Vehicle positioning method, electronic equipment and storage medium

Abstract

The invention provides a vehicle positioning method, electronic equipment and a storage medium, and the method comprises the steps: obtaining an environment perception image collected by a vehicle, landmark observation information, and a landmark real position of a landmark in a map, and determining the confidence coefficient of each element in the map according to the environment perception image; the position of the vehicle at the current moment is speculated according to the actual vehicle position at the previous moment, a plurality of predicted vehicle positions at the current moment are obtained, and then the landmark observation information, the landmark real position and the confidence coefficient of each element in the map are combined for each predicted vehicle position; according to the method, the matching degree between the predicted vehicle position and the actual vehicle position is determined, so that the actual vehicle position at the current moment is determined based on the matching degree corresponding to each predicted vehicle position, and vehicle positioning is realized. And whether each element of the map is distorted is fully considered, so that the vehicle positioning precision is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to a vehicle positioning method, electronic equipment and storage medium. Background Art

[0002] With the rapid development of autonomous driving technology, valet memory parking has gradually become an important part of the assisted driving system. In this function, environment-based memory mapping is particularly important, which can improve the convenience and safety of user parking.

[0003] However, in the process of building or updating the map, the bird's-eye view of the IPM (Inverse Perspective Mapping) projected by the camera itself is distorted. In other words, intuitively, the farther away from the center of the camera, the greater the distortion. This distortion will affect the confidence of the elements in the map.

[0004] The influence of such distortion is usually not considered in the prior art, resulting in a decrease in positioning performance. Summary of the invention

[0005] In view of the above-mentioned defects or deficiencies in the prior art, the present application aims to provide a vehicle positioning method, an electronic device and a storage medium to solve the problem of low vehicle positioning accuracy caused by low confidence of some elements in the map in the prior art.

[0006] The present invention provides a vehicle positioning method, which includes:

[0007] Obtaining environmental perception images collected by the vehicle, landmark observation information, and the actual location of the landmark in the map;

[0008] Determining the confidence of each element in the map based on the environmental perception image, and determining a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment;

[0009] For each predicted vehicle position, determining a degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark true position of the landmark, and the confidence of each element in the map;

[0010] Based on the matching degree corresponding to each of the predicted vehicle positions, the actual vehicle position at the current moment is determined.

[0011] Optionally, determining the confidence of each element in the map based on the environment perception image includes:

[0012] Performing an inverse perspective mapping transformation on the environment perception image to map each pixel in the environment perception image to an element in the map;

[0013] For each element in the map, the confidence of the element is determined based on the position of the pixel point corresponding to the element in the environmental perception image.

[0014] Optionally, determining the confidence of the element based on a position of a pixel point corresponding to the element in the environment perception image includes:

[0015] Determine a pixel error, a pitch angle error, and a road height error corresponding to the element based on a position of a pixel point corresponding to the element in the environment perception image;

[0016] The confidence of the element is determined according to the pixel error, the pitch angle error, and the road height error.

[0017] Optionally, determining a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment includes:

[0018] generating a plurality of initial vehicle positions of the vehicle based on the actual vehicle position of the vehicle at a previous moment;

[0019] For each initial vehicle position, a predicted vehicle position at a current moment is determined based on the control amount of the vehicle and the initial vehicle position.

[0020] Optionally, determining a degree of match between the predicted vehicle position and the actual vehicle position at a current moment based on the predicted vehicle position, the landmark observation information, the landmark real position of the landmark, and the confidence of each element in the map includes:

[0021] Based on the predicted vehicle position, the coordinates of the landmark in the landmark observation information in the vehicle body coordinate system are converted into the map to obtain the landmark observation position corresponding to the predicted vehicle position in the map;

[0022] For each landmark, based on the landmark observation position, searching the map for the confidence of the element corresponding to the landmark observation position;

[0023] The degree of match between the predicted vehicle position and the actual vehicle position at the current moment is determined based on the landmark observation position, the landmark's true position, and the found confidence level.

[0024] Optionally, determining the degree of match between the predicted vehicle position and the actual vehicle position at the current moment according to the landmark observation position, the landmark real position of each landmark, and the found confidence level includes:

[0025] For each landmark, determine the distance between the landmark observation position and the landmark real position of the landmark, and determine the likelihood between the landmark observation position and the landmark real position based on the distance and the found confidence level;

[0026] The degree of match between the predicted vehicle position and the actual vehicle position at the current moment is determined based on the likelihood between the landmark observed positions of all landmarks and the landmark real positions.

