High-precision map data revision method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a high-precision map data revision method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring high-precision map data and driving environment information of a current driving lane of a vehicle; the high-precision map data comprises virtual lane lines, and the current driving lane comprises actual lane lines; generating a judgment result for expressing whether the virtual lane line conforms to the actual lane line or not through the high-precision map data and the driving environment information; and revising the high-precision map data on the basis of the driving environment information when it is determined that the virtual lane line does not conform to the actual lane line according to the determination result. Real-time verification and correction of high-precision map data are realized. And when the high-precision map data is inconsistent with the actual road condition, correction can be carried out in time, so that the safe and reliable driving of the automatic driving vehicle can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic map drawing, and in particular, to a method for revising high-precision map data, a device for revising high-precision map data, an electronic device, and a computer-readable storage medium. Background Art

[0002] High-precision maps not only provide local decision-making support but also enable global planning. Vehicles can make decisions such as turning, lane-changing, and decelerating in advance based on a global perspective, improving driving efficiency and safety. For example, when approaching a complex or congested section, the system can plan the optimal route in advance to avoid unnecessary stops and starts, saving fuel and time.

[0003] Since road conditions change rapidly, for example, new roads are built, traffic signs are updated, and road construction occurs frequently. To ensure the accuracy of high-precision maps, timely updates are required. However, the actual update frequency often fails to keep up with the change speed, resulting in possible lag in map information, which in turn affects the accuracy of navigation and assisted driving.

[0004] The pilot assisted driving function highly depends on high-precision maps, and map information directly affects the decision-making of the automatic driving system. If the high-precision map information does not match the actual road conditions, it may cause the system to malfunction or make incorrect judgments, and may even cause the vehicle to deviate from the safe driving area, posing a serious safety hazard. Summary of the Invention

[0005] Embodiments of the present invention provide a method, a device, an electronic device, and a computer-readable storage medium for revising high-precision map data to overcome or at least partially solve the above problems.

[0006] Embodiments of the present invention disclose a method for revising high-precision map data, including:

[0007] Obtain high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines;

[0008] Generate a determination result for expressing whether the virtual lane line matches the actual lane line based on the high-precision map data and the driving environment information;

[0009] When it is determined through the determination result that the virtual lane line does not match the actual lane line, revise the high-precision map data based on the driving environment information.

[0010] Optionally, the step of generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the high-precision map data and the driving environment information includes:

[0011] Obtain information about the vehicle ahead and road information data based on the driving environment information; the information about the vehicle ahead includes the driving trajectory of the vehicle ahead.

[0012] When it is determined that there is a vehicle ahead through the information about the vehicle ahead, generate a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the driving trajectory of the vehicle ahead and the high-precision map data.

[0013] When it is determined that there is no vehicle ahead through the information about the vehicle ahead, generate a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the road information data and the high-precision map data.

[0014] Optionally, the step of generating a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the driving trajectory of the vehicle ahead and the high-precision map data includes:

[0015] Generate a decision trajectory based on the virtual lane line.

[0016] Calculate the curvature difference, lateral offset, and heading angle difference between the decision trajectory and the driving trajectory of the vehicle ahead.

[0017] Determine a loss function based on the curvature difference, the lateral offset, and the heading angle difference.

[0018] When the loss function is greater than a preset threshold, generate a determination result for expressing that the virtual lane line does not conform to the actual lane line.

[0019] Optionally, the step of generating a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the road information data and the high-precision map data includes:

[0020] Determine the actual lane number, actual lane width, actual lane shape, and actual lane attributes from the road information data.

[0021] Determine the virtual lane number, virtual lane width, virtual lane shape, and virtual lane attributes from the high-precision map data.

[0022] Generate a difference result for the actual lane number and virtual lane number, the actual lane width and the virtual lane width, the actual lane shape and the virtual lane shape, and the actual lane attributes and the virtual lane attributes.

[0023] Use the difference result to determine a determination result for expressing whether the virtual lane line conforms to the actual lane line.

[0024] Optionally, it further includes:

[0025] When it is determined through the determination result that the virtual lane line does not match the actual lane line, re-plan the driving path of the vehicle.

[0026] Optionally, the step of re-planning the driving path of the vehicle includes:

[0027] Set the oncoming lane type of the virtual lane line as a curb;

[0028] Generate a driving path for controlling the vehicle to drive away from the curb at a preset distance threshold.

[0029] Optionally, it further includes:

[0030] When it is determined that an effective driving path cannot be generated, extract the driving trajectory of the vehicle in front from the information of the vehicle in front;

[0031] Determine the driving trajectory of the vehicle in front as the driving path of the vehicle.

[0032] An embodiment of the present invention also discloses a high-precision map data revision device, including:

[0033] A data acquisition module for acquiring high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines;

[0034] A determination result generation module for generating a determination result for expressing whether the virtual lane line matches the actual lane line through the high-precision map data and the driving environment information;

[0035] A high-precision map data revision module for revising the high-precision map data based on the driving environment information when it is determined through the determination result that the virtual lane line does not match the actual lane line.

[0036] An embodiment of the present invention also discloses an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0037] The memory is used for storing a computer program;

[0038] The processor is used for implementing the method as described in the embodiment of the present invention when executing the program stored on the memory.

[0039] An embodiment of the present invention also discloses a computer-readable storage medium, on which instructions are stored, and when executed by one or more processors, cause the processors to execute the method as described in the embodiment of the present invention.

[0040] The embodiments of the present invention have the following advantages:

[0041] In the embodiments of the present invention, by obtaining high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines; based on the high-precision map data and the driving environment information, a determination result for expressing whether the virtual lane lines match the actual lane lines is generated; when it is determined through the determination result that the virtual lane lines do not match the actual lane lines, the high-precision map data is revised based on the driving environment information. Real-time verification and correction of the high-precision map data are achieved. When it is found that the high-precision map data does not match the actual road conditions, it can be corrected in a timely manner, so as to ensure that the autonomous driving vehicle can drive safely and reliably. Description of the Drawings

[0042] Figure 1 is a flowchart of the steps of a method for revising high-precision map data provided in the embodiments of the present invention;

[0043] Figure 2 is a schematic flowchart of a method for revising high-precision map data and vehicle path planning provided in the embodiments of the present invention;

[0044] Figure 3 is a structural block diagram of a device for revising high-precision map data provided in the embodiments of the present invention;

[0045] Figure 4 is a hardware structural block diagram of an electronic device provided in the embodiments of the present invention;

[0046] Figure 5 is a schematic diagram of a computer-readable medium provided in the embodiments of the present invention. Detailed Embodiments

[0047] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0048] In practical applications, with the rapid development of intelligent technologies, intelligent driving technologies have been widely applied to vehicles. The popularization of intelligent driving has significantly changed the traffic mode, and more and more people choose intelligent driving to improve travel efficiency and safety. High-precision maps play a crucial role in highway pilot assist driving.

