Method and apparatus for evaluating high-precision map data requirements for autonomous vehicles

Abstract

The invention provides a method for evaluating a high-precision map data demand of an automatic driving vehicle. The method comprises the following steps: S1, obtaining a sensing range of a vehicle-mounted sensing system of the vehicle; s2, obtaining an intersection range of the intersection and road and / or lane information communicated with the intersection; s3, under the condition of considering the road and / or lane information, comparing the sensing range of the vehicle with an intersection range to check whether the sensing range of the vehicle meets a preset condition: when the vehicle does not enter the intersection, the sensing range of the vehicle can extend to the road and / or lane to be driven after the vehicle passes through the intersection; and S4, if the sensing range of the vehicle meets the preset condition, determining that the high-precision map data is not needed at the intersection, and if the sensing range of the vehicle does not meet the preset condition, determining that the high-precision map data is needed at the intersection. The invention further provides a corresponding device and a computer program product.

Description

Technical Field

[0001] The present application relates to a method for evaluating the high-precision map data requirements of an autonomous driving vehicle. The present application also relates to a method for collecting high-precision map data, a method for providing a light map, a method for autonomous driving of a vehicle, a device for evaluating the high-precision map data requirements of an autonomous driving vehicle, a device for collecting high-precision map data, a device for providing a light map, a method for autonomous driving of a vehicle, and a computer program product. Background Art

[0002] In urban intelligent driving solutions, the detailed road geometry data provided by high-precision maps is crucial for the navigation decision-making and driving control of the autonomous driving system. However, the complexity and dynamics of the urban environment pose challenges to the collection, updating and maintenance of high-precision maps. At the same time, the huge amount of data involved in high-precision maps also places high demands on the storage and processing capabilities of vehicles.

[0003] With the development of technology, autonomous driving systems are gradually evolving towards a map-light approach, which aims to reduce the reliance of autonomous driving vehicles on high-precision maps and instead make greater use of the vehicle's own perception system and basic navigation maps to complete autonomous driving. However, how to accurately identify road scenes that can achieve autonomous driving without the support of high-precision maps remains a technical challenge. Summary of the invention

[0004] The purpose of the present application is to provide a method for evaluating the high-precision map data requirements of an autonomous driving vehicle, a method for collecting high-precision map data, a method for providing a light map, a method for autonomous driving of a vehicle, a device for evaluating the high-precision map data requirements of an autonomous driving vehicle, a device for collecting high-precision map data, a device for providing a light map, a device for autonomous driving of a vehicle, and a computer program product, so as to at least solve some of the problems in the prior art.

[0005] According to a first aspect of the present application, a method for evaluating high-precision map data requirements of an autonomous driving vehicle is provided, the method comprising the following steps:

[0006] Step S1: obtaining the sensing range of the vehicle's onboard sensing system;

[0007] Step S2: Obtaining the intersection range of the intersection and the road and / or lane information connected to the intersection;

[0008] Step S3: considering the road and / or lane information, comparing the vehicle's perception range with the intersection range to check whether the vehicle's perception range meets a preset condition: before the vehicle enters the intersection, the vehicle's perception range can be extended to the road and / or lane to be entered after the vehicle passes the intersection; and

[0009] Step S4: If the perception range of the vehicle meets the preset conditions, it is determined that high-precision map data is not required at the intersection; if the perception range of the vehicle does not meet the preset conditions, it is determined that high-precision map data is required at the intersection.

[0010] This application particularly includes the following technical concepts: By comparing the vehicle perception range with the intersection range, an intuitive and effective evaluation method is established to check whether the autonomous driving vehicle can pre-identify the key scene elements necessary for crossing the intersection in the intersection traffic environment. This method can accurately screen which specific road scenes do require HD map support and which scenes do not rely on HD maps. This precise definition of the demand for HD map data helps to reduce the dependence of the autonomous driving system on HD maps while ensuring safety, and achieve a more economical and practical autonomous driving solution.

[0011] In an exemplary embodiment, the preset conditions further include: the vehicle's perception range covers all lanes of the road to be entered in width; and / or, for all potential passage schemes of the vehicle at the intersection, when the vehicle has not yet entered the intersection, the vehicle's perception range can be extended to the road and / or lane to be entered after the vehicle passes the intersection; and / or, while the vehicle is driving within the intersection, the vehicle's perception range always remains able to extend to the road and / or lane to be entered after the vehicle passes the intersection. Thus, it can be ensured that when a high-precision map is not used, the vehicle can perceive all potential driving paths and all lane mapping relationships within the intersection. In addition, it can be ensured that when a high-precision map is not used, even if the position and orientation of the perception range change significantly due to the vehicle turning in the intersection, the road and / or lane to be entered are always visible to the vehicle. Through these stricter restrictions on the perception range, it can be fully ensured that driving safety is taken into account when evaluating the high-precision map data requirements.

[0012] In an exemplary embodiment, in step S3, it is checked whether the perception range of the vehicle meets the preset conditions in the following manner: obtaining the overlapping area between the perception range of the vehicle and the intersection range when the vehicle has not yet entered the intersection; checking whether the overlapping area extends continuously between the current position of the vehicle and the local geometric boundary of the intersection range, and the local geometric boundary is directly connected to the road and / or lane to be entered after the vehicle passes through the intersection; and if the overlapping area extends continuously between the current position of the vehicle and the local geometric boundary of the intersection range, it is confirmed that the perception range of the vehicle meets the preset conditions. By confirming the continuity of the overlapping area, it can be ensured that the virtual path in the intersection is pre-visible to the vehicle, thereby more effectively proving whether it is necessary to use a high-precision map.