[0027] Optionally, determining the actual vehicle position at the current moment based on the matching degree corresponding to each of the predicted vehicle positions includes:

[0028] Based on the matching degree corresponding to each predicted vehicle position, updating each predicted vehicle position;

[0029] Returning to the step of determining, for each predicted vehicle position, a degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark true position of the landmark, and the confidence of each element in the map, until an iteration stop condition is reached;

[0030] The predicted vehicle position with the highest matching degree is taken as the actual vehicle position at the current moment.

[0031] Optionally, updating each predicted vehicle position based on the matching degree corresponding to each predicted vehicle position includes:

[0032] Eliminate the predicted vehicle positions whose matching degree is lower than the preset elimination threshold;

[0033] For a predicted vehicle position whose matching degree is greater than a preset sampling threshold, position sampling is performed in an area adjacent to the predicted vehicle position to obtain a new predicted vehicle position.

[0034] The present application also provides a vehicle positioning device, which includes:

[0035] An acquisition module, used to acquire an environmental perception image collected by a vehicle, current observation information of the vehicle for each landmark in the map, and a real location of each landmark in the map;

[0036] A prediction module, used to determine the confidence of each element in the map based on the environmental perception image, and determine a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment;

[0037] a matching module, for determining, for each predicted vehicle position, a matching degree between the predicted vehicle position and the actual vehicle position at a current moment based on the predicted vehicle position, the current observation information, the true position of each landmark in the map, and the confidence of each element in the map;

[0038] The determination module is used to determine the actual vehicle position at the current moment based on the matching degree corresponding to each of the predicted vehicle positions.

[0039] The present application also provides an electronic device, the electronic device comprising:

[0040] Processor and memory;

[0041] The processor is used to execute the steps of the vehicle positioning method provided in any embodiment of the present application by calling the program or instruction stored in the memory.

[0042] An embodiment of the present application also provides a computer-readable storage medium, which stores a program or instruction, and the program or instruction enables a computer to execute the steps of the vehicle positioning method provided by any embodiment of the present application.

[0043] In summary, the present application proposes a vehicle positioning method, which obtains an environmental perception image collected by the vehicle, landmark observation information observed by the vehicle for a landmark in a map, and the real position of the landmark in the map, determines the confidence of each element in the map according to the environmental perception image, and infers the position of the vehicle at the current moment according to the actual vehicle position at the previous moment, obtains multiple predicted vehicle positions at the current moment, and then for each predicted vehicle position, combines the landmark observation information, the real position of the landmark in the map, and the confidence of each element in the map to determine the matching degree between the predicted vehicle position and the actual vehicle position, so as to The matching degree determines the actual vehicle position at the current moment, thereby realizing vehicle positioning. This method determines the confidence degree of each element in the map. In the process of determining the matching degree corresponding to the predicted vehicle position based on the landmark observation information and the landmark's true position, the confidence degree of each element in the map is introduced, and whether each element of the map is distorted is fully considered. When the corresponding position of the landmark is distorted, a smaller confidence degree can be used, and when the corresponding position of the landmark is not distorted, a larger confidence degree can be used. Compared with only determining the matching degree corresponding to the predicted vehicle position through the landmark observation information and the landmark's true position, the matching degree can be determined in combination with whether the landmark is distorted in the map, thereby improving the accuracy of vehicle positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 is a flow chart of a vehicle positioning method provided by an embodiment of the present application;

[0046] Figure 2 is a schematic diagram of a landmark observation position and a landmark real position provided in an embodiment of the present application;

[0047] Figure 3 is a structural schematic diagram of a vehicle positioning device provided in an embodiment of the present application;

[0048] Figure 4 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0050] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0051] As mentioned in the background technology, in response to the problems in the prior art, this application proposes a vehicle positioning method. Figure 1 This is a flow chart of a vehicle positioning method provided in an embodiment of the present application. The method can be applicable to autonomous driving vehicles. Specifically, it can be applicable to the meeting assistance function, valet parking function, path planning function, etc. of autonomous driving vehicles. Taking the valet parking function as an example, the vehicle can be controlled based on the positioning result (i.e. the actual vehicle position at the current moment).

[0052] See also Figure 1 , the vehicle positioning method specifically includes:

[0053] S110, obtaining environment perception images collected by the vehicle, landmark observation information, and the real location of the landmark in the map.

[0054] The environment perception image may be a perception image obtained by collecting images of the vehicle's surrounding environment by a camera mounted on the vehicle.

[0055] The landmark observation information may be information obtained by sensors (such as radar sensors, laser sensors, etc.) carried on the vehicle to perceive landmarks in the map. In the embodiment of the present application, since the landmark observation information is obtained by observing the vehicle as the center, the landmark observation information may include the observation coordinates of each landmark in the vehicle body coordinate system.

