Lane level positioning method for high-precision map

Abstract

The invention relates to a lane level positioning method for a high-precision map, and the method comprises the following steps: S1, collecting the track point position information of a vehicle in real time through a vehicle-mounted electronic device in the driving process of the vehicle, and recognizing the marking information of a left lane and a right lane of a lane where the vehicle is located; s2, converting the vector high-precision map stored in the vehicle-mounted electronic equipment into a double-layer mask rasterized high-precision map with lane groups and lanes; s3, according to the position of the track point and the left and right lane marking information of the current lane, calculating the geometric and transmission probability from the vehicle position to the grid high-precision map; and S4, calculating a lane-level map matching result of the vehicle according to the geometric probability and the transmission probability, and determining lane-level positioning information of the vehicle. The vehicle can obtain more accurate real-time position service, so that a driver or an automatic driving system is helped to better understand and predict the driving environment, and the method has wide applicability and market potential.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving intelligent transportation technology, and more specifically, to a lane-level positioning method for high-precision maps. Background Art

[0002] High-precision maps refer to electronic maps with high precision, high freshness and high richness, covering the precise attributes and geometric information of geographical elements such as roads, lanes and traffic facilities. High-precision maps can help vehicles achieve high-precision positioning and obtain road environment information in advance, thereby assisting vehicles in real-time perception and dynamic planning. Lane-level positioning is a key prerequisite for achieving high-precision location services, path planning and navigation. Existing lane-level positioning methods usually rely on complex environmental perception algorithms to build local semantic maps, and achieve high-precision positioning of vehicles by matching features with high-precision maps. This method not only requires expensive on-board sensors, but also consumes a lot of computing resources. It is only suitable for smart driving cars with high computing power. In my country, the number of smart driving cars above level L2 has increased year by year, but ordinary cars with low computing power still dominate the market. These two types of vehicles will coexist for a long time in the future. At present, the coverage of my country's high-precision maps has exceeded 1.5 million kilometers. Making full use of high-precision maps to achieve lane-level positioning for ordinary low-computing power cars will help improve road traffic efficiency and reduce energy consumption.

[0003] Existing lane-level positioning methods usually require building a local semantic map on the vehicle side and achieving high-precision positioning by matching features with high-precision maps. These methods rely on expensive sensors, such as lidar and high-resolution cameras, and need to process a large amount of real-time data. There is a problem of high computational load, which makes the existing methods unsuitable for ordinary cars with low computing power. At present, the first problem is high computational load, and the second is low efficiency. Existing map matching methods usually use directed graph methods to describe maps, which is more convenient for traditional navigation maps. However, the amount of lane information data contained in high-precision maps is very large. Using directed graphs to represent high-precision maps will result in excessive computational load, making it difficult to effectively support batch processing and real-time applications. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a lane-level positioning method for high-precision maps, which reduces the demand for computing resources by optimizing algorithms, does not rely on expensive sensors, and can achieve efficient and accurate lane-level positioning on ordinary cars with limited computing resources.

[0005] The technical solution adopted by the present invention to solve the technical problem is: constructing a lane-level positioning method for high-precision maps, comprising the following steps: S1. During the driving process of the vehicle, the vehicle's trajectory point position information is collected in real time through the vehicle's onboard electronic equipment, and the marking information of the left and right lanes of the lane where the vehicle is located is identified; S2, converting the vector high-precision map stored in the on-board electronic device into a double-layer mask rasterized high-precision map having lane groups and lanes; S3. Calculate the geometric probability and transmission probability of the vehicle position to the grid high-precision map based on the position of the trajectory point and the left and right lane marking information of the current lane; S4. Calculate the lane-level map matching result of the vehicle based on the geometric probability and the transmission probability, and determine the lane-level positioning information of the vehicle.

[0006] According to the above scheme, in step S1, the vehicle trajectory positioning information collected by the on-board electronic device includes the horizontal and vertical coordinates of the vehicle in the world coordinate system. The identified left and right lane marking information includes the coordinates of the starting and ending points of the lane markings in the image coordinate system: ,

[0007] In the formula, , are the horizontal and vertical coordinates of the starting point of the left lane line, are the horizontal and vertical coordinates of the end point of the left lane line, are the horizontal and vertical coordinates of the starting point of the right lane line, They are the horizontal and vertical coordinates of the end point of the right lane line respectively.