[0049] High-precision maps contain detailed road information, such as lane types, widths, curvatures, slopes, speed limit signs, and the positions of traffic lights. This information enables autonomous vehicles to better understand and adapt to the current road conditions, thus making more accurate driving decisions. For example, the system can adjust the speed and steering angle in advance based on the lane curvature and slope information provided by the high-precision map to ensure the stable driving of the vehicle under various complex road conditions. In addition, information such as speed limit signs and the positions of traffic lights contained in the high-precision map can help autonomous vehicles comply with traffic rules and improve driving safety.

[0050] High-precision maps not only provide local decision support but also enable global path planning. Based on the global perspective of the high-precision map, vehicles can make decisions such as turning, changing lanes, and decelerating in advance, improving driving efficiency and overall safety. For example, when approaching a complex or congested section, the system can plan the optimal path in advance to avoid unnecessary stops and starts, saving fuel and reducing travel time.

[0051] However, the production and maintenance of high-precision maps require a large amount of human and material resources. High-precision sensor devices (such as lidar and high-resolution cameras) are used to collect accurate road data, and complex data processing and storage infrastructure are used to ensure the accuracy and real-time nature of the maps. Therefore, the production and maintenance costs of high-precision maps are very high.

[0052] In addition, road conditions are dynamically changing. For example, new roads are built, traffic signs are updated, and road construction occurs frequently. To ensure the accuracy of high-precision maps, frequent updates are required. However, the actual update frequency often fails to keep up with the rate of change, resulting in map information that may lag behind the actual situation, thus affecting the accuracy of navigation and assisted driving.

[0053] Since the navigation assist driving function highly depends on high-precision maps, their information directly affects the decisions of the autonomous driving system. If the information in the high-precision map does not match the actual road conditions, it may cause the system to malfunction or make incorrect judgments, and may even cause the vehicle to deviate from the safe driving area, posing serious safety hazards.

[0054] Therefore, in response to this situation, the autonomous driving system should adopt some strategies to update and adjust the driving path in a timely manner according to the actual road conditions, avoid the occurrence of dangerous situations during autonomous driving, and at the same time improve the safety and reliability of the autonomous driving system.

[0055] The embodiments of the present invention propose a scenario where the virtual lane lines given by the high-precision map do not match the actual road. By combining real-time road information, the path selection strategy is updated to drive at a certain distance from the roadside, enhancing the safety of the autonomous driving system. In addition, the embodiments of the present invention implement data processing by deploying an artificial intelligence learning model in the vehicle, without relying on communication devices on the road infrastructure. The following functions can be achieved only by the vehicle itself and related sensors: 1. Real-time monitoring of road conditions and adjustment of path planning according to actual situations to ensure the safe driving of the vehicle under various road conditions; 2. Utilizing on-vehicle sensors and computing devices to process and analyze road information in real time to ensure the accuracy and timeliness of high-precision map information; 3. Reducing the dependence on external communication devices and infrastructure and improving the independence and reliability of the system. It overcomes dangerous driving behaviors caused by the mismatch between the high-precision map and the actual road, such as driving along the roadside, and significantly enhances the safety of the autonomous driving system.

[0056] Referring to Figure 1 , the flowchart of steps of a method for revising high-precision map data provided in an embodiment of the present invention is shown, which may specifically include the following steps:

[0057] Step 101, obtaining high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines;

[0058] Step 102, generating a determination result for expressing whether the virtual lane lines match the actual lane lines through the high-precision map data and the driving environment information;

[0059] Step 103, when it is determined through the determination result that the virtual lane lines do not match the actual lane lines, revising the high-precision map data based on the driving environment information.

[0060] In practical applications, the embodiments of the present invention can be applied to vehicles.

[0061] The embodiments of the present invention can obtain high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines;

[0062] In specific implementation, high-precision map data is an important information source for autonomous vehicles. It contains rich road information, including virtual lane lines, road signs, traffic lights, etc. These information are crucial for vehicle positioning, path planning, and decision-making.

[0063] Embodiments of the present invention can also obtain the driving environment information of the current driving lane of the vehicle to understand the road environment where the vehicle is currently located, including information such as the position, type, and curvature of the actual lane lines. This information can be obtained in real time through on-vehicle sensors (such as cameras, lidar, etc.); Embodiments of the present invention can also distinguish between virtual lane lines and actual lane lines. The high-precision map may contain virtual lane lines, which do not actually exist but are generated for the convenience of map data expression. Distinguishing between virtual lane lines and actual lane lines is important for subsequent judgment and processing.

[0064] In embodiments of the present invention, by obtaining high-precision map data and the driving environment information of the current driving lane of the vehicle, a necessary data basis is provided for subsequent judgment and processing. By distinguishing between virtual lane lines and actual lane lines, confusion and errors in subsequent processing can be avoided.

[0065] Embodiments of the present invention can also generate a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the high-precision map data and the driving environment information;

[0066] Example 1: Lane line deviation on the highway;

[0067] Scenario: The vehicle is driving on the highway. The high-precision map shows that the current lane line is a dotted line, but the actual lane line is a solid line and is slightly offset in position.

[0068] High-precision map data: Includes information such as the lane line type (dotted line) and the lane line position (based on the map coordinate system).

[0069] Driving environment information:

[0070] Visual perception: The image captured by the on-vehicle camera shows that the actual lane line is a solid line and its position deviates from the record in the high-precision map.

[0071] Lidar: The point cloud data scanned by the lidar also shows that the position of the actual lane line deviates from the record in the high-precision map.

[0072] Determination method:

[0073] Lane line type comparison: Comparing the lane line type (dotted line) in the high-precision map with the lane line type (solid line) perceived visually, it is found that they are inconsistent.