[0056] The real location of the landmark may be the location of the landmark in the map, that is, the location of the landmark in the spatial coordinate system corresponding to the map. Specifically, the landmark may be a point of interest in the map, such as a landmark including but not limited to a road sign, a lane line, a traffic light, etc. In an embodiment of the present application, the number of landmarks may be one or more.

[0057] S120, determining the confidence of each element in the map based on the environment perception image, and determining multiple predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment.

[0058] In an embodiment of the present application, since the map is determined based on the positioning results of the image and the vehicle, in order to obtain whether each element in the map is distorted, the confidence of each element in the map can also be determined through the environmental perception image, so as to describe the degree of distortion of the element, that is, the uncertainty of the element, through the confidence.

[0059] For example, the confidence of each element in the map can be determined based on the calibrated intrinsic parameters and extrinsic parameters of the camera in the vehicle, combined with the environmental perception image; wherein, the confidence of the element closer to the center position is higher, and the confidence of the element farther from the center position is lower.

[0060] In a specific implementation, determining the confidence of each element in the map based on the environment perception image includes the following steps:

[0061] Step 11: Perform inverse perspective mapping transformation on the environment perception image to map each pixel in the environment perception image to an element in the map;

[0062] Step 12: For each element in the map, determine the confidence of the element based on the position of the pixel point corresponding to the element in the environment perception image.

[0063] Among them, in step 11, the environment perception image can be transformed by inverse perspective mapping, and the pixel points in the environment perception image can be mapped to the map to obtain the 3D geometric information of each element in the map; the element can be understood as a location point in the map.

[0064] Furthermore, in step 12, for each element in the map, the confidence of the element can be calculated according to the position of its corresponding pixel point in the environment perception image. For example, the confidence of the element can be determined according to the distance between the corresponding pixel point and the center point of the environment perception image, and the greater the distance, the smaller the confidence.

[0065] In the embodiment of the present application, considering that the elements in the map are very sensitive to pixel errors, road irregularities and vehicle posture changes, the confidence of the elements can be determined by combining pixel errors, road irregularities and vehicle posture changes.

[0066] With respect to the above step 12, in one example, determining the confidence of the element based on the position of the pixel point corresponding to the element in the environment perception image includes the following steps:

[0067] Step 121, based on the position of the pixel point corresponding to the element in the environment perception image, determine the pixel error, pitch angle error and road height error corresponding to the element;

[0068] Step 122: Determine the confidence of the element according to the pixel error, the pitch angle error, and the road height error.

[0069] In step 121, based on the position of the pixel point corresponding to the element in the environment perception image, the pixel error corresponding to the element is determined, which can satisfy the following formula:

[0070]

[0071] Where u and v are the horizontal and vertical coordinates of the pixel coordinates of the corresponding pixel point, that is, the position of the corresponding pixel point in the environment perception image, x and y are the coordinates of the corresponding pixel point in the camera coordinate system, and f y is f / dy, f z It is f / dz, where f represents the focal length of the camera.

[0072] In addition, in the embodiment of the present application, it can be assumed that the virtual camera coordinate system Parallel to the road surface, the pitch angle θ of the vehicle relative to the ground will change due to road irregularities and vehicle acceleration, making this assumption not completely valid. Therefore, the pitch angle error can be defined as follows:

[0073]

[0074] Furthermore, the virtual camera coordinate system is assumed to be The pitch angle error will greatly affect the position of the element obtained by IPM. Therefore, based on the position of the pixel point corresponding to the element in the environment perception image, the pitch angle error corresponding to the element is determined, which can satisfy the following formula:

[0075]

[0076] In addition, in the embodiment of the present application, the road height error can be assumed to be the value obtained from the virtual camera coordinate system. The distance error to the actual road surface may be caused by road irregularities or changes in vehicle height D. Therefore, based on the position of the pixel point corresponding to the element in the environment perception image, the road height error corresponding to the element is determined, which can satisfy the following formula:

[0077]

[0078] It should be noted that, for the above pixel error, pitch angle error, and road height error, it can be considered that these errors all obey a Gaussian distribution with a mean of zero. After obtaining the pixel error, pitch angle error, and road height error, further, in step 122, the confidence of the element can be calculated to describe the uncertainty of the element in the IPM through the confidence.

[0079] For example, the confidence of an element may be calculated by using pixel error, pitch angle error, road height error, and the variance corresponding to each error, as shown in the following formula:

[0080]

[0081] In the formula, Represents the confidence of the element, They are the variances of the Gaussian distribution corresponding to pixel error, pitch angle error, and road height error, respectively.