[0008] According to the above scheme, in step S2, the method for converting the vector high-precision map stored in the on-board electronic device into a double-layer mask rasterized high-precision map having lane groups and lanes includes the following steps: S201, rasterizing the lane groups in the high-precision map, converting each lane group into a grid unit with consistent attributes, defining the connectivity relationship between grids by the eight-neighborhood method, and constructing the first mask layer of the rasterized high-precision map, i.e., the lane group grid layer; S202. Rasterize the lane information in the high-precision map to form the second mask layer of the raster high-precision map: the lane grid. Each lane is represented as a grid unit in the lane group. The rasterized lane information includes lane type, driving direction, and lane restriction. The lane information is expressed in detail as an attribute of the grid.

[0009] According to the above scheme, in step S3, determining the candidate matching lane group grid and lane grid of the vehicle trajectory point includes the following steps: S301, according to the position of the trajectory point, search for lane group grids and lane grids within a range of 0-30m around the trajectory point, and add them to the candidate matching lane group and lane set; S302, calculating the azimuth of the trajectory according to the positions of the front and rear trajectory points, the calculation formula is as follows:

[0010] In the formula, is the azimuth of the trajectory point, For the The values ​​of the horizontal and vertical coordinates of the trajectory points, The previous trajectory point The horizontal and vertical coordinate values ​​of ; S303, calculating the angle difference between the trajectory azimuth and the candidate lanes, and deleting the lane groups and lanes in the candidate matching lane groups and lane sets whose azimuth angle difference with the trajectory is greater than a threshold of 120°; S304, calculating the geometric probability between the vehicle and each candidate lane group and lane according to the position information and direction information of the vehicle trajectory point and the left and right marking information of the current lane; S305: Calculate the transmission probability between the front and rear trajectory point candidate matching lane group grids.

[0011] According to the above scheme, in step S304, the method for calculating the geometric probability between the vehicle and each candidate lane group and lane according to the position information, direction information of the vehicle trajectory point and the left and right marking information of the current lane includes the following steps: S304a, calculating vehicle trajectory The location and candidate grid Vertical distance between lane center points , the probability density of zero-mean Gaussian distribution is used to measure the probability of similarity between the position of the vehicle trajectory and the distance between the candidate lanes. The specific formula is as follows:

[0012] In the formula, represents the distance similarity probability, Indicates the number of the trajectory point, Indicates the number of the candidate grid, Represents trajectory points The position and candidate lane grid The vertical distance between the center points, represents the standard deviation of vehicle positioning error; S304b, using the cosine probability function to calculate the similarity between the trajectory direction and the driving direction between the candidate lane grids, the specific formula is as follows:

[0013] In the formula, represents the probability of similarity in direction, Indicates The trajectory and The angle between candidate lane grids; S304c, based on the left and right lane marking information collected by the electronic device , , calculate the slope of the left and right lane markings, the specific formula is as follows:

[0014]

[0015] In the formula, is the slope of the left lane marking, is the slope of the right lane marking; S304a, according to the slopes of the lane markings on the left and right sides, obtain the ratio of the distance from the vehicle to the left lane line to the entire lane width , the specific formula is as follows:

[0016] S304d, using lane width information provided by high-precision maps Ratio , calculate the vehicle's position in the lane , the specific formula is as follows: ; S304e. Calculate the lateral position of the vehicle on the overall road based on the change in the lateral position of the vehicle. Specifically, output the lane where the vehicle is located when it starts, and calculate the lateral position of the vehicle in the current lane. When the lateral position of the vehicle gradually increases and then suddenly decreases, and the jump distance exceeds the threshold, it is considered that the vehicle has changed lanes to the right, and the lateral displacement of the vehicle on the overall road increases by one lane width based on the original position; when the lateral position of the vehicle gradually decreases and then suddenly increases, and the jump distance exceeds the threshold, it is considered that the vehicle has increased to the left, and the lateral displacement of the vehicle on the overall road decreases by one lane width based on the original position, and the lateral position of the vehicle on the overall road is obtained. Finally, the lateral deviation probability is used to measure the matching probability between the vehicle position and different lane grids. The specific formula is as follows:

[0017] In the formula, is the lateral deviation of the candidate lane grid centerline, is the standard deviation of the lateral deviation model error, is the probability of lateral deviation; S304f, obtaining the distance similarity probability between the vehicle trajectory point and the candidate grid lane , similarity probability of direction , lateral deviation probability Then, calculate the overall geometric probability , the specific formula is as follows: .