[0013] In an exemplary embodiment, step S2 includes determining the intersection range of the intersection based on the high-precision map, wherein: extracting the stop line and / or zebra crossing related to the intersection from the high-precision map; extracting the road boundary information related to the intersection from the high-precision map; generating feature points related to the intersection based on the stop line and / or zebra crossing and the road boundary information; connecting the feature points in sequence according to a determined connection order to form a closed figure surrounding the intersection; and determining the closed figure as the intersection range of the intersection. As a result, the high-precision map can provide richer road elements and detailed information related to the intersection, thereby significantly improving the accuracy of the intersection range definition.

[0014] In an exemplary embodiment, feature points associated with the intersection are determined in the following manner: the stop line is extended so that its extended line intersects with the road boundary line, and the intersection point of the stop line and its extended line with the road boundary line is determined as the feature point associated with the intersection; and / or, the endpoint of the zebra crossing or the intersection point of the zebra crossing and the road boundary line is determined as the feature point associated with the intersection.

[0015] In an exemplary embodiment, step S2 includes determining the intersection range of the intersection based on the high-precision map, wherein: extracting all nodes related to the intersection from the high-precision map, the nodes representing the starting point and the end point of the road segment in the high-precision map at the intersection; connecting the nodes in sequence according to the determined connection order to form a closed figure surrounding the intersection; and determining the closed figure as the intersection range of the intersection. Thus, even for areas not covered by the high-precision map, the intersection range can be effectively defined.

[0016] In an exemplary embodiment, in step S2, only the precision maps that meet the following mapping specifications are used to determine the intersection range of the intersection: under the mapping specifications, the nodes related to the intersection are located outside the intersection, and / or the positions of the nodes related to the intersection are located at the positions where the corresponding stop lines or zebra crossings are located. Through this screening of mapping specifications, it is possible to avoid using those precision maps with high abstraction or less details to construct the intersection range, thereby improving the accuracy of the intersection range determination.

[0017] In an exemplary embodiment, step S2 further includes: after forming a closed figure surrounding the intersection, determining the correspondence between the geometric boundary or local geometric boundary of the closed figure and each exit road at the intersection; marking the geometric boundary or local geometric boundary of the closed figure with the correspondence; and determining the closed figure with the marked correspondence as the intersection range of the intersection. Thus, the relationship between the geometric boundary and the exit road is clarified, which can simplify the geometric comparison process between the perception range and the intersection range.

[0018] In an exemplary embodiment, step S2 further includes: obtaining static obstacle information, slope and / or road coverage relationship at the intersection from the high-precision map and / or the standard map; and in step S3, considering the occlusion of the vehicle's perception range caused by the static obstacle information, slope and / or road coverage relationship, comparing the vehicle's perception range with the intersection range, and checking whether the vehicle's perception range meets the preset conditions. By considering the influence of factors such as static obstacles, the authenticity and effectiveness of the perception range evaluation are improved.

[0019] In an exemplary embodiment, step S3 further includes: importing the perception range of the vehicle, the intersection range of the intersection, and the road and / or lane information connected to the intersection into the simulation platform; simulating different postures of the vehicle in front of multiple intersections with the help of the simulation platform, wherein the different postures include different lane positions and / or vehicle head orientations; and, for each intersection, in each simulated posture, comparing the perception range of the vehicle with the intersection range. By simulating different postures on the simulation platform, the perception performance of the vehicle in various situations can be predicted and evaluated before actual road testing. This method significantly reduces the cost and time investment required for real vehicle testing and improves the efficiency and economy of the evaluation process.

[0020] In an exemplary embodiment, step S3 further includes: stopping the vehicle in different postures before the intersection, and acquiring in real time sensor data collected by the vehicle's onboard perception system about a determined point at the intersection, the determined point being outside the intersection and on the road and / or lane to be entered after the vehicle passes through the intersection; analyzing the signal strength, reflectivity and / or other measurable attributes of the sensor data; and checking whether the determined point is within the perception range of the vehicle based on the results of the analysis. By analyzing the sensor data in real time, it is ensured that the vehicle's perception range and the intersection range can be compared even when a high-precision map and an applicable standard-precision map are missing.

[0021] In an exemplary embodiment, the method further includes step S5: deploying a test vehicle with the perception range to attempt autonomous driving at least for the portion of intersections determined in step S4 as not requiring high-precision map data; checking whether the test vehicle can successfully drive through the portion of intersections; and, if the test vehicle can successfully drive through the portion of intersections, verifying that high-precision map data is not required at the portion of intersections. The accuracy of the evaluation results is improved through the actual vehicle test verification process.

[0022] In an exemplary embodiment, step S5 further includes: if the verification result is inconsistent with the result determined in step S4, adjusting the modeling parameters or measurement parameters of the intersection range of some intersections and / or the perception range of the vehicle-mounted perception system, and re-performing steps S3 and S4 until the re-determined result is consistent with the verification result. In this way, it can be ensured that the evaluation mechanism is continuously optimized and adapted to actual road conditions.

[0023] In an exemplary embodiment, step S4 further includes: marking the high-precision map data requirements for each intersection on the autonomous driving map covering multiple intersections; and / or storing the results of the determined high-precision map data requirements in the local storage device and / or cloud server of the vehicle and / or the original equipment manufacturer. This facilitates the understanding and application of the evaluation results, and is also conducive to the management and call of the evaluation results.