[0082] Through the above examples, we can combine the fact that map elements are very sensitive to pixel errors, road irregularities and vehicle posture changes. According to pixel errors, pitch angle errors and road height errors, we can evaluate the uncertainty of each element in the map. This can accurately reflect the degree of distortion of map elements and ensure the accuracy of subsequent vehicle positioning combined with confidence.

[0083] In the embodiment of the present application, in addition to determining the confidence of each element in the map, it is also necessary to determine multiple predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment.

[0084] For example, the vehicle's current position can be predicted based on the actual vehicle position at the previous moment and combined with the vehicle's current state, such as speed, acceleration, heading angle, etc., to obtain multiple predicted vehicle positions at the current moment. The actual vehicle position and predicted vehicle position both refer to the vehicle's position in the map.

[0085] In a specific implementation, determining a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment includes:

[0086] Based on the actual vehicle position of the vehicle at the previous moment, multiple initial vehicle positions of the vehicle are generated; for each initial vehicle position, based on the control amount of the vehicle and the initial vehicle position, the predicted vehicle position at the current moment is determined.

[0087] Specifically, an area can be divided in the map with the actual vehicle position of the vehicle at the last moment as the center, and then multiple points in the area can be randomly selected as the initial vehicle positions respectively; or, multiple initial vehicle positions can be determined in the map based on the actual vehicle position of the vehicle at the last moment and Gaussian noise.

[0088] Furthermore, for each initial vehicle position, the vehicle's current position can be predicted by combining the vehicle's control quantity through the state transfer equation to obtain the predicted vehicle position at the current moment. In the process of predicting through the state transfer equation, process noise can also be introduced. For example:

[0089]

[0090] In the formula, is the predicted vehicle position at the current moment, is the initial vehicle position determined based on the actual vehicle position at the previous moment, ω i is the process noise, and f is the state transfer function.

[0091] Through the above implementation, the generation of the predicted vehicle position at the current moment can be realized. By generating multiple predicted vehicle positions as samples and then determining the actual vehicle position by recursion, the accuracy of vehicle positioning can be guaranteed. Moreover, by generating multiple predicted vehicle positions at the current moment through the actual vehicle position and control amount at the previous moment, compared with randomly generating multiple predicted vehicle positions in the entire map, the predicted vehicle position can be closer to the actual vehicle position, avoiding the influence of too many generated samples on the positioning efficiency and improving the speed of vehicle positioning.

[0092] S130. For each predicted vehicle position, determine the degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, landmark observation information, the landmark's true position, and the confidence of each element in the map.

[0093] After obtaining multiple predicted vehicle positions at the current moment, the quality of each predicted vehicle position can be evaluated. Among them, the observed landmarks can be matched one by one with the landmarks in the map. The premise of pairing is that the landmarks involved in the matching need to be in the same coordinate system. In the embodiment of the present application, the landmark observation information is in the vehicle body coordinate system, and the landmark's actual position is in the spatial coordinate system corresponding to the map. Therefore, the two coordinate systems need to be unified first.

[0094] Specifically, for each predicted vehicle position, the landmark observation information can be converted from the vehicle body coordinate system to the spatial coordinate system corresponding to the map based on the landmark observation information, that is, the coordinates of the landmark in the vehicle body coordinate system, and the predicted vehicle position, to obtain the landmark observation position at the predicted vehicle position in the map, and then combined with the landmark's true position in the map and the confidence of each element in the map to determine the degree of match between the predicted vehicle position and the actual vehicle position at the current moment.

[0095] Among them, the matching degree between the predicted vehicle position and the actual vehicle position can reflect the degree to which the predicted vehicle position is close to the actual vehicle position. The higher the matching degree, the closer the predicted vehicle position is to the actual vehicle position.

[0096] In a specific implementation, based on the predicted vehicle position, the landmark observation information, the landmark's true position, and the confidence of each element in the map, determining the matching degree between the predicted vehicle position and the actual vehicle position at the current moment includes the following steps:

[0097] Step 21: based on the predicted vehicle position, the coordinates of the landmark in the landmark observation information in the vehicle body coordinate system are converted into the map to obtain the landmark observation position corresponding to the predicted vehicle position in the map;

[0098] Step 22: for each landmark, based on the landmark observation position, find the confidence of the element corresponding to the landmark observation position in the map;

[0099] Step 23: Determine the degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the landmark observation position, the landmark's true position, and the found confidence level.

[0100] Among them, in step 21, the coordinate transformation of the landmark observation information can be achieved by the following formula:

[0101]

[0102] In the formula, x m ,y m is the landmark observation position, x c ,y c is the coordinate in the landmark observation information, that is, the landmark position observed in the vehicle coordinate system, x p ,y p is the predicted vehicle position in the map, and θ is the vehicle orientation in the map.