[0074] Lane line position comparison: Comparing the lane line position in the high-precision map with the lane line position obtained by visual perception / lidar, it is found that there is an offset.

[0075] Determination result: The virtual lane line does not conform to the actual lane line.

[0076] Example 2: Blurred lane lines on urban roads;

[0077] Scenario: The vehicle is driving on an urban road. The high-precision map shows that the current lane lines are clearly visible, but the actual lane lines are blurred due to wear and other reasons.

[0078] High-precision map data: Includes information such as lane line type (solid / dashed) and lane line position.

[0079] Driving environment information:

[0080] Visual perception: The image captured by the in-vehicle camera shows that the actual lane lines are blurred and difficult to accurately identify.

[0081] Judgment method:

[0082] Lane line clarity judgment: Analyze the camera image through an image processing algorithm to determine whether the clarity of the actual lane lines meets the requirements.

[0083] Match with high-precision map data: Even if the actual lane lines are blurred, if their general position matches the lane line position in the high-precision map, it can be preliminarily judged as consistent.

[0084] Judgment result: Within a certain allowable error range, it can be considered that the virtual lane lines are basically consistent with the actual lane lines. However, it should be noted that due to the blurred actual lane lines, the system may need to rely more on high-precision map data for auxiliary decision-making.

[0085] Example 3: Road construction causes lane line changes;

[0086] Scenario: The vehicle drives to a section of the road under construction. The high-precision map shows that there are two lane lines in this section, but actually one of the lane lines has been closed due to construction.

[0087] High-precision map data: Includes information such as the position and type of two lane lines.

[0088] Driving environment information:

[0089] Visual perception: The image captured by the in-vehicle camera shows that there is actually only one lane line and there are construction signs.

[0090] Judgment method:

[0091] Lane line quantity comparison: Compare the number of lane lines in the high-precision map with the number of lane lines perceived visually and find that they are inconsistent.

[0092] Construction sign recognition: Identify construction signs through an image recognition algorithm to confirm the road construction information.

[0093] Judgment result: The virtual lane lines do not match the actual lane lines.

[0094] In practical applications, due to problems such as untimely updates or data errors in high-precision maps, the virtual lane lines may not match the actual lane lines. Therefore, it is necessary to compare and judge the two.

[0095] Generating a judgment result can express the comparison result in a clear form, such as "the virtual lane line matches the actual lane line" or "the virtual lane line does not match the actual lane line". By judging whether the virtual lane line and the actual lane line match, problems existing in the high-precision map can be discovered in a timely manner.

[0096] In an embodiment of the present invention, when it is determined through the judgment result that the virtual lane line does not match the actual lane line, the high-precision map data is revised based on the driving environment information.

[0097] In a specific implementation, when it is found that the virtual lane line in the high-precision map does not match the actual lane line, it can be determined that the high-precision map data needs to be corrected to ensure its accuracy. Revision based on driving environment information: The basis for correction is the driving environment information sensed by the vehicle in real time, and this information can more accurately reflect the actual road conditions.

[0098] The vehicle in the embodiment of the present invention can be equipped with on-vehicle sensors for obtaining driving environment information, such as lidar, cameras, millimeter-wave radars, etc. These sensors can sense the surrounding environment in real time, including information such as the geometry of the road, the position of the lane lines, traffic signs, obstacles, etc.

[0099] Exemplarily, the revision of the high-precision map data based on the driving environment information can be achieved in the following manner.

[0100] Example 1: Correct the type of incorrect virtual lane line;

[0101] Scenario: The vehicle is driving on a highway, and the high-precision map shows that the current lane line is a dashed line, but the actual lane line is a solid line.

[0102] Driving environment information:

[0103] Visual perception: The image captured by the on-vehicle camera clearly shows that the actual lane line is a solid line.

[0104] Revision method: Lane line type comparison, and the system recognizes that the lane line type (dashed line) in the high-precision map is inconsistent with the actual lane line type (solid line).

[0105] Data update: The system modifies the type of the lane line in the high-precision map from a dashed line to a solid line.

[0106] Example 2: Correct the offset of the lane line position;

[0107] Scenario: The vehicle is driving on an urban road, and there is a deviation between the position of the current lane line shown on the high-precision map and the actual lane line position.

[0108] Driving environment information:

[0109] LiDAR: The point cloud data scanned by the LiDAR accurately shows the position of the actual lane line.

[0110] Revision method:

[0111] Lane line position comparison. The system compares the lane line position in the high-precision map with the actual lane line position obtained by the LiDAR and finds a deviation.

[0112] Data update: The system adjusts the position of the lane line in the high-precision map according to the LiDAR data to align it with the actual lane line.

[0113] Example 3: Delete the incorrect virtual lane line;

[0114] Scenario: The vehicle drives to a section of the road under construction. The high-precision map shows that there are two lane lines on this section, but actually one of the lane lines has been closed due to construction.

[0115] Driving environment information:

[0116] Visual perception: The image captured by the in-vehicle camera shows that there is actually only one lane line and there are construction signs.

[0117] Revision method: Lane line quantity comparison. The system identifies that the number of lane lines in the high-precision map is inconsistent with the actual number of lane lines.

[0118] Data update: The system deletes the closed lane line in the high-precision map.

[0119] Example 4: Add the missing road signs;

[0120] Scenario: The vehicle drives to an intersection. There is no speed limit sign in the high-precision map of this intersection, but there is actually a speed limit sign.

[0121] Driving environment information:

[0122] Visual perception: The image captured by the in-vehicle camera shows that there is a speed limit sign at the intersection.

[0123] Revision method: Road sign recognition. The system identifies that there is a speed limit sign at the actual intersection, but it is missing in the high-precision map.

[0124] Data update: The system adds the information (position, speed limit value, etc.) of this speed limit sign to the high-precision map.

[0125] In an embodiment of the present invention, when it is determined that the high-precision map data needs to be corrected, the high-precision map data is revised based on the driving environment information, improving the accuracy of the high-precision map and making it consistent with the actual road conditions. This provides a more reliable information source for autonomous vehicles, enhancing driving safety and reliability.

[0126] In an embodiment of the present invention, high-precision map data and driving environment information of the current driving lane of the vehicle are obtained; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines; based on the high-precision map data and the driving environment information, a determination result for expressing whether the virtual lane lines match the actual lane lines is generated; when it is determined through the determination result that the virtual lane lines do not match the actual lane lines, the high-precision map data is revised based on the driving environment information. Real-time verification and correction of the high-precision map data are achieved. When it is found that the high-precision map data does not match the actual road conditions, timely correction can be made, thus ensuring the safe and reliable driving of autonomous vehicles.