[0018] According to the above scheme, in step S305, the specific steps of calculating the transmission probability between the front and rear trajectory point candidate matching lane group grids are: S305a, first calculate the front and rear trajectory points using the distance formula and The distance between them is as follows:

[0019] In the formula, For trajectory points The horizontal and vertical coordinates of For trajectory points The horizontal and vertical coordinates of S305b, calculate the distance between the front and rear trajectory point candidate matching lane group grids through the A* algorithm, and the specific formula is as follows:

[0020] In the formula, For the The trajectory point candidate matching lane group grids, For the The grid candidate matching lane group grids; S305c, calculating front and rear trajectory points and The distance between the candidate matching lane group grid , Distance The Difference , the specific formula is as follows:

[0021] S305d, finally, based on the difference in distance between the trajectory points and the grids of the candidate lane group Calculate the overall transmission probability , the specific formula is as follows:

[0022] In the formula, Current track point The candidate matching lane group grid, The previous track point The candidate matching lane group grid.

[0023] According to the above scheme, in step S4, the specific steps of calculating the lane-level map matching result of the vehicle based on the geometric probability and the transmission probability and determining the lane-level positioning information of the vehicle are as follows: S401: The trajectory point of the vehicle Match the lane grid with its candidate With Grid Group Multiply the geometric probability and transmission probability between them, and take the lane grid corresponding to the maximum value as the final lane-level matching result of the vehicle. The specific formula is as follows: ; S402: Based on the lane-level matching result of the vehicle, the vehicle trajectory projection point The lane-level positioning result of the vehicle is calculated at the position within the matching lane. The specific formula is as follows:

[0024]

[0025] In the formula, The projection points The horizontal and vertical coordinates of It is the lane azimuth provided by the high-precision map.

[0026] The lane-level positioning method for high-precision maps of the present invention has the following beneficial effects: 1. Based on the location information collected by the on-board electronic equipment and the left and right lane line information of the vehicle's current lane, the present invention combines the rasterized high-precision map representation method to achieve lane-level map matching on low-computing power devices, significantly reducing the demand for computing resources. Ordinary vehicles can efficiently complete accurate lane-level positioning without being equipped with expensive hardware equipment, laying the foundation for the widespread application of high-precision maps, especially in intelligent transportation systems, and can achieve economical and efficient lane-level navigation and control; 2. With the map matching technology of the present invention, vehicles can perform more accurate lane-level path planning with the support of high-precision maps, which means that the vehicle can adjust the driving path in real time according to the current lane and the actual conditions of the road ahead, thereby improving driving safety and efficiency. This is of great significance for realizing autonomous driving and assisted driving, and can effectively cope with complex road environments, especially in scenarios such as multi-lane highways or complex intersections in cities; 3. The vehicle of the present invention can obtain more accurate real-time location services, thereby helping the driver or the autonomous driving system to better understand and predict the driving environment. This precise lane-level matching and positioning helps to avoid traffic accidents caused by lane departure, emergency lane change, etc., thereby improving the overall safety of road traffic. The vehicle's driving decision-making also becomes more efficient, and it can more reasonably use road resources, reduce traffic congestion and optimize the vehicle's energy consumption performance; 4. The present invention can be implemented using common vehicle-mounted electronic equipment without the need for expensive sensor configuration. It is not only suitable for high-end autonomous driving vehicles, but can also be applied to driving assistance systems of ordinary vehicles. It has wide applicability and market potential. The low-cost and efficient solution enables high-precision positioning services to benefit more users and promote the popularization of smart transportation and smart driving technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 It is a flow chart of the lane-level positioning method for high-precision maps of the present invention; Figure 2 is a schematic diagram of the structure of the lane group grid of the present invention; Figure 3 is a schematic diagram of a lane grid of the present invention; Figure 4 It is the vector high-precision map of the present invention; Figure 5 is a schematic diagram of a lane group mask of the present invention; Figure 6 Schematic diagram of lane mask of the present invention. DETAILED DESCRIPTION