[0103] After obtaining the landmark observation positions of each landmark in the map, further, in step 22, for each landmark observation position, the confidence level at the landmark observation position may be searched in the map.

[0104] In step 23, the landmark observation position can be first associated with the landmark actual position, that is, the landmark observation position in the map can be matched with the landmark actual position, and then for each matched pair of landmark observation position and landmark actual position, the matching degree between the predicted vehicle position and the actual vehicle position can be determined in combination with the found confidence level.

[0105] Among them, associating the observed position of a landmark with the real position of the landmark in the map can be understood as the association between points. For example, the association between points can be achieved using the nearest neighbor method, for example, based on the distance between the observed position of the landmark and the real position of the landmark, the point closest to the current point can be used as the associated point.

[0106] With respect to the above step 23, in one example, determining the degree of match between the predicted vehicle position and the actual vehicle position at the current moment according to the landmark observation position, the landmark real position of each landmark, and the found confidence level includes the following steps:

[0107] Step 231: for each landmark, determine the distance between the landmark observation position and the landmark real position, and determine the likelihood between the landmark observation position and the landmark real position based on the distance and the found confidence level;

[0108] Step 232: Determine the degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the likelihood between the landmark observed positions and the landmark real positions of all landmarks.

[0109] Specifically, in step 231, for each landmark, that is, the associated landmark observation position and the landmark true position, the distance between the landmark observation position and the landmark true position can be calculated, and then the distance is divided by the confidence level corresponding to the landmark observation position, and the likelihood between the landmark observation position and the landmark true position is determined based on the division result.

[0110] For example, Figure 2 As shown, Figure 2 It is a schematic diagram of a landmark observation position and a landmark actual position provided in an embodiment of the present application. In the figure, the solid five-pointed star and the dotted five-pointed star are associated points. The solid five-pointed star represents the landmark observation position, and the dotted five-pointed star represents the landmark actual position. At this time, the distance between the two can be calculated, and the confidence corresponding to the solid five-pointed star is found in the confidence matrix to be 0.9, and then the distance is divided by 0.9.

[0111] The likelihood can reflect the similarity between the observed position of the landmark and the real position of the landmark, that is, the degree to which the observed position of the landmark is close to the real position of the landmark. For example, the likelihood between the observed position of the landmark and the real position of the landmark can be expressed by the following formula:

[0112]

[0113] In the formula, z t is the real position of the landmark, m(t) is the observed position of the landmark, p(z t |m(t)) is the likelihood between the observed position of the landmark and the true position of the landmark, σ is the standard deviation of the observation noise, and h(z t -m(t)) is the distance between the landmark observation position and the landmark true position divided by the confidence level corresponding to the landmark observation position.

[0114] It should be noted that in the embodiment of the present application, the purpose of determining the likelihood between the landmark observation position and the landmark real position based on the distance between the landmark observation position and the landmark real position and the confidence corresponding to the landmark observation position is that: compared with simply relying on the distance to determine the likelihood, the way of considering the landmark confidence can make the likelihood more accurate. For example, for the landmark at the center, since it is not distorted, the confidence can be close to 1. At this time, the likelihood formula can be simplified to determine the likelihood based on the distance. Since the confidence is high, the accuracy of determining the likelihood by distance is high at this time; for the landmark at the edge, since it has a certain degree of distortion, the confidence is much less than 1. At this time, the likelihood needs to be determined by combining the distance and the confidence at the same time. Since the confidence is low, the accuracy of determining the likelihood based solely on the distance is low, and it needs to be adjusted in combination with the confidence to reduce the likelihood.

[0115] Furthermore, in step 232, if the likelihood between the observed position of all landmarks and the actual position of the landmarks is obtained, all likelihoods can be multiplied to obtain the matching degree between the predicted vehicle position and the actual vehicle position, so as to reflect whether the predicted vehicle position is close to the actual vehicle position through the matching degree.

[0116] Through the above example, coordinate transformation can be used to calculate the likelihood in the same coordinate system to evaluate whether the landmarks observed at the predicted vehicle position match the landmarks in the real space well enough, and then the likelihood of all landmarks is combined to determine the match between the predicted vehicle position and the actual vehicle position, which can accurately evaluate the possibility that the predicted vehicle position is close to the actual vehicle position.

[0117] S140. Determine the actual vehicle position at the current moment based on the matching degree corresponding to each predicted vehicle position.

[0118] Specifically, after obtaining the matching degree corresponding to each predicted vehicle position, if the number of predicted vehicle positions obtained during the sampling process is sufficient, the predicted vehicle position with the highest matching degree can be directly selected as the actual vehicle position.