[0127] Based on the above embodiments, a variant embodiment of the above embodiments is proposed. Here, it should be noted that for the sake of brevity of description, only the differences from the above embodiments are described in the variant embodiments.

[0128] In an optional embodiment of the present invention, the step of generating a determination result for expressing whether the virtual lane lines match the actual lane lines based on the high-precision map data and the driving environment information includes:

[0129] Obtain vehicle information and road information data through the driving environment information; the vehicle information includes the driving trajectory of the vehicle ahead.

[0130] When it is determined through the vehicle information that there is a vehicle ahead, based on the driving trajectory of the vehicle ahead and the high-precision map data, a determination result for expressing whether the virtual lane lines match the actual lane lines is generated;

[0131] When it is determined through the vehicle information that there is no vehicle ahead, based on the road information data and the high-precision map data, a determination result for expressing whether the virtual lane lines match the actual lane lines is generated.

[0132] In an embodiment of the present invention, by comprehensively considering the vehicle information ahead and the road information data, it is possible to more flexibly determine whether the virtual lane lines in the high-precision map match the actual lane lines. It is divided into two cases: there is a vehicle ahead and there is no vehicle ahead.

[0133] Case 1: There is a vehicle ahead;

[0134] Obtain information: Obtain vehicle information ahead through the driving environment information, including the driving trajectory of the vehicle ahead.

[0135] Judgment method: Based on the driving trajectory of the leading vehicle and the high-precision map data, a judgment result is generated to express whether the virtual lane line conforms to the actual lane line.

[0136] When the driving trajectory of the leading vehicle highly coincides with the virtual lane line provided by the high-precision map, even if there is a certain deviation between the virtual lane line and the actual lane line, it can be preliminarily judged that the virtual lane line basically conforms to the actual lane line.

[0137] When the driving trajectory of the leading vehicle is significantly different from the virtual lane line provided by the high-precision map, it is judged that the virtual lane line does not conform to the actual lane line.

[0138] By using the driving trajectory of the leading vehicle as a reference, the accuracy of the high-precision map can be judged more accurately.

[0139] Even if there are certain errors in the high-precision map, as long as the driving trajectory of the leading vehicle is basically consistent with the actual lane line, the driving safety of the host vehicle can be ensured.

[0140] Case 2: There is no leading vehicle;

[0141] Obtain information: Obtain road information data through the driving environment information, including the number of lanes, lane width, lane shape, and lane attributes, etc.

[0142] Judgment method: Based on the road information data and the high-precision map data, a judgment result is generated to express whether the virtual lane line conforms to the actual lane line. Specifically, the lane information provided by the high-precision map can be compared with the actual road information to judge whether they are consistent. For example, compare the number of lanes, lane width, lane line type, etc.

[0143] By generating a judgment result to express whether the virtual lane line conforms to the actual lane line based on the road information data and the high-precision map data when it is determined that there is no leading vehicle, even without a leading vehicle, the difference between the high-precision map and the actual road can be detected in time, ensuring the driving safety of the vehicle in various situations.

[0144] In an embodiment of the present invention, vehicle information and road information data are obtained through the driving environment information; the vehicle information includes the driving trajectory of the vehicle ahead; when it is determined that there is a vehicle ahead through the vehicle information, based on the driving trajectory of the vehicle ahead and the high-precision map data, a determination result is generated for expressing whether the virtual lane line conforms to the actual lane line; when it is determined that there is no vehicle ahead through the vehicle information, based on the road information data and the high-precision map data, a determination result is generated for expressing whether the virtual lane line conforms to the actual lane line. By comprehensively considering the two situations of whether there is a vehicle ahead or not, the accuracy and robustness of the judgment are improved. By comparing the driving trajectory of the vehicle ahead or the actual road information, the accuracy of the high-precision map can be judged more effectively, and the path decision can be updated in time to ensure the safety of vehicle driving.

[0145] In an optional embodiment of the present invention, the step of generating a determination result for expressing whether the virtual lane line conforms to the actual lane line based on the driving trajectory of the vehicle ahead and the high-precision map data includes:

[0146] Generate a decision trajectory based on the virtual lane line;

[0147] Calculate the curvature difference, lateral offset, and orientation angle difference between the decision trajectory and the driving trajectory of the vehicle ahead;

[0148] Determine a loss function based on the curvature difference, the lateral offset, and the orientation angle difference;

[0149] When the loss function is greater than a preset threshold, generate a determination result for expressing that the virtual lane line does not conform to the actual lane line.

[0150] In an embodiment of the present invention, when it is determined that there is a vehicle ahead of the vehicle, the determination result can be generated in the following manner.

[0151] Obtain information: Obtain the lane line type of the current lane from the high-precision map; retain the driving trajectory of the vehicle ahead closest to the own vehicle.

[0152] In practical applications, the virtual lane line in the high-precision map may not conform to the actual lane line, resulting in a too narrow or too wide passable lane and affecting the passing experience.

[0153] Judgment method:

[0154] Compare the driving trajectory of the vehicle ahead of the vehicle ahead and the decision trajectory generated by the virtual lane line provided by the high-precision map, calculate the curvature difference, lateral offset, and orientation angle difference between the decision trajectory and the driving trajectory of the vehicle ahead, and form a loss function based on the curvature difference, the lateral offset, and the orientation angle difference.

[0155] If the loss function is greater than the preset threshold, it indicates that the two decision-making trajectories are quite different, and the virtual lane lines given by the high-precision map do not match the actual situation.

[0156] Optionally, if the loss function is less than or equal to the preset threshold, it indicates that the two decision-making trajectories are not very different, and it is determined that the virtual lane lines given by the high-precision map match the actual situation.

[0157] In the embodiments of the present invention, by generating a decision-making trajectory based on the virtual lane lines; calculating the curvature difference, lateral offset, and orientation angle difference between the decision-making trajectory and the driving trajectory of the vehicle in front; determining a loss function based on the curvature difference, the lateral offset, and the orientation angle difference; when the loss function is greater than the preset threshold, generating a determination result for expressing that the virtual lane lines do not match the actual lane lines. By comparing the driving trajectory of the vehicle in front, the accuracy of the high-precision map can be judged more accurately, and the situation where the high-precision map does not match the actual situation can be discovered in time, avoiding the vehicle from making wrong decisions.