[0024] According to the second aspect of the present application, a method for collecting high-precision map data is provided, wherein the method comprises the following steps: obtaining high-precision map data requirements for a preset area, wherein the high-precision map data requirements are determined according to the method described in the first aspect of the present application; and collecting high-precision map data for at least one intersection in the preset area that requires high-precision map data; and not collecting high-precision map data for at least another intersection in the preset area that does not require high-precision map data. Thus, the collection and maintenance costs of high-precision maps are reduced.

[0025] According to the third aspect of the present application, a method for providing a lightweight map is provided, wherein the method comprises the following steps: collecting high-precision map data for a preset area according to the method described in the second aspect of the present application; creating a lightweight map for the preset area, wherein the lightweight map includes high-precision map data of at least one intersection in the preset area, and does not include high-precision map data of at least another intersection in the preset area. This lightweight map solution not only reduces the burden of storing and processing map data for autonomous driving vehicles, but also improves data processing efficiency, making the operation of the autonomous driving system more efficient.

[0026] According to the fourth aspect of the present application, a method for automatic driving of a vehicle is provided, the method comprising the following steps: obtaining the high-precision map data demand of a preset area, the high-precision map data demand is determined according to the method described in the first aspect of the present application; when the vehicle travels to a preset distance from at least one intersection requiring high-precision map data in the preset area, automatically triggering the download of high-precision map data of the at least one intersection; and controlling the vehicle to automatically drive through the at least one intersection based on the downloaded high-precision map data. This on-demand downloading method not only optimizes the efficiency of data use, but also reduces unnecessary data storage and processing burdens.

[0027] According to the fifth aspect of the present application, a method for automatic driving of a vehicle is provided, the method comprising the following steps: obtaining a light map provided by the method described in the third aspect of the present application; and controlling the automatic driving of the vehicle in a preset area based on the light map, wherein, for at least one intersection with high-precision map data in the light map, the vehicle is controlled to automatically drive through the at least one intersection based on the high-precision map data, and for at least another intersection in the light map that does not have high-precision map data, the vehicle is controlled to automatically drive through the at least another intersection based on the vehicle-mounted perception system. In this way, the dependence of the automatic driving vehicle on high-precision map data can be reduced, and the response speed and efficiency of the automatic driving system can be improved.

[0028] According to the sixth aspect of the present application, there is provided a device for evaluating the high-precision map data requirements of an autonomous driving vehicle, the device comprising a memory and a processor, the memory storing computer program instructions, and when the computer program instructions are executed by the processor, the processor is capable of executing the method described in the first aspect of the present application.

[0029] According to the seventh aspect of the present application, a device for high-precision map data collection is provided, the device comprising a memory and a processor, the memory storing computer program instructions, when the computer program instructions are executed by the processor, the processor can execute the method described in the second aspect of the present application.

[0030] According to the eighth aspect of the present application, a device for providing a light map is provided, the device comprising a memory and a processor, the memory storing computer program instructions, when the computer program instructions are executed by the processor, the processor is capable of executing the method described in the third aspect of the present application.

[0031] According to the ninth aspect of the present application, an automatic driving device for a vehicle is provided, the automatic driving device comprising a memory and a processor, the memory storing computer program instructions, when the computer program instructions are executed by the processor, the processor is capable of executing the method described in the fourth aspect or the fifth aspect of the present application.

[0032] According to the tenth aspect of the present application, a computer program product is provided, comprising computer program instructions, wherein the computer program instructions, when executed by one or more processors, enable the one or more processors to execute the method described in the first aspect, the second aspect, the third aspect, the fourth aspect and / or the fifth aspect of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The present application will be described in more detail below with reference to the accompanying drawings, so that the principles, features and advantages of the present application can be better understood. The accompanying drawings include:

[0034] Figure 1 A flow chart of a method for evaluating high-precision map data requirements of an autonomous driving vehicle according to an exemplary embodiment of the present application is shown;

[0035] Figures 2A to 2F A schematic diagram showing how to check whether the vehicle's perception range meets preset conditions by comparing the vehicle's perception range with the intersection range in different scenarios;

[0036] Figure 3A and Figure 3B A schematic diagram showing a method of determining an intersection range of an intersection based on a high-precision map according to an exemplary embodiment of the present application is shown;

[0037] Figure 4A A schematic diagram showing a method of determining an intersection range of an intersection based on a precise map according to an exemplary embodiment of the present application is shown;

[0038] Figure 4B A schematic diagram of determining the intersection range of an intersection based on a precise map according to another exemplary embodiment of the present application is shown.

[0039] Figure 4C A schematic diagram showing a precision map that does not meet cartographic specifications;

[0040] Figure 5A flow chart of a method for evaluating high-precision map data requirements of an autonomous driving vehicle according to another exemplary embodiment of the present application is shown;

[0041] Figure 6 A flow chart of a method for collecting high-precision map data according to an exemplary embodiment of the present application is shown;

[0042] Figure 7 A flow chart of a method for providing a light map according to an exemplary embodiment of the present application is shown;

[0043] Figure 8 A flow chart showing a method for autonomous driving of a vehicle according to an exemplary embodiment of the present application is shown;

[0044] Fig. 9 A flow chart showing a method for autonomous driving of a vehicle according to another exemplary embodiment of the present application is shown;

[0045] Fig.10 A block diagram of an apparatus for evaluating high-precision map data requirements of an autonomous driving vehicle according to an exemplary embodiment of the present application is shown;

[0046] Fig.11 A block diagram of a device for high-precision map data collection according to an exemplary embodiment of the present application is shown;

[0047] Fig.12 A block diagram of an apparatus for providing a light image according to an exemplary embodiment of the present application is shown;

[0048] Fig.13 A block diagram of an automatic driving device for a vehicle according to an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0049] In order to make the technical problems, technical solutions and beneficial technical effects to be solved by the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and multiple exemplary embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the scope of protection of the present application.