[0028] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0029] like Figure 1-4 As shown, the lane-level positioning method for high-precision maps of the present invention includes the following steps: S1. Lane-level positioning device and data collection, specifically: An electronic device with integrated positioning and visual perception functions is installed on the front windshield of the vehicle. The device can collect the vehicle's positioning data and visual information of the surrounding environment in real time. It can also collect the vehicle's trajectory point location information in real time during the vehicle's driving process, and identify the marking information of the left and right lanes of the vehicle's lane. The electronic device has built-in vector high-precision map information of the road. The map content includes the lane group information, lane marking information, lane centerline information and road boundary information of the road. The algorithm built into the electronic device can identify the left and right lane marking information of the vehicle's current lane in real time. The vehicle trajectory positioning information collected by the electronic device includes the horizontal and vertical coordinates of the vehicle in the world coordinate system. The identified left and right lane marking information includes the coordinates of the starting and ending points of the lane markings in the image coordinate system: ,

[0030] In the formula, , are the horizontal and vertical coordinates of the starting point of the left lane line, are the horizontal and vertical coordinates of the end point of the left lane line, are the horizontal and vertical coordinates of the starting point of the right lane line, They are the horizontal and vertical coordinates of the end point of the right lane line respectively.

[0031] S2. Rasterize the vector high-precision map stored in the on-board electronic device and convert it into a double-layer mask high-precision map including a lane group raster layer and a lane raster layer. The specific steps are as follows: S201. In a high-precision map, one or more lanes on a road perpendicular to the road's travel direction are represented as a lane group. The number of lanes in a lane group is constant and has the same travel direction. In order to achieve rasterization of the map, the lane groups in the high-precision map are first rasterized, and each lane group is converted into a grid unit with consistent attributes, such as Figure 2 As shown. The connectivity between grids is defined by the eight-neighborhood method to ensure that adjacent grids can correctly express the spatial structure and topological relationship of the lane group. Finally, the first mask layer of the rasterized high-precision map, namely the lane group grid layer, is constructed.

[0032] S202, rasterize the lane information in the high-precision map to form the second mask layer of the raster high-precision map: lane grid. Each lane is independently represented as a grid unit in the lane group. The rasterized lane information includes attributes such as lane type, driving direction, lane restrictions, etc. This information is expressed in detail as the attributes of the grid, such as Figure 3 shown.

[0033] S3. Based on the position of the vehicle trajectory point, the information of the left and right lane markings, and the rasterized high-precision map generated in step 2, the candidate matching grids of the trajectory point are first calculated. By analyzing the distance similarity, direction similarity and lateral deviation probability between the trajectory point and each candidate matching grid, the geometric probability of the trajectory point and the candidate grid as a whole is integrated. Subsequently, the transmission probability between the candidate grids of the trajectory is evaluated by calculating the shortest path between the candidate matching grid groups of the front and rear trajectory points. The specific steps are as follows: First, determine the candidate matching lane group grid and lane grid of the vehicle trajectory point. The specific steps are as follows: S301. According to the position of the trajectory point, search for lane group grids and lane grids within 30m around the trajectory point, and add them to the candidate matching lane group and lane set.

[0034] S302: Calculate the azimuth of the trajectory according to the positions of the front and rear trajectory points. The specific calculation formula is as follows:

[0035] In the formula, is the azimuth of the trajectory point, For the The values ​​of the horizontal and vertical coordinates of the trajectory points, The previous trajectory point The horizontal and vertical coordinate values ​​of .

[0036] S303: Calculate the angle difference between the trajectory azimuth and the candidate lanes, and delete the lane groups and lanes in the candidate matching lane groups and lane sets whose azimuth angle difference with the trajectory is greater than a threshold of 120°.