[0119] In an embodiment of the present application, in order to further improve the accuracy of vehicle positioning, resampling can be performed according to the matching degree corresponding to each predicted vehicle position to perform multiple iterations, and finally the actual vehicle position is determined based on the matching degree corresponding to the last batch of predicted vehicle positions.

[0120] In a specific implementation, determining the actual vehicle position at the current moment based on the matching degree corresponding to each predicted vehicle position includes the following steps:

[0121] Step 31: updating each predicted vehicle position based on the matching degree corresponding to each predicted vehicle position;

[0122] Step 32, returning to execute the step of determining the matching degree between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark's true position, and the confidence of each element in the map for each predicted vehicle position, until the iteration stop condition is reached;

[0123] Step 33: The predicted vehicle position with the highest matching degree is used as the actual vehicle position at the current moment.

[0124] In step 31, the vehicle positions may be resampled according to the matching degrees corresponding to all predicted vehicle positions to obtain new predicted vehicle positions.

[0125] With respect to the above step 31, in one example, updating each predicted vehicle position based on the matching degree corresponding to each predicted vehicle position includes:

[0126] The predicted vehicle positions with a matching degree lower than a preset elimination threshold are eliminated; for the predicted vehicle positions with a matching degree greater than a preset sampling threshold, position sampling is performed in the vicinity of the predicted vehicle position to obtain a new predicted vehicle position.

[0127] Among them, the preset rejection threshold can be a preset low matching critical value, below which it can be considered that the actual vehicle position is unlikely to be the preset rejection threshold; the preset sampling threshold can be a preset high matching critical value, above which it can be considered that the probability of the actual vehicle position is high.

[0128] Specifically, all predicted vehicle positions with a matching degree lower than a preset elimination threshold may be eliminated. At the same time, multiple points may be randomly selected around the predicted vehicle position with a matching degree greater than a preset sampling threshold, i.e., in the vicinity, as the new predicted vehicle position; or, multiple points may be randomly selected near the initial vehicle position corresponding to the predicted vehicle position (i.e., the initial vehicle position from which the predicted vehicle position is derived) as the new initial vehicle position, and a new predicted vehicle position may be derived.

[0129] Through the above example, the predicted vehicle positions with low matching degrees can be deleted, and new predicted vehicle positions can be re-collected near the predicted vehicle positions with high matching degrees. During the whole process, the number of predicted vehicle positions in each batch can be kept unchanged. In this way, the predicted vehicle positions can be made closer to the actual vehicle positions, further improving the accuracy of vehicle positioning.

[0130] After the predicted vehicle position is updated, i.e., resampling, the step of determining the degree of match between each predicted vehicle position and the actual vehicle position at the current moment can be further returned to repeat the above process until the iteration stop condition is reached. The iteration stop condition may be that the number of iterations reaches a preset number of times, or the degree of match is higher than a preset matching degree threshold, etc.

[0131] After the iteration stop condition is reached, the predicted vehicle position with the highest matching degree can be selected as the actual vehicle position at the current moment.

[0132] Through the above steps 31 to 33, resampling and iteration of the predicted vehicle position are achieved, so that the predicted vehicle position can be closer to the actual vehicle position during the iteration process, further improving the accuracy of vehicle positioning.

[0133] In an embodiment of the present application, after obtaining the actual vehicle position of the vehicle at the current moment, the map can also be updated in combination with the environmental perception image.

[0134] The vehicle positioning method provided in the embodiment of the present application obtains the environment perception image collected by the vehicle, the landmark observation information observed by the vehicle for the landmark in the map, and the real location of the landmark in the map, determines the confidence of each element in the map according to the environment perception image, and infers the location of the vehicle at the current moment according to the actual vehicle position at the previous moment, and obtains multiple predicted vehicle positions at the current moment, and then for each predicted vehicle position, combines the landmark observation information, the real location of the landmark in the map, and the confidence of each element in the map to determine the matching degree between the predicted vehicle position and the actual vehicle position, thereby The actual vehicle position at the current moment is determined to realize vehicle positioning. This method determines the confidence of each element in the map. In the process of determining the matching degree corresponding to the predicted vehicle position based on the landmark observation information and the real position of the landmark, the confidence of each element in the map is introduced, and whether each element of the map is distorted is fully considered. When the corresponding position of the landmark is distorted, a smaller confidence can be used, and when the corresponding position of the landmark is not distorted, a larger confidence can be used. Compared with only determining the matching degree corresponding to the predicted vehicle position through the landmark observation information and the real position of the landmark, the matching degree can be determined in combination with whether the landmark is distorted in the map, thereby improving the accuracy of vehicle positioning.