[0050] Figure 1 A flow chart of a method for evaluating high-precision map data requirements of an autonomous driving vehicle according to an exemplary embodiment of the present application is shown. The method exemplarily includes steps S1 to S4.

[0051] In step S1, the sensing range of the vehicle's onboard sensing system is obtained.

[0052] The vehicle perception system consists of a variety of environmental sensors, including but not limited to radar, lidar, camera, ultrasonic sensor and infrared sensor. These sensors are installed around the vehicle body (such as the front, side, rear and rearview mirror of the vehicle) to ensure comprehensive perception of the vehicle's surrounding environment.

[0053] like Figure 2A As shown in the example, by understanding the parameter configuration of the vehicle perception system, the vehicle's safety radius rs can be determined, which refers to the radius of a circular area around the vehicle, representing the safety space required by the vehicle during driving. At the same time, the visual radius rc of the front side of the vehicle can also be determined, which is expressed as a fan-shaped area in front of the vehicle, representing the perceived depth in front of the vehicle. The angle α describes the field of view of the sensor on the front side of the vehicle, that is, the horizontal viewing angle range that the sensor can cover.

[0054] By integrating the sensing ranges of each sensor, the overall sensing range of the vehicle can be constructed. FIG. 2A to FIG. 2C If there are road environment objects within this range, such as other vehicles, pedestrians or obstacles, the onboard perception system can detect and identify these objects to ensure the safe operation of the autonomous driving system.

[0055] In step S2, the intersection range and the road and / or lane information connected to the intersection are obtained.

[0056] When determining the scope of an intersection, either a map-based method or a method based on actual measurements can be used.

[0057] For intersections covered by HD maps, the intersection range can be determined directly based on the HD map. In the case where HD map data is missing or outdated, the intersection range can be determined based on the standard HD map. In both cases, appropriate graphics algorithms can be used to construct polygons or other geometric shapes to accurately identify the intersection range, which will be further explained below.

[0058] When determining the scope of an intersection through actual measurement, the scope of the intersection can be visually delineated by setting up markers or reflectors on the spot, or a rough intersection model can be first constructed on a map, and then this model can be mapped to the actual geographical environment and adjusted through on-site measurement.

[0059] The information of roads and / or lanes connected to the intersection is usually stored in a high-precision map or a standard-precision map, and can be directly obtained by accessing the map. In addition, this information can also be obtained from the Internet or a road supervision platform in combination with vehicle positioning. This information includes not only the original road and / or lane data, but may also involve processed data, such as the correspondence between the specific direction of the intersection and the road or lane, the correspondence between the intersection boundary or local boundary and the road or lane, and the number of roads and / or lanes connected to the intersection and the specified driving direction. In step S3, taking into account the road and / or lane information, the perception range of the vehicle is compared with the intersection range to check whether the perception range of the vehicle meets the preset conditions: when the vehicle has not yet entered the intersection, the perception range of the vehicle can be extended to the road and / or lane to be entered after the vehicle passes through the intersection.

[0060] "Considering the road and / or lane information" means, for example, that the comparison process needs to be performed according to the direction of a specific road and / or lane or the expected trajectory with it as the target, rather than randomly comparing two geometric figures in any direction. For example, in the scenario of going straight through an intersection, the goal is to enter the straight lane facing the road in front after passing the intersection. At this time, the comparison process needs to be performed along the direction of the straight lane or the expected trajectory related to it. Similarly, in the right turn or left turn scenario, the comparison process needs to be performed along the direction or expected trajectory of the right turn or left turn lane. "The vehicle has not entered the intersection" means that any part of the vehicle body is outside the intersection range. In order to evaluate the worst case, the above comparison can be performed when the vehicle just reaches the boundary of the intersection range but has not entered the intersection range. For example, when the vehicle stops at the stop line or zebra crossing before the intersection, this situation is met. For another example, when the vehicle is still in the straight lane, right turn lane or left turn lane before the intersection, this situation is also met. "The vehicle passes the intersection" means that the vehicle has completely driven through the intersection and entered the opposite lane.

[0061] In one embodiment, the comparison can be performed by means of simulation, and it can be checked whether the perception range of the vehicle meets the preset conditions. Specifically, the perception range of the vehicle, the intersection range, and the road and / or lane information can be imported into the simulation platform. In the simulation platform, a detailed vehicle model is created. This model is built according to the actual size and shape of the actual vehicle and is equipped with perception devices such as cameras, radars, and lidars. The parameter configurations of these devices match the performance of the real devices. Then, based on the imported intersection range, the geographical fence area of ​​the intersection is marked in the simulation environment. By setting different simulation scenarios, a variety of possible lane positions and vehicle head orientations before the vehicle approaches the intersection, as well as the entire process of the vehicle approaching, passing, and leaving the intersection can be simulated. Under each simulated vehicle posture, the perception range of the vehicle is compared with the intersection range. Exemplarily, an overlapping area between the vehicle's perception range and the intersection range can be created. By evaluating the continuity of the overlapping area between the vehicle's current position and the specific geometric boundary of the intersection, it can be determined whether the vehicle's perception range meets the preset conditions. This will be combined with the following FIG. 2A to FIG. 2F Further elaboration.