[0037] S304, then calculate the geometric probability between the vehicle and each candidate lane group and lane according to the position information, direction information of the vehicle trajectory point and the left and right marking information of the current lane, the specific steps are: a. First calculate the vehicle trajectory The location and candidate grid Vertical distance between lane center points , and then use the probability density of zero-mean Gaussian distribution to measure the probability of similarity between the position of the vehicle trajectory and the candidate lane. The specific formula is as follows:

[0038] In the formula, represents the distance similarity probability, Indicates the number of the trajectory point, Indicates the number of the candidate grid, Represents trajectory points The position and candidate lane grid The vertical distance between the center points, Represents the standard deviation of vehicle positioning error.

[0039] b. Then, the cosine probability function is used to calculate the similarity between the trajectory direction and the driving direction between the candidate lane grids. The specific formula is as follows:

[0040] In the formula, represents the probability of similarity in direction, Indicates The trajectory and The angle between the candidate lane grids.

[0041] c. Based on the left and right lane marking information collected by electronic equipment , , calculate the slope of the left and right lane markings, the specific formula is as follows:

[0042]

[0043] In the formula, is the slope of the left lane marking, is the slope of the right lane marking.

[0044] d. According to the slope of the lane markings on the left and right sides, get the ratio of the distance from the vehicle to the left lane line to the entire lane width , the specific formula is as follows:

[0045] e. Use lane width information provided by high-precision maps Ratio , calculate the vehicle's position in the lane , the specific formula is as follows:

[0046] f. Calculate the lateral position of the vehicle on the overall road based on the change in the lateral position of the vehicle. The specific process is as follows: Output the lane the vehicle is in when it starts, and calculate the lateral position of the vehicle in the current lane. If the lateral position of the vehicle gradually increases and then suddenly decreases, and the jump distance exceeds the threshold, it is considered that the vehicle has changed lanes to the right, and the lateral displacement of the vehicle on the overall road increases by one lane width based on the original position. If the lateral position of the vehicle gradually decreases and then suddenly increases, and the jump distance exceeds the threshold, it is considered that the vehicle has increased to the left, and the lateral displacement of the vehicle on the overall road decreases by one lane width based on the original position. Get the lateral position of the vehicle on the overall road Finally, the lateral deviation probability is used to measure the matching probability between the vehicle position and different lane grids. The specific formula is as follows:

[0047] In the formula, is the lateral deviation of the candidate lane grid centerline, is the standard deviation of the lateral deviation model error, is the probability of lateral deviation.

[0048] g. Obtain the similarity probability of the distance between the vehicle trajectory point and the candidate grid lane , similarity probability of direction , lateral deviation probability Then, calculate the overall geometric probability , the specific formula is as follows:

[0049] S305: Calculate the transmission probability between the front and rear trajectory point candidate matching lane group grids. The specific steps are: 1) First, calculate the front and rear trajectory points using the distance formula and The distance between them is as follows:

[0050] In the formula, For trajectory points The horizontal and vertical coordinates of For trajectory points The horizontal and vertical coordinates of .

[0051] 2) The distance between the candidate matching lane groups of the front and rear trajectory points is calculated by the A* algorithm. The specific formula is as follows:

[0052] In the formula, It is The trajectory point candidate matching lane group grids, It is The grid candidate matching lane group grids.

[0053] 3) Calculate the front and rear trajectory points and The distance between the candidate matching lane group grid , Distance The Difference , the specific formula is as follows:

[0054] 4) Finally, based on the difference in distance between the trajectory points and the candidate lane group grids Calculate the overall transmission probability , the specific formula is as follows:

[0055] In the formula, Current track point The candidate matching lane group grid, The previous track point The candidate matching lane group grid.

