute positioning poses in the absolute positioning trajectory is less than m or if a number of relative positioning poses in the relative positioning trajectory is less than m, taking a last frame absolute positioning pose in the absolute positioning trajectory as the positioning result of the target vehicle.

6. The method of claim 1, wherein, The method further comprises: if the absolute positioning pose is located in a preset prior region, enabling the relative positioning pose of the target vehicle at the current time point.

7. The method of claim 3, wherein, The method further comprises: If the absolute positioning pose is not located in the preset prior region, and a first average score of the absolute positioning pose determined based on the time sequence trajectory evaluation model is less than a preset score threshold, the relative positioning pose of the target vehicle at the current time is enabled.

8. The method of claim 1, wherein, The determination of the actual difference between the absolute positioning pose and the relative positioning pose comprises: determining a pose alignment matrix according to the absolute positioning pose and the relative positioning pose, and converting the relative positioning pose into a first absolute positioning pose according to the pose alignment matrix; determining the actual difference between the first absolute positioning pose and the absolute positioning pose.

9. The method of claim 8, wherein, The determination of the pose alignment matrix according to the absolute positioning pose and the relative positioning pose comprises: taking the absolute positioning pose and the relative positioning pose of the same frame as a positioning pose pair, and determining an initial pose alignment matrix according to the positioning pose pair when the positioning pose pair is not less than N; optimizing the initial pose alignment matrix according to the absolute positioning pose and the relative positioning pose until the Euclidean distance difference between the absolute positioning pose and the relative positioning pose of the current frame is less than a first threshold, and taking the optimized initial pose alignment matrix of the current frame as the pose alignment matrix.

10. The method of claim 1, wherein, The normalization processing of the first sampling trajectories and the second sampling trajectories by the pre-processing module comprises: For each first sampling trajectory and each second sampling trajectory, the timestamps of other positioning poses in the sampling trajectory are normalized based on the timestamp of the first positioning pose in the sampling trajectory.

11. The method of claim 1, wherein, The normalization processing of the first sampling trajectories and the second sampling trajectories by the pre-processing module comprises: For each first sampling trajectory and each second sampling trajectory, other positioning poses in the sampling trajectory are normalized based on the first positioning pose in the sampling trajectory.

12. A vehicle positioning apparatus characterized by comprising: It comprises: a pose acquisition module configured to acquire an absolute positioning pose of a target vehicle at a current time; a judgment module configured to determine whether to enable a relative positioning pose of the target vehicle at the current time according to the absolute positioning pose; an evaluation result determination module configured to determine an actual difference between the absolute positioning pose and the relative positioning pose when the relative positioning pose of the target vehicle at the current time is enabled; In the case where the actual difference is greater than a preset threshold, an absolute positioning trajectory comprising absolute positioning poses of at least m historical times and a relative positioning trajectory comprising relative positioning poses of at least m historical times are input into a time sequence trajectory evaluation model, each first sampling trajectory of the absolute positioning trajectory and each second sampling trajectory of the relative positioning trajectory are determined by a pre-processing module in the time sequence trajectory evaluation model using a sliding window, and the first sampling trajectories and the second sampling trajectories are normalized by the pre-processing module; the normalized results are sequentially input into a backbone network and a specific operator in the time sequence trajectory evaluation model to obtain first scores of the first sampling trajectories and second scores of the second sampling trajectories; A positioning result determination module is configured to determine a positioning result of the target vehicle according to the first scores and the second scores.

13. An electronic device, comprising: The electronic device includes: one or more processors; a memory device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method in any one of claims 1-11.

14. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method in any one of claims 1-11. The program is executed by the processor to implement the method in any one of claims 1-11.

Images