[0062] In another embodiment, the comparison can also be performed by means of real vehicle testing. Specifically, a marker or reflective object can be pre-set at the exit road of the actual intersection (or at the specified geometric boundary) to define the intersection range. Then, a test vehicle (or fleet) with a specific perception range is stopped in different postures before the intersection, and the sensor data collected by the vehicle-mounted perception system of the vehicle about the determined point E at the intersection is obtained in real time. The determined point E is outside the intersection range and is on the road and / or lane to be entered after the vehicle passes through the intersection. By analyzing the signal strength, reflectivity and / or other measurable properties of the sensor data, it can be determined whether the determined point E is within the perception range of the vehicle. For example, if the number or density of the point cloud data collected about the determined point E exceeds a preset threshold, it can be considered that the perception range of the vehicle can cover point E. For another example, if the clarity of the image taken by the vehicle about point E reaches a predetermined standard, it can also be considered that the perception range meets the preset conditions.

[0063] In another embodiment, the occlusion of the vehicle's perception range caused by static obstacle information, slope and / or road coverage relationship can also be considered in step S3. For example, the three-dimensional features extracted from the high-precision map or the standard-precision map can be integrated into the simulation platform, and it can be analyzed whether specific three-dimensional features will cause occlusion to the vehicle's perception range. Such occlusion may cause the vehicle to be unable to perceive the road and / or lane to be entered on the opposite side of the intersection in advance. In this way, the perception ability of the vehicle in the actual traffic environment can be more accurately evaluated, ensuring the accuracy of the evaluation of the high-precision map data requirements in a complex road environment.

[0064] In another embodiment, the preset condition may further include: the vehicle's perception range covers all lanes of the road to be entered in width. This not only ensures that the vehicle can see at least one lane to be entered in advance, but also ensures that the vehicle can perceive all lanes to be entered in advance, so that all drivable paths in the intersection can be provided to the vehicle even in the absence of high-precision map data.

[0065] In another embodiment, the preset conditions may further include: for all potential traffic plans of the vehicle at the intersection, before the vehicle enters the intersection, the perception range of the vehicle can be extended to the road and / or lane to be entered after the vehicle passes the intersection. This means that no matter whether the vehicle intends to go straight, turn right or turn left, before the vehicle enters the intersection, the perception range of the vehicle should be able to cover the corresponding road and / or lane to be entered. Additionally or alternatively, no matter from which direction the vehicle enters the intersection, its perception range can cover the corresponding road and / or lane to be entered.

[0066] In another embodiment, the preset conditions may further include: during the period when the vehicle is driving within the intersection, the perception range of the vehicle is always able to extend to the road and / or lane to be entered after the vehicle passes through the intersection. For example, when the vehicle turns left or right, the direction and coverage area of ​​the perception range will change as the vehicle turns. By ensuring that the road and / or lane to be entered is always within the perception range of the vehicle, the safety of the vehicle autonomously passing through the intersection can be improved.

[0067] In step S4, if the perception range of the vehicle meets the preset conditions, it is determined that high-precision map data is not required at the intersection; if the perception range of the vehicle does not meet the preset conditions, it is determined that high-precision map data is required at the intersection.

[0068] In one embodiment, such demand can be visualized in an appropriate manner. For example, the high-precision map data demand can be marked for each intersection on an autonomous driving map covering multiple intersections. Such visualization means enable autonomous driving developers, vehicle users, and map suppliers to intuitively understand the data demand for each intersection, thereby making more informed decisions.

[0069] In another embodiment, the evaluation results of the demand for high-precision map data can be stored in the local storage device of the vehicle and / or the original equipment manufacturer and / or the cloud server, thereby ensuring that the information can be easily retrieved and called when needed.

[0070] In another embodiment, the evaluation result in step S4 may be updated regularly in an additional step. Considering that the vehicle's hardware equipment may undergo upgrades and updates over time, and the road layout and planning of the intersection may also change, these changes may affect the vehicle's perception range and the determination of the intersection range. Therefore, the vehicle's perception range and intersection range can be re-determined after a certain period of time, and the high-precision map data requirements can be re-evaluated based on the comparison of these two ranges. If the new evaluation result is inconsistent with the previous evaluation result, the new evaluation result can be used to update the original evaluation result to ensure that the evaluation of the high-precision map requirements is always kept up to date to adapt to changes in vehicle hardware and road environment.

[0071] FIG. 2A to FIG. 2F A schematic diagram is shown for checking whether the perception range of a vehicle meets a preset condition by comparing the perception range of the vehicle with the intersection range in different scenarios. In the intersection scenario shown, intersection 2 consists of roads in four directions, each of which is connected to an exit road and an entrance road.

[0072] FIG. 2A to FIG. 2D The diagram shows a scenario in which a vehicle 1 is expected to go straight through an intersection 2. In this example scenario, the vehicle 1 is located in the straight lane and is about to enter the intersection 2. Four different examples of vehicle perception ranges 10 are depicted in the figure.

[0073] exist Figure 2A In the first case shown, the perception range 10 of the vehicle 1 only extends to the center of the intersection 2 and fails to reach the opposite road and / or lane to be entered. This means that without the assistance of a high-precision map, the vehicle 2 cannot obtain enough information to safely pass through the intersection 2.