[0056] S4. Combining the geometric probability and transmission probability between the vehicle trajectory point and the candidate matching grid, select the candidate grid with the maximum comprehensive probability value, determine the lane-level map matching result of the vehicle, and calculate the precise positioning information of the vehicle on the corresponding lane based on the matching result. The specific steps are: S401: The trajectory point of the vehicle Match the lane grid with its candidate With Grid Group Multiply the geometric probability and transmission probability between them, and take the lane grid corresponding to the maximum value as the final lane-level matching result of the vehicle. The specific formula is as follows:

[0057] S402: Based on the lane-level matching result of the vehicle, the vehicle trajectory projection point The lane-level positioning result of the vehicle is calculated at the position within the matching lane. The specific formula is as follows:

[0058]

[0059] In the formula, The projection points The horizontal and vertical coordinates of It is the lane azimuth provided by the high-precision map.

[0060] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A lane-level positioning method for high-precision maps, characterized in that: The following steps are included: S1. During the driving process of the vehicle, the vehicle's trajectory point position information is collected in real time through the vehicle's onboard electronic equipment, and the marking information of the left and right lanes of the lane where the vehicle is located is identified; S2, converting the vector high-precision map stored in the on-board electronic device into a double-layer mask rasterized high-precision map having lane groups and lanes; S3. Calculate the geometric probability and transmission probability of the vehicle position to the grid high-precision map based on the position of the trajectory point and the left and right lane marking information of the current lane; S4. Calculate the lane-level map matching result of the vehicle based on the geometric probability and the transmission probability, and determine the lane-level positioning information of the vehicle.

2. The lane-level positioning method for high-precision maps according to claim 1 is characterized in that: In step S1, the vehicle trajectory positioning information collected by the on-board electronic device includes the horizontal and vertical coordinates of the vehicle in the world coordinate system. The identified left and right lane marking information includes the coordinates of the starting and ending points of the lane markings in the image coordinate system: , In the formula, , are the horizontal and vertical coordinates of the starting point of the left lane line, are the horizontal and vertical coordinates of the end point of the left lane line, are the horizontal and vertical coordinates of the starting point of the right lane line, They are the horizontal and vertical coordinates of the end point of the right lane line respectively.

3. The lane-level positioning method for high-precision maps according to claim 1 is characterized in that: In step S2, the method for converting the vector high-precision map stored in the on-board electronic device into a double-layer mask rasterized high-precision map having lane groups and lanes comprises the following steps: S201, rasterizing the lane groups in the high-precision map, converting each lane group into a grid unit with consistent attributes, defining the connectivity relationship between grids by the eight-neighborhood method, and constructing the first mask layer of the rasterized high-precision map, i.e., the lane group grid layer; S202. Rasterize the lane information in the high-precision map to form the second mask layer of the raster high-precision map: the lane grid. Each lane is represented as a grid unit in the lane group. The rasterized lane information includes lane type, driving direction, and lane restriction. The lane information is expressed in detail as an attribute of the grid.

4. The lane-level positioning method for high-precision maps according to claim 1 is characterized in that: In step S3, determining the candidate matching lane group grid and lane grid of the vehicle trajectory point includes the following steps: S301, according to the position of the trajectory point, search for lane group grids and lane grids within a range of 0-30m around the trajectory point, and add them to the candidate matching lane group and lane set; S302, calculating the azimuth of the trajectory according to the positions of the front and rear trajectory points, the calculation formula is as follows: In the formula, is the azimuth of the trajectory point, For the The values ​​of the horizontal and vertical coordinates of the trajectory points, The previous trajectory point The horizontal and vertical coordinate values ​​of ; S303, calculating the angle difference between the trajectory azimuth and the candidate lanes, and deleting the lane groups and lanes in the candidate matching lane groups and lane sets whose azimuth angle difference with the trajectory is greater than a threshold of 120°; S304, calculating the geometric probability between the vehicle and each candidate lane group and lane according to the position information and direction information of the vehicle trajectory point and the left and right marking information of the current lane; S305: Calculate the transmission probability between the front and rear trajectory point candidate matching lane group grids.