[0158] In an optional embodiment of the present invention, the step of generating a determination result for expressing whether the virtual lane lines match the actual lane lines based on the road information data and the high-precision map data includes:

[0159] Determining the actual lane number, actual lane width, actual lane shape, and actual lane attributes from the road information data;

[0160] Determining the virtual lane number, virtual lane width, virtual lane shape, and virtual lane attributes from the high-precision map data;

[0161] Generating difference results for the actual lane number and the virtual lane number, the actual lane width and the virtual lane width, the actual lane shape and the virtual lane shape, and the actual lane attributes and the virtual lane attributes;

[0162] Using the difference results to determine a determination result for expressing whether the virtual lane lines match the actual lane lines.

[0163] In the embodiments of the present invention, when it is determined that there is no vehicle in front of the vehicle, the determination result can be generated in the following manner.

[0164] Obtain information:

[0165] Obtaining all the virtual lane numbers, virtual lane widths, virtual lane shapes, and virtual lane attributes (curvature, type, length, width) in front of the vehicle from the high-precision map data.

[0166] The road information data collected by the vehicle-mounted camera includes the number of all actual lanes in front of the vehicle, the actual lane width, the actual lane shape, and the actual lane attributes (curvature, type, length, width).

[0167] Judgment method:

[0168] The similarity between the two sets of data obtained from the high-precision map and the vehicle-mounted camera is judged by using the generative adversarial network GAN.

[0169] Processing method:

[0170] If the output result of the network GAN indicates that for the number of actual lanes and the number of virtual lanes, the actual lane width and the virtual lane width, the actual lane shape and the virtual lane shape, and the difference result of the actual lane attributes and the virtual lane attributes shows that the virtual lane line does not match the actual lane line, then the path decision needs to be updated.

[0171] Exemplarily, the similarity between the two sets of data obtained from the high-precision map and the vehicle-mounted camera can be judged by using the generative adversarial network GAN in the following way.

[0172] 1. Output result of the GAN network:

[0173] The generative adversarial network (GAN) consists of two networks: a generator and a discriminator. In this application scenario:

[0174] Generator: Receives the high-precision map data and attempts to generate "fake" data similar to the vehicle-mounted camera data.

[0175] Discriminator: Receives the high-precision map data and the vehicle-mounted camera data and attempts to distinguish which data is "real" (from the vehicle-mounted camera) and which is "fake" (generated by the generator).

[0176] The training objective of the GAN network is to make the "fake" data generated by the generator as realistic as possible so that the discriminator cannot distinguish between true and false. After training, the discriminator can be used to judge whether the high-precision map data and the vehicle-mounted camera data are similar.

[0177] The output result of the discriminator is usually a probability value indicating the likelihood that the input data comes from the real data. For example:

[0178] Output close to 1: Indicates that the input data is very likely to come from the vehicle-mounted camera (real data).

[0179] Output close to 0: Indicates that the input data is very likely to come from the high-precision map (possibly "fake" data).

[0180] 2. How to determine that the virtual lane line does not match the actual lane line based on the output result:

[0181] In this application scenario, we hope to determine whether the virtual lane lines in the high-precision map match the actual lane lines. Therefore, we can input the high-precision map data and in-vehicle camera data into the trained GAN network and observe the output result of the discriminator.

[0182] If the output result of the discriminator is close to 0: This indicates that there is a large difference between the high-precision map data and the in-vehicle camera data, that is, the virtual lane lines in the high-precision map do not match the actual lane lines.

[0183] If the output result of the discriminator is close to 1: This indicates that there is a high similarity between the high-precision map data and the in-vehicle camera data, that is, the virtual lane lines in the high-precision map basically match the actual lane lines.

[0184] 3. Example illustration:

[0185] Suppose the vehicle drives to an intersection. The high-precision map shows that there are two lane lines at this intersection, but actually, due to construction, one of the lane lines has been closed.

[0186] High-precision map data: Includes information such as the positions and types of the two lane lines.

[0187] In-vehicle camera data: The captured image shows that there is actually only one lane line and there are construction signs.

[0188] Input these two sets of data into the GAN network. If the output result of the discriminator is close to 0, it indicates that there is a large difference between the high-precision map data and the in-vehicle camera data, that is, the two lane lines in the high-precision map do not match the situation where there is actually only one lane line.

[0189] In the embodiment of the present invention, by using the generative adversarial network GAN to judge the similarity of two sets of data obtained from the high-precision map and the in-vehicle camera, it is possible to more effectively judge the difference between the high-precision map and the actual road information. Even without a vehicle in front, the problems of the high-precision map can be discovered in time to ensure driving safety.

[0190] In an optional embodiment of the present invention, it further includes:

[0191] When it is determined that the virtual lane line does not match the actual lane line based on the determination result, re-plan the driving path of the vehicle.

[0192] The embodiment of the present invention can re-plan the driving path of the vehicle when it is determined that the virtual lane line does not match the actual lane line through the determination result. The purpose is:

[0193] Ensuring driving safety: When it is determined that the virtual lane line does not match the actual lane line, it means that if the vehicle continues to drive according to the instructions of the high-precision map, it may deviate from the actual lane and even pose a danger. The primary purpose of re-planning the driving path is to ensure the driving safety of the vehicle.

[0194] Adapting to environmental changes: The actual road environment is dynamically changing, and the high-precision map may not be able to fully reflect these changes in real time. By re-planning the path, the vehicle can better adapt to the actual road environment, such as avoiding construction areas and bypassing obstacles.

[0195] Optimizing driving efficiency: Re-planning the path can help the vehicle select a better driving route, such as avoiding congested sections and choosing a shorter path, thereby improving driving efficiency.

[0196] In the embodiments of the present invention, when it is determined that the virtual lane line does not match the actual lane line through the determination result, the driving path of the vehicle is re-planned, achieving the following beneficial effects:

[0197] Improving safety: Timely detecting the difference between the high-precision map and the actual road and re-planning the path can effectively reduce the risk of traffic accidents.

[0198] Enhancing adaptability: Re-planning the path enables the vehicle to better adapt to the dynamically changing road environment and improves the robustness of the automatic driving system.