[0135] Figure 3 3 is a schematic diagram of the structure of a vehicle positioning device provided in an embodiment of the present application. The device includes an acquisition module 310, a prediction module 320, a matching module 330 and a determination module 340, wherein:

[0136] An acquisition module 310 is used to acquire an environment perception image collected by a vehicle, current observation information of the vehicle for each landmark in the map, and a real location of each landmark in the map;

[0137] A prediction module 320, configured to determine the confidence of each element in the map based on the environment perception image, and determine a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment;

[0138] A matching module 330, for determining, for each predicted vehicle position, a matching degree between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the current observation information, the true position of each landmark in the map, and the confidence of each element in the map;

[0139] The determination module 340 is used to determine the actual vehicle position at the current moment based on the matching degree corresponding to each of the predicted vehicle positions.

[0140] Based on the above embodiments, optionally, the prediction module 320 includes a confidence determination unit, which is used to:

[0141] An inverse perspective mapping transformation is performed on the environmental perception image to map each pixel in the environmental perception image to an element in the map; for each element in the map, a confidence level of the element is determined based on a position of a pixel corresponding to the element in the environmental perception image.

[0142] Based on the above implementations, optionally, the confidence determination unit is specifically used to:

[0143] Based on the position of the pixel point corresponding to the element in the environmental perception image, the pixel error, pitch angle error and road height error corresponding to the element are determined; and the confidence of the element is determined according to the pixel error, the pitch angle error and the road height error.

[0144] Based on the above-mentioned embodiments, optionally, the prediction module 320 includes a position prediction unit, which is used to generate multiple initial vehicle positions of the vehicle based on the actual vehicle position of the vehicle at the previous moment; for each initial vehicle position, the predicted vehicle position at the current moment is determined based on the control amount of the vehicle and the initial vehicle position.

[0145] On the basis of the above embodiments, optionally, the matching module 330 includes a coordinate conversion unit, a confidence search unit and a matching degree calculation unit, wherein:

[0146] A coordinate conversion unit, configured to convert the coordinates of the landmark in the landmark observation information in the vehicle body coordinate system into the map based on the predicted vehicle position, so as to obtain the landmark observation position corresponding to the predicted vehicle position in the map;

[0147] A confidence search unit, configured to search, for each landmark, based on the landmark observation position, the confidence of an element corresponding to the landmark observation position in the map;

[0148] The matching degree calculation unit is used to determine the matching degree between the predicted vehicle position and the actual vehicle position at the current moment according to the landmark observation position, the landmark real position of the landmark and the found confidence level.

[0149] On the basis of the above implementations, optionally, the matching degree calculation unit is specifically used to:

[0150] For each landmark, determine the distance between the landmark observation position and the landmark's true position, and based on the distance and the found confidence, determine the likelihood between the landmark observation position and the landmark's true position; based on the likelihood between the landmark observation position and the landmark's true position of all landmarks, determine the degree of match between the predicted vehicle position and the actual vehicle position at the current moment.

[0151] Based on the above implementations, optionally, the determination module 340 is specifically configured to:

[0152] Based on the matching degree corresponding to each predicted vehicle position, each predicted vehicle position is updated; returning to execute the step of determining the matching degree between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark's true position and the confidence degree of each element in the map for each predicted vehicle position, until the iteration stop condition is reached; and taking the predicted vehicle position with the highest matching degree as the actual vehicle position at the current moment.

[0153] On the basis of the above-mentioned embodiments, optionally, the determination module 340 is further used to eliminate the predicted vehicle positions whose matching degree is lower than a preset elimination threshold; for the predicted vehicle positions whose matching degree is greater than a preset sampling threshold, position sampling is performed in the vicinity of the predicted vehicle positions to obtain new predicted vehicle positions.

[0154] The vehicle positioning device provided in the embodiment of the present application can execute the steps in the vehicle positioning method provided in the method embodiment of the present application, and the execution steps and beneficial effects are no longer repeated here.

[0155] Figure 4 Schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown, the electronic device 400 includes one or more processors 401 and a memory 402 .

[0156] The processor 401 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 400 to perform desired functions.

[0157] The memory 402 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 401 may run the program instructions to implement the vehicle positioning method of any embodiment of the present application described above and / or other desired functions. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0158] In one example, the electronic device 400 may further include: an input device 403 and an output device 404, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown). The input device 403 may include, for example, a keyboard, a mouse, etc. The output device 404 may output various information to the outside, including early warning information, braking force, etc. The output device 404 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0159] Of course, to simplify, Figure 4 Only some of the components related to the present application in the electronic device 400 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application conditions, the electronic device 400 may also include any other appropriate components.