[0074] exist Figure 2B , Figure 2C and Figure 2D In the other three cases shown, the perception range 10 of vehicle 1 can extend from its current position to the opposite lane, forming a continuous overlapping area 11 with the intersection range 20. The overlapping area 11 continuously covers the local geometric boundary of intersection 2 from the current position of vehicle 1, indicating that before entering intersection 2, vehicle 1 can use its perception system to "observe" the road and / or lane to be entered after passing intersection 2 in advance. In these cases, even without a high-precision map, vehicle 1 can rely on its own perception system and, if necessary, combine with a standard precision map to autonomously pass through the intersection.

[0075] Specifically, Figure 2BThe second case shown represents the most unfavorable case (the so-called "worst-case"). In this case, the edge of the perception range 10 of the vehicle 1 is completely aligned with the local geometric boundary of the intersection range 20, that is, the fan-shaped visible area in front of the vehicle 1 just touches the boundary of the intersection range 20 along the expected driving direction, for example, reaches a part of the lane of the road to be entered. This ensures that the perception range 10 of the vehicle 1 just covers the path required for it to go straight through the intersection 2.

[0076] exist Figure 2C In the third case shown, the front viewing angle of the vehicle 1 is wider. In this case, the perception range 10 of the vehicle 1 not only extends to the road to be entered, but also covers all lanes of the road in width. This perception range 10 enables the vehicle 1 to monitor multiple possible driving paths in advance, increasing the reliability when passing through the intersection 2.

[0077] exist Figure 2D In the fourth case shown, the perception range 10 of the vehicle 1 exceeds the geometric boundary of the intersection range 20, providing a larger safety margin, further ensuring that the vehicle 1 can identify the road and / or lane to be entered in advance.

[0078] exist Figure 2E , a situation in which vehicle 1 is located in the right-turn lane and is about to turn right through intersection 2 is shown. At this time, the overlapping area 11 formed between the perception range 10 of vehicle 1 and the intersection range 20 continuously covers from the current position of vehicle 1 to the local geometric boundary on the right side of intersection 2. This continuous coverage indicates that vehicle 1 can use its perception system to "observe" the road and / or lane to be entered after the right turn in advance before the actual right turn action occurs, thereby grasping the right turn trajectory at intersection 2 in advance. Through the simulation of the simulation platform, different postures of vehicle 1 in the right-turn lane can be simulated, including the state of the front of the vehicle slightly turning to the right, etc., to verify whether vehicle 1 can identify the road and / or lane to be entered in advance in various postures.

[0079] Figure 2F The figure shows a situation where vehicle 1 is in the left-turn lane and is about to turn left through intersection 2. At this time, the perception range 10 of vehicle 1 extends from the current position of the left-turn lane to the lane that vehicle 1 will enter after turning left. The overlapping area 11 formed by the perception range 10 of vehicle 1 and the intersection range 20 also covers continuously from the current position of vehicle 1 to the local geometric boundary on the left side of intersection 2. This shows that before implementing the left turn, the perception system can be used to identify the road and / or lane to be entered after the left turn in advance.

[0080] In an embodiment not shown, Figure 2FIn the illustrated situation, the simulation platform can also be used to simulate the state of vehicle 1 on the left-turn waiting line, as well as different postures of the vehicle 1 turning left while driving in intersection 2. These simulations help verify whether the vehicle perception system can always maintain the road and / or lane to be entered extending to the left side of intersection 2, ensuring that vehicle 1 can safely complete the left turn.

[0081] It should be noted that FIG. 2A to FIG. 2F The shapes and sizes of the perception range 10 and the intersection range 20 of the vehicle 1 shown in the figure are only exemplary. In actual applications, the exact shapes and sizes of these ranges may vary depending on the hardware configuration of the perception system carried by the vehicle and the method selected to construct the intersection range 20.

[0082] It should also be noted that although Figures 2A to 2F Only the case where a vehicle enters the intersection from an entrance road in a specific direction is shown, but in actual applications, vehicles may also enter the intersection from entrance roads in other directions. Therefore, when evaluating the high-precision map data requirements of autonomous vehicles, the case where vehicle 1 enters intersection 2 from various possible entrance roads can also be considered.

[0083] Figure 3A and Figure 3B A schematic diagram is shown of determining an intersection range 20 of an intersection 2 based on a high-precision map according to an exemplary embodiment of the present application.

[0084] exist Figure 3A In the illustrated embodiment, the stop line 21 associated with the intersection 2 can be extracted from the high-precision map. In addition, road boundary information associated with the intersection 2 can also be extracted from the high-precision map, which includes, for example, road boundary lines 23 and 25 of each road connected to the intersection.

[0085] In the roads in each direction, usually only the entrance road L1 of the intersection 2 is provided with a stop line 21, while the exit road L2 is not provided. For the entrance road L1 in each direction, the intersection points A1 and A2 of the stop line 21 and the road boundary lines 23 and 24 already exist. By extending the stop line 21 to intersect with the outer road boundary line 25 of the exit road L2, an additional intersection point A3 can be constructed. Repeating this step for the roads in each direction, the intersection points A1, A2, and A3 of the stop line 21 and its extension line 22 with the road boundary lines 23, 24, and 25 can be obtained in all directions.

[0086] Then, an appropriate graphics algorithm (such as a convex hull algorithm or a minimum bounding rectangle algorithm) can be used to connect all these intersections in sequence to form a closed figure. The constructed closed figure is, for example, a minimum polygon surrounding all intersections of the intersection 2, which can be used to approximately represent the intersection range 20.