[0199] Enhancing the user experience: By optimizing the driving path, a safer, more comfortable and efficient travel experience can be provided for users.

[0200] In an optional embodiment of the present invention, the step of re-planning the driving path of the vehicle includes:

[0201] Setting the opposite lane type of the virtual lane line as the roadside;

[0202] Generating a driving path for controlling the vehicle to drive away from the roadside at a preset distance threshold.

[0203] In practical applications, the purpose of path re-planning is to re-plan a safe and feasible driving path when it is detected that the high-precision map does not match the actual road.

[0204] The principle of path re-planning is that the new path should maintain a certain distance from the roadside to ensure driving safety. The preset distance threshold can be calibrated according to actual needs.

[0205] Exemplarily, the driving path can be generated in the following way:

[0206] Modify the high-precision map information: Modify the oncoming lane type of the virtual lane line given in the high-precision map data to the curb. For example, if the left lane line type in the high-precision map is a virtual lane line, then modify the right lane line type to the curb. This is to enable the path planning algorithm to correctly identify the curb and generate a path that maintains a distance from it.

[0207] Use the A* algorithm to search for a path: The A* algorithm is a commonly used path search algorithm. By adding a loss function for the distance from the curb, the algorithm can be guided to generate a planned path that maintains a certain distance from the curb.

[0208] Curb distance loss function: The function serves to penalize paths that are too close to the curb and encourage the algorithm to select paths that are farther from the curb. The preset distance threshold for being far away can be set according to actual needs.

[0209] In the embodiment of the present invention, by setting the oncoming lane type of the virtual lane line to the curb; generating a driving path, the driving path is used to control the vehicle to drive away from the curb according to a preset distance threshold, and its beneficial effects are as follows:

[0210] Improve the accuracy of path planning: Setting the oncoming lane type of the virtual lane line to the curb can help the path planning algorithm more accurately identify the road boundary. Since the virtual lane line may deviate from the actual lane line or even be completely inconsistent, setting its oncoming lane type to the curb can force the algorithm to use the curb as a constraint condition for path planning, ensuring that the generated path does not deviate from the actual road.

[0211] Enhance driving safety: Control the vehicle to drive away from the curb. By controlling the distance between the vehicle and the curb through a preset distance threshold, it can effectively prevent the vehicle from getting too close to the curb, thereby reducing the risk of collision or lane departure. This is crucial for ensuring the driving safety of autonomous vehicles in complex or uncertain road environments.

[0212] Improve road adaptability: Adapt to the errors in the high-precision map. Since the high-precision map may have problems such as untimely updates or data errors, resulting in inconsistencies between the virtual lane line and the actual lane line. This step can effectively cope with the errors in the high-precision map by setting the oncoming lane type of the virtual lane line to the curb and controlling the vehicle to drive away from the curb, improving the adaptability of the autonomous driving system under various road conditions.

[0213] Enhance the user experience: By controlling the vehicle to maintain a certain distance from the curb, it can avoid situations where the vehicle gets too close to the curb or frequently adjusts lanes during driving, thereby improving the smoothness and comfort of driving and enhancing the user experience.

[0214] Optionally, it further includes:

[0215] When it is determined that an effective driving path cannot be generated, extract the driving trajectory of the vehicle ahead from the information of the vehicle ahead;

[0216] Determine the driving trajectory of the vehicle ahead as the driving path of the vehicle.

[0217] In practical applications, there may also be situations where no path can be searched. If, due to certain reasons (such as overly narrow roads, excessive obstacles, etc.), a path that meets the requirements cannot be searched, the vehicle can follow the driving trajectory of the vehicle ahead. This is a conservative strategy that can ensure that the vehicle will not drive close to the roadside in this scenario.

[0218] In the embodiments of the present invention, by extracting the driving trajectory of the vehicle ahead from the information of the vehicle ahead when it is determined that an effective driving path cannot be generated; and determining the driving trajectory of the vehicle ahead as the driving path of the vehicle, the following beneficial effects are achieved:

[0219] Coping with complex or unknown road conditions: In some complex or unknown road environments (such as road closures, sudden obstacles, missing map data, etc.), the autonomous driving system may not be able to generate an effective driving path through conventional path planning algorithms.

[0220] Advantages of following the trajectory of the vehicle ahead: The vehicle ahead is usually driven by an experienced driver, and its driving trajectory reflects to a certain extent the actual road conditions and drivable areas. Following the driving trajectory of the vehicle ahead can help the autonomous driving vehicle find a drivable path in complex or unknown road conditions.

[0221] Improving driving safety: When the best driving path cannot be determined, following the driving trajectory of the vehicle ahead is a relatively conservative and safe strategy. It can prevent the autonomous driving vehicle from blindly exploring in unknown road conditions and reduce the possibility of danger.

[0222] Utilizing the experience of the vehicle ahead: The driving experience of the driver of the vehicle ahead can provide a reference for the autonomous driving vehicle to better cope with complex road conditions, such as avoiding obstacles and selecting an appropriate vehicle speed.

[0223] Enhancing robustness: This strategy can enhance the fault tolerance of the autonomous driving system in complex environments. Even if the path planning algorithm fails, the vehicle can still maintain a certain driving ability by following the driving trajectory of the vehicle ahead.

[0224] Improving adaptability: By following the driving trajectory of the vehicle ahead, the autonomous driving vehicle can better adapt to various complex road conditions and improve its robustness in the actual road environment.

[0225] Simplifying the decision-making process: In the case where an effective path cannot be generated, directly following the driving trajectory of the vehicle ahead can simplify the decision-making process, reduce the calculation amount, and improve the response speed.

[0226] Quick decision-making: This strategy can provide a feasible driving plan for an autonomous vehicle in a short time, ensuring that it can respond to emergencies in a timely manner.

[0227] Embodiment 2

[0228] Reference Figure 2 , Figure 2 is a schematic flowchart of a method for high-precision map data revision and vehicle path planning provided in an embodiment of the present invention.

[0229] S1. Information collection: The vehicle collects its own chassis information, obstacle information, high-precision map data, driving environment information including visual lane line information, and local positioning information in real time. The data can be sourced from the vehicle CAN bus and other autonomous driving algorithm modules.

[0230] The settings of the two parameters of "data collection frequency" and "vehicle speed" are crucial for the information processing and decision-making of the autonomous driving system.