[0160] In addition to the above methods and devices, the embodiments of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the vehicle positioning method provided by any embodiment of the present application.

[0161] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0162] In addition, the embodiments of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the vehicle positioning method provided by any embodiment of the present application.

[0163] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, 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.

[0164] It should be noted that the terms used in this application are only for describing specific embodiments, rather than limiting the scope of this application. As shown in the specification and claims of this application, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the presence of other identical elements in the process, method or device including the elements.

[0165] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0166] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and its core ideas of this application. The above is only the preferred implementation method of this application. It should be pointed out that due to the limitations of textual expression and the objective existence of infinite specific structures, ordinary technicians in this technical field can make several improvements, modifications or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the concept and technical solution of the invention to other occasions without improvement, should be regarded as the scope of protection of this application.

Claims

1. A vehicle positioning method, characterized in that: include: Obtaining environmental perception images collected by the vehicle, landmark observation information, and the actual location of the landmark in the map; Determining the confidence of each element in the map based on the environmental perception image, and determining a plurality of predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment; For each predicted vehicle position, determining a degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark true position of the landmark, and the confidence of each element in the map; Based on the matching degree corresponding to each of the predicted vehicle positions, the actual vehicle position at the current moment is determined.

2. The method according to claim 1, characterized in that Determining the confidence of each element in the map based on the environmental perception image includes: Performing an inverse perspective mapping transformation on the environment perception image to map each pixel in the environment perception image to an element in the map; For each element in the map, the confidence of the element is determined based on the position of the pixel point corresponding to the element in the environmental perception image.

3. The method according to claim 2, characterized in that Determining the confidence of the element based on the position of the pixel point corresponding to the element in the environment perception image includes: Determine a pixel error, a pitch angle error, and a road height error corresponding to the element based on a position of a pixel point corresponding to the element in the environment perception image; The confidence of the element is determined according to the pixel error, the pitch angle error, and the road height error.

4. The method according to claim 1, characterized in that: Determine multiple predicted vehicle positions at the current moment based on the actual vehicle position at the previous moment, including: generating a plurality of initial vehicle positions of the vehicle based on the actual vehicle position of the vehicle at a previous moment; For each initial vehicle position, a predicted vehicle position at a current moment is determined based on the control amount of the vehicle and the initial vehicle position.

5. The method according to claim 1, characterized in that Determining a degree of match between the predicted vehicle position and the actual vehicle position at a current moment based on the predicted vehicle position, the landmark observation information, the landmark real position of the landmark, and the confidence of each element in the map, including: Based on the predicted vehicle position, the coordinates of the landmark in the landmark observation information in the vehicle body coordinate system are converted into the map to obtain the landmark observation position corresponding to the predicted vehicle position in the map; For each landmark, based on the landmark observation position, searching the map for the confidence of the element corresponding to the landmark observation position; The degree of match between the predicted vehicle position and the actual vehicle position at the current moment is determined based on the landmark observation position, the landmark's true position, and the found confidence level.

6. The method according to claim 5, characterized in that Determining the matching degree between the predicted vehicle position and the actual vehicle position at the current moment according to the landmark observation position, the landmark real position of each landmark and the found confidence level, including: For each landmark, determine the distance between the landmark observation position and the landmark real position of the landmark, and determine the likelihood between the landmark observation position and the landmark real position based on the distance and the found confidence level; The degree of match between the predicted vehicle position and the actual vehicle position at the current moment is determined based on the likelihood between the landmark observed positions and the landmark real positions of all landmarks.

7. The method according to claim 1, characterized in that Determining the actual vehicle position at the current moment based on the matching degree corresponding to each of the predicted vehicle positions includes: Based on the matching degree corresponding to each predicted vehicle position, updating each predicted vehicle position; Returning to the step of determining, for each predicted vehicle position, a degree of match between the predicted vehicle position and the actual vehicle position at the current moment based on the predicted vehicle position, the landmark observation information, the landmark true position of the landmark, and the confidence of each element in the map, until an iteration stop condition is reached; The predicted vehicle position with the highest matching degree is taken as the actual vehicle position at the current moment.

8. The method according to claim 7, characterized in that Based on the matching degree corresponding to each predicted vehicle position, updating each predicted vehicle position includes: Eliminate the predicted vehicle positions whose matching degree is lower than the preset elimination threshold; For a predicted vehicle position whose matching degree is greater than a preset sampling threshold, position sampling is performed in an area adjacent to the predicted vehicle position to obtain a new predicted vehicle position.

9. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The processor is used to execute the steps of the vehicle positioning method according to any one of claims 1 to 8 by calling the program or instruction stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the steps of the vehicle positioning method according to any one of claims 1 to 8.

Images