[0087] exist Figure 3AIn the illustrated embodiment, each vertex of the closed figure corresponds to the intersection formed by the stop line at the intersection and its extension line and the road boundary line. In other embodiments not shown, this polygon can also be slightly expanded outwards according to actual needs so that its vertices are not strictly limited to the intersection points, but can be the adjacent points of these intersection points.

[0088] exist Figure 3B In the illustrated embodiment, a zebra crossing 27 associated with the intersection 2 and road boundary lines 23, 25 of each road connected to the intersection 2 can be extracted from the high-precision map. The zebra crossing 27 can be defined by its two endpoints A1, A2, and these endpoints A1, A2 can be used as feature points of the intersection 2. For example, a point (e.g., a midpoint) on these line segments can be taken as the endpoints A1, A2. In another embodiment, the intersection formed by the zebra crossing 27 and the road boundary lines 23, 25 or their extensions can also be determined as the feature point of the intersection 2.

[0089] For the above situation, the generated feature points can be connected in a certain order to form a closed figure surrounding the intersection. This can also be achieved with the help of a graphics algorithm such as a convex hull algorithm or a minimum enclosing rectangle algorithm. Figure 3A Similar to the illustrated case, the closed figure thus formed can be defined as the intersection range 20 of the intersection.

[0090] Figure 4A A schematic diagram is shown of determining the intersection range of an intersection based on a precise map according to an exemplary embodiment of the present application.

[0091] In a precise map, a road network consists of nodes and road segments connecting these nodes. Nodes usually represent intersections or branch points of roads. By analyzing the locations of these nodes, the geometry and scope of the intersections 2 can be inferred.

[0092] like Figure 4A As shown, first, all nodes A1, A2, B1, B2, C1, C2, D1, and D2 related to intersection 2 can be extracted from the precision map. These nodes represent the starting point and end point of the road segment at intersection 2. Figure 4A In the illustrated accurate map of the mapping specification, the nodes are located outside the actual intersection 2 and correspond, for example, to the stop line position of the intersection 2. At the intersection 2, the road segment along the same driving direction is interrupted by two nodes 9 (for example, A1 and C2).

[0093] Then, these nodes A1, A2, B1, B2, C1, C2, D1, and D2 can be connected in sequence according to the determined connection order to form a closed figure surrounding the intersection. In one embodiment, the connection order of the nodes can be determined in a clockwise or counterclockwise order. In another embodiment, the geometric center of the intersection can also be determined based on the node connection line, and the connection order of the nodes can be determined based on the angle between each node and the geometric center connection line and the reference direction.

[0094] like Figure 4A As shown in the example, the roads in each direction are marked by two nodes. For example, the roads 41 and 42 below the intersection 2 are marked by nodes A1 and A2, respectively, where A1 represents the end point of the entrance road 41 and A2 represents the starting point of the exit road 42. Similarly, the left road is marked by nodes B1 and B2, the upper road is marked by nodes C1 and C2, and the right road is marked by nodes D1 and D2. In this way, a total of eight nodes related to the intersection 2 are obtained.

[0095] In order to determine the connection order of these nodes and construct a closed figure around the intersection, the geometric center point O of these nodes may be calculated first. Then, the connection order is determined according to the angle between the line connecting each node to the geometric center point O and a common reference direction. In the illustrated embodiment, the connection order of the nodes may be, for example, A1→A2→B1→B2→C1→C2→D1→D2→A1, and such an order ensures that the polygon is closed in a clockwise or counterclockwise direction, and the closed figure thus formed is determined as the intersection range 20 of the intersection 2.

[0096] Figure 4B A schematic diagram of determining the intersection range of an intersection based on a precise map according to another exemplary embodiment of the present application is shown.

[0097] Similar to Figure 4A The method described in Figure 4B In the example, a closed figure 20 around the intersection 2 is constructed based on the precise map. In addition, Figure 4B Further, for each geometric boundary or geometric boundary segment (also referred to as a local geometric boundary), a corresponding relationship with each exit road at the intersection 2 is marked, and a corresponding relationship with each entrance road is also optionally marked.

[0098] Specifically, the characteristics of each node can be determined by analyzing the road segments connected to each node in the precision map to determine whether these nodes are associated with entry roads or exit roads. For example, node A1 represents the end point of entry road L1, while node A2 represents the starting point of exit road. Based on this, the right half of the geometric boundary of the line between nodes A1 and A2 can be defined as corresponding to the entry road and assigned the number g1, while the left half of the geometric boundary is defined as corresponding to the exit road and assigned the number g2.

[0099] In this way, all geometric boundaries can be assigned numbers. Figure 4B In the figure, the geometric boundary segments corresponding to the entrance roads are assigned odd numbers g1, g3, g5, g7, and the geometric boundary segments corresponding to the exit roads are assigned even numbers g2, g4, g6, g8. In this way, not only is the intersection range 20 constructed by the closed graph, but also the corresponding relationship between each geometric boundary and a specific exit road (and entrance road) is incorporated into the definition of the intersection range 20.

[0100] For example, if a vehicle is about to enter intersection 2 from entrance road L1, and plans to enter exit road L6 after leaving the intersection, then before the vehicle enters the intersection, if the vehicle's perception range can extend continuously from geometric boundary segment g1 to geometric boundary segment g6, it can be determined that the vehicle's perception range meets the preset conditions.