[0231] 1. The meaning of a data collection frequency of 10Hz is that 10Hz means collecting data 10 times per second.

[0232] In practical applications, the autonomous driving system needs to perceive the changes in the surrounding environment in real time, including the vehicle's own state, obstacle information, lane line information, etc. A collection frequency of 10Hz can provide relatively continuous and real-time environmental information, meeting the real-time requirements of the autonomous driving system. A collection frequency of 10Hz can achieve a balance between the amount of information and the amount of calculation. An excessively high frequency will generate a large amount of data, increasing the computational burden; an excessively low frequency may miss important environmental change information. Thus, it ensures that the autonomous driving system can obtain the surrounding environmental information in a timely manner, make accurate judgments and decisions, and provide a sufficient information basis for subsequent data processing and analysis.

[0233] 2. The meaning of the vehicle speed being the planned speed: The planned speed refers to the driving speed preset by the autonomous driving system according to factors such as the current road conditions, traffic rules, and vehicle state.

[0234] The purpose of the vehicle speed being the planned speed lies in:

[0235] Safety: The planned speed needs to comprehensively consider various safety factors, such as road speed limits, traffic flow, and obstacle distances, to ensure the safe driving of the vehicle.

[0236] Efficiency: The planned speed also needs to take into account driving efficiency. Under the premise of ensuring safety, an appropriate vehicle speed should be selected as much as possible to improve traffic efficiency.

[0237] Comfort: The planned speed also needs to consider the comfort of the driver and passengers, avoiding frequent acceleration and deceleration and maintaining a smooth drive.

[0238] Setting the vehicle speed to the planned speed enables the autonomous vehicle to travel at a predetermined speed, ensuring driving safety and improving the driving efficiency and comfort of the autonomous vehicle.

[0239] Therefore, the data acquisition frequency can be set to 10 Hz and the vehicle speed to the planned speed.

[0240] S2. Determine whether the virtual lane line matches the actual lane line: Combining the obstacle information and visual lane line information obtained above, determine whether the virtual lane line given by the high-precision map matches the actual road. The lane line information provided by the high-precision map and visual perception can be based on the ego-vehicle coordinate system.

[0241] When there is a leading vehicle: First, obtain the lane line type of the current lane from the high-precision map. If it is a virtual lane line and there is a leading vehicle in front of the ego-vehicle, retain the driving trajectory of the leading vehicle closest to the ego-vehicle. Compare the driving trajectory of the leading vehicle with the decision-making trajectory generated by the virtual lane line provided by the high-precision map, and calculate the curvature difference, lateral offset, and orientation angle difference between the two trajectories to form a loss function. If the loss function is large, it indicates that the two trajectories are significantly different, and the virtual lane line given by the high-precision map does not match the actual situation, and the path decision-making strategy needs to be updated.

[0242] When there is no leading vehicle: Collect road information data such as the number, width, shape, and attributes (curvature, type, length, width) of all lanes in front of the ego-vehicle from the high-precision map and the on-vehicle camera respectively, and use the generative adversarial network GAN to judge the similarity of the two sets of data. If the network output result indicates that the two sets of data are different, the path decision needs to be updated.

[0243] S3. Path replanning and high-precision map update: If it is determined that the virtual lane line matches the actual lane line, replan the path and modify the high-precision map information to plan a path that maintains a certain distance (calibratable) from the road edge.

[0244] Modify the oncoming lane type of the virtual lane line in the high-precision map to the road edge (for example, if the left lane line type in the high-precision map is a virtual lane line, modify the right lane line type to the road edge).

[0245] Use the A* algorithm to search for a path, add a road edge distance loss function, and generate a planned path that maintains a certain distance from the road edge.

[0246] If no path can be searched, follow the driving trajectory of the leading vehicle to ensure that the vehicle does not drive close to the road edge in this scenario.

[0247] S4. Optimal path planning: If it is determined that the virtual lane line does not match the actual lane line, plan the optimal path based on the high-precision map and real-time positioning information.

[0248] S5. Path Smoothing: Smooth the path between the previous and next paths frame by frame, and output the smoothed path trajectory.

[0249] Path smoothing is an important part of autonomous driving technology. Its main purpose is to make the vehicle's driving trajectory smoother and more natural, improving the riding comfort. Frame-by-frame smoothing, as the name implies, is to smooth the paths of the vehicle at different time points (frames).

[0250] Exemplarily, it can be achieved in the following way.

[0251] Data Input:

[0252] Original Path Trajectory: This refers to the sequence of original path points generated by the vehicle during driving based on sensor data and path planning algorithms. These path points may have some noise or jitter, resulting in an uneven trajectory.

[0253] Time Information: Each path point corresponds to a timestamp indicating the time when the vehicle reaches that point.

[0254] Smoothing Algorithm:

[0255] Select an appropriate smoothing algorithm: Commonly used smoothing algorithms include:

[0256] Moving Average Filtering: Average the coordinates of multiple adjacent path points to obtain the smoothed path points.

[0257] Polynomial Fitting: Fit the original path points with a polynomial function to obtain a smooth curve.

[0258] Spline Interpolation: Connect the path points through spline functions (such as B-spline, cubic spline) to generate a smooth curve.

[0259] Algorithm Parameter Adjustment: According to actual requirements, adjust the parameters of the smoothing algorithm, such as the window size of the moving average filtering, the order of the polynomial fitting, etc. These parameters will affect the smoothing effect.

[0260] Smoothing Process:

[0261] Point-by-Point Smoothing: Smooth each path point in sequence according to the time order.

[0262] Consider the Time Factor: In the smoothing process, the time factor needs to be considered to ensure that the smoothed path trajectory is also continuous in time.

[0263] Output the Smoothed Path Trajectory:

[0264] Take the sequence of smoothed path points as the vehicle's final driving trajectory.

[0265] For example, assume that during the driving of a vehicle, due to the influence of sensor noise, there are some jitters in the generated original path trajectory.

[0266] Original path trajectory: There are some irregular bends.

[0267] Smoothing algorithm: Select the cubic spline interpolation algorithm.

[0268] Smoothing process: Connect the original path points through the cubic spline function to generate a smooth curve.

[0269] Smoothed path trajectory: More stable and natural, eliminating the jitters in the original trajectory.