5. The lane-level positioning method for high-precision maps according to claim 4 is characterized in that: In step S304, the method for calculating the geometric probability between the vehicle and each candidate lane group and lane according to the position information, direction information of the vehicle trajectory point and the left and right marking information of the current lane includes the following steps: S304a, calculating vehicle trajectory The location and candidate grid Vertical distance between lane center points , the probability density of zero-mean Gaussian distribution is used to measure the probability of similarity between the position of the vehicle trajectory and the distance between the candidate lanes. The specific formula is as follows: In the formula, represents the distance similarity probability, Indicates the number of the trajectory point, Indicates the number of the candidate grid, Represents trajectory points The position and candidate lane grid The vertical distance between the center points, represents the standard deviation of vehicle positioning error; S304b, using the cosine probability function to calculate the similarity between the trajectory direction and the driving direction between the candidate lane grids, the specific formula is as follows: In the formula, represents the probability of similarity in direction, Indicates The trajectory and The angle between candidate lane grids; S304c, based on the left and right lane marking information collected by the electronic device , , calculate the slope of the left and right lane markings, the specific formula is as follows: In the formula, is the slope of the left lane marking, is the slope of the right lane marking; S304a, according to the slopes of the lane markings on the left and right sides, obtain the ratio of the distance from the vehicle to the left lane line to the entire lane width , the specific formula is as follows: S304d, using lane width information provided by high-precision maps Ratio , calculate the position of the vehicle in the lane , the specific formula is as follows: ； S304e. Calculate the lateral position of the vehicle on the overall road based on the change in the lateral position of the vehicle. Specifically, output the lane where the vehicle is located when it starts, and calculate the lateral position of the vehicle in the current lane. When the lateral position of the vehicle gradually increases and then suddenly decreases, and the jump distance exceeds the threshold, it is considered that the vehicle has changed lanes to the right, and the lateral displacement of the vehicle on the overall road increases by one lane width based on the original position; when the lateral position of the vehicle gradually decreases and then suddenly increases, and the jump distance exceeds the threshold, it is considered that the vehicle has increased to the left, and the lateral displacement of the vehicle on the overall road decreases by one lane width based on the original position, and the lateral position of the vehicle on the overall road is obtained. Finally, the lateral deviation probability is used to measure the matching probability between the vehicle position and different lane grids. The specific formula is as follows: In the formula, is the lateral deviation of the candidate lane grid centerline, is the standard deviation of the lateral deviation model error, is the probability of lateral deviation; S304f, obtaining the distance similarity probability between the vehicle trajectory point and the candidate grid lane , similarity probability of direction , lateral deviation probability Then, calculate the overall geometric probability , the specific formula is as follows: 。 6. The lane-level positioning method for high-precision maps according to claim 1 is characterized in that: In step S305, the specific steps of calculating the transmission probability between the front and rear trajectory point candidate matching lane group grids are: S305a, first calculate the front and rear trajectory points using the distance formula and The distance between them is as follows: In the formula, For trajectory points The horizontal and vertical coordinates of For trajectory points The horizontal and vertical coordinates of S305b, calculate the distance between the front and rear trajectory point candidate matching lane group grids through the A* algorithm, and the specific formula is as follows: In the formula, For the The trajectory point candidate matching lane group grids, For the The grid candidate matching lane group grids; S305c, calculating front and rear trajectory points and The distance between the candidate matching lane group grid , Distance The Difference , the specific formula is as follows: S305d, finally, based on the difference in distance between the trajectory points and the grids of the candidate lane group Calculate the overall transmission probability , the specific formula is as follows: In the formula, Current track point The candidate matching lane group grid, The previous track point The candidate matching lane group grid.

7. The lane-level positioning method for high-precision maps according to claim 1 is characterized in that: In step S4, the specific steps of calculating the lane-level map matching result of the vehicle according to the geometric probability and the transmission probability and determining the lane-level positioning information of the vehicle are as follows: S401: The trajectory point of the vehicle Match the lane grid with its candidate With Grid Group Multiply the geometric probability and transmission probability between them, and take the lane grid corresponding to the maximum value as the final lane-level matching result of the vehicle. The specific formula is as follows: ； S402: Based on the lane-level matching result of the vehicle, the vehicle trajectory projection point The lane-level positioning result of the vehicle is calculated at the position within the matching lane. The specific formula is as follows: In the formula, The projection points The horizontal and vertical coordinates of It is the lane azimuth provided by the high-precision map.

Images