[0101] In an embodiment not shown, the correspondence between each geometric boundary of a closed figure and the exit road can also be marked based on a high-precision map. For example, the relationship between each geometric boundary and the stop line can be analyzed, and it can be determined whether the geometric boundary is composed of a stop line or an extension of the stop line. If a geometric boundary is composed of a stop line, this usually means that it corresponds to an entry road; and if a geometric boundary is composed of an extension of a stop line, it means that it corresponds to an exit road. In addition, this correspondence can also be determined by analyzing the relative position of the geometric boundary on the high-precision map with the entry road and the exit road.

[0102] Figure 4C A schematic diagram showing a precision map that does not meet cartographic specifications.

[0103] exist Figure 4C In the precise map of the illustrated mapping specification, internal nodes A1, A2, A3, and A4 are used to model the intersection 2. These nodes A1, A2, A3, and A4 are located inside the intersection and are mainly used to abstractly represent the intersection of roads in different directions, rather than representing the actual starting and ending points of the road segments at the intersection. In this mapping specification, although these nodes identify the intersection 2 on the map, their positions do not directly correspond to the actual positions of the stop line or zebra crossing of the intersection 2.

[0104] Since this precise map uses a highly abstract and simplified representation method, its intersection modeling method fails to provide sufficient details, so it is difficult to accurately restore the approximate scope of intersection 2 based on this precise map.

[0105] Figure 5 A flow chart of a method for evaluating the high-precision map data requirements of an autonomous driving vehicle according to another exemplary embodiment of the present application is shown. In this embodiment, Figure 1 The method shown also comprises an additional step S5, which further comprises sub-steps S51 to S55.

[0106] exist Figure 5 In the illustrated embodiment, for example, in step S3, all intersection ranges within a preset area and known ranges of vehicles are compared in advance with the aid of simulation tests, and in step S4, based on the simulation test results, it is concluded that some intersections within the preset area do not require high-precision map data.

[0107] Next, in step S5, these simulation test results can be further verified in combination with a real vehicle test process.

[0108] Specifically, after step S4, sub-step S51 is first executed, in which a test vehicle with the sensing range is deployed to attempt autonomous driving at least for some intersections determined in step S4 as not requiring high-precision map data. For example, the autonomous driving parameters of the vehicle may be configured, and the vehicle may be allowed to attempt to autonomously pass through the intersection.

[0109] In sub-step S52, it is checked whether the test vehicle can successfully drive through the said part of the intersection. For example, if the vehicle can successfully cross the intersection from the current position and reach the road and / or lane to be entered without any safety incident or need for manual intervention during the process, it can be considered that the perception range of the vehicle at this part of the intersection is sufficient to meet the needs of autonomous driving without relying on high-precision map data.

[0110] If the test vehicle can successfully drive through these partial intersections automatically, then in sub-step S53, it is verified that high-precision map data is indeed not required at these partial intersections.

[0111] If the test vehicle fails to successfully drive through these partial intersections, then in sub-step S54, it is verified that high-precision map data is actually required at these partial intersections. In this case, the verification result of step S54 is inconsistent with the evaluation result of the simulation test in step S4. Therefore, the modeling parameters or measurement parameters of the intersection range and / or the perception range of the on-board perception system of the partial intersection can be adjusted in sub-step S55. After the adjustment, it is necessary to re-execute steps S3 and S4 based on the updated perception range and intersection range to re-evaluate the demand for high-precision map data of the autonomous driving vehicle. This process will be repeated until the result of the re-evaluation is consistent with the result of the actual test verification, and then the parameter adjustment can be considered reasonable and effective. Once it is confirmed that the parameter adjustment is effective, these adjusted parameters can be used to continue the subsequent intersection range determination, perception range acquisition, comparison and evaluation work. Through this iterative verification method, the accuracy and reliability of the evaluation results can be ensured, thereby providing a more accurate high-precision map data demand assessment for the autonomous driving system.

[0112] Figure 6 A flow chart of a method for high-precision map data collection according to an exemplary embodiment of the present application is shown. The method exemplarily includes step S601 and step S602.

[0113] In step S601, the high-precision map data requirements of the preset area are obtained. The high-precision map data requirements are based on Figure 1 Determined by the method described.

[0114] For example, high-precision map data requirements for a certain geographical range can be retrieved from the local storage device or cloud server of the vehicle or OEM. These requirements usually exist in the form of an autonomous driving map with annotated information, which not only includes the location information of each intersection in the area, but also clearly marks which intersections require high-precision map data support.

[0115] In step S602, high-precision map data is collected for at least one intersection in the preset area that requires high-precision map data, and high-precision map data is not collected for at least another intersection in the preset area that does not require high-precision map data.

[0116] According to the acquisition requirements, all the intersections that need to collect HD map data can be screened out, and a HD map collection list can be created based on their location or number. This list can then be sent to the map supplier so that the required HD map data can be purchased only for the intersections included in the list or HD map maintenance services can be subscribed regularly (such as monthly or annually). Such a process optimizes data collection efficiency and ensures that resources are reasonably allocated to areas that really need HD map support.

[0117] In an optional embodiment, although high-precision map data is not collected for those intersections in the preset area that are determined not to require high-precision map data, the vehicle's own perception system can be used to dynamically construct local map data for these areas. This method allows the vehicle to collect and generate map information of intersections in real time through its onboard sensor technology during driving, thereby supplementing or replacing traditional high-precision map data. In this way, it can not only reduce the dependence on pre-collected high-precision maps, but also improve the real-time and accuracy of map data.

[0118] Figure 7 A flow chart of a method for providing a light map according to an exemplary embodiment of the present application is shown. The method exemplarily includes step S701 and step S702.

[0119] In step S701, according to Figure 6 The method shown collects high