[0270] By performing inter-frame smoothing on the front and rear paths, the output of the smoothed path trajectory improves driving smoothness: The smoothed path trajectory is more stable, reducing the bumps and shakes of the vehicle during driving and improving ride comfort. Reducing control error: The smoothed path trajectory is more in line with the kinematic characteristics of the vehicle, reducing the control error of the vehicle when tracking the path and improving driving accuracy. Improving safety: The smooth path trajectory can reduce the risk of the vehicle skidding or rolling over, improving driving safety.

[0271] Path smoothing is an important link in autonomous driving technology, which can improve the driving smoothness, control accuracy and safety of the vehicle. Inter-frame smoothing is a commonly used method for path smoothing. By smoothing the front and rear paths, a more natural and comfortable driving trajectory can be obtained.

[0272] It should be noted that for the method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the embodiments of the present invention are not limited by the described action sequence, because according to the embodiments of the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential for the embodiments of the present invention.

[0273] The embodiment of the present application also provides a high-precision map data revision device 30. Refer to Figure 3, which shows a structural block diagram of a high-precision map data revision device provided in an embodiment of the present invention, including: a data acquisition module 310, configured to acquire high-precision map data and driving environment information of the current driving lane of the vehicle; the high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines; a determination result generation module 320, configured to generate a determination result for expressing whether the virtual lane line matches the actual lane line through the high-precision map data and the driving environment information; a high-precision map data revision module 330, configured to, when it is determined through the determination result that the virtual lane line does not match the actual lane line, revise the high-precision map data based on the driving environment information.

[0274] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant parts, please refer to the partial description of the method embodiment.

[0275] In addition, an embodiment of the present application further provides an electronic device 40, please refer to Figure 4 , including a processor 410 and a memory 420, wherein the memory 410 is configured to store a computer program; the processor 420 is configured to execute the program stored on the memory 410 to implement the high-precision map data revision method introduced in any embodiment of the present application.

[0276] As Figure 5 shown, in another embodiment provided by the present invention, a computer-readable storage medium 501 is further provided. Instructions are stored in the computer-readable storage medium, and when it runs on a computer, it causes the computer to execute the high-precision map data revision method described in the above embodiments.

[0277] In the present application, multiple means two or more.

[0278] In the present application, unless otherwise clearly defined, the terms "installation", "connection", and "connection" 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 directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific situations.

[0279] The terms "first", "second", "third", "fourth", etc. (if any) in the present application are used to distinguish similar objects and do not have to be used to describe a specific order or sequence.

[0280] In this application, the term "and / or" is merely a description of the relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. Additionally, in this application, the character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0281] If there is no special instruction, all steps of this application can be carried out sequentially or randomly. For example, the method includes steps A and B, which means that the method can include steps A and B carried out sequentially, or steps B and A carried out sequentially. For example, when it is mentioned that the method may further include step C, it means that step C can be added to the method in any order. For example, the method can include steps A, B, and C, or steps A, C, and B, or steps C, A, and B, etc.

[0282] The above are only the preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent replacements, and improvements made within the spirit and principles of this application shall be included within the protection scope of this application.

Claims

1. A method for revising high-precision map data, characterized in that, Including: Obtaining high-precision map data and driving environment information of the current driving lane of the vehicle; The high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines; Generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the high-precision map data and the driving environment information; When it is determined that the virtual lane line does not match the actual lane line through the determination result, the high-precision map data is revised based on the driving environment information.

2. The method according to claim 1, wherein The step of generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the high-precision map data and the driving environment information includes: Obtaining the leading vehicle information and road information data through the driving environment information; the leading vehicle information includes the driving trajectory of the leading vehicle; When it is determined that there is a leading vehicle through the leading vehicle information, generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the driving trajectory of the leading vehicle and the high-precision map data; When it is determined that there is no leading vehicle through the leading vehicle information, generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the road information data and the high-precision map data.

3. The method according to claim 2, wherein The step of generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the driving trajectory of the leading vehicle and the high-precision map data includes: Generating a decision trajectory based on the virtual lane line; Calculating the curvature difference, lateral offset, and heading angle difference between the decision trajectory and the driving trajectory of the leading vehicle; Determining a loss function based on the curvature difference, the lateral offset, and the heading angle difference; When the loss function is greater than a preset threshold, generating a determination result for expressing that the virtual lane line does not match the actual lane line.

4. The method according to claim 2, wherein The step of generating a determination result for expressing whether the virtual lane line matches the actual lane line based on the road information data and the high-precision map data includes: Determining the actual lane number, actual lane width, actual lane shape, and actual lane attributes from the road information data; Determining the virtual lane number, virtual lane width, virtual lane shape, and virtual lane attributes from the high-precision map data; Generating difference results for the actual lane number and virtual lane number, the actual lane width and the virtual lane width, the actual lane shape and the virtual lane shape, and the actual lane attributes and the virtual lane attributes; Using the difference results to determine a determination result for expressing whether the virtual lane line matches the actual lane line.

5. The method according to any one of claims 1-4, characterized in that, Also including: When it is determined that the virtual lane line does not match the actual lane line through the determination result, re-planning the driving path of the vehicle.

6. The method according to any one of claims 5, characterized in that The step of re-planning the driving path of the vehicle includes: Setting the opposite lane type of the virtual lane line to a road edge; Generating a driving path for controlling the vehicle to drive away from the road edge at a preset distance threshold.

7. The method according to any one of claims 6, characterized in that Also including: When it is determined that an effective driving path cannot be generated, extracting the driving trajectory of the leading vehicle from the leading vehicle information; Determine the driving trajectory of the preceding vehicle as the driving path of the vehicle.

8. A high-precision map data revision device, characterized in that, It includes: A data acquisition module, configured to acquire high-precision map data and driving environment information of the current driving lane of the vehicle; The high-precision map data includes virtual lane lines, and the current driving lane includes actual lane lines; A determination result generation module, configured to generate a determination result for expressing whether the virtual lane line matches the actual lane line through the high-precision map data and the driving environment information; A high-precision map data revision module, configured to revise the high-precision map data based on the driving environment information when it is determined through the determination result that the virtual lane line does not match the actual lane line.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used to store computer programs; When the processor is configured to execute the program stored on the memory, it implements the method according to any one of claims 1-7.

10. A computer-readable storage medium, on which instructions are stored, and when executed by one or more processors, cause the processor to execute the method according to any one of claims 1-7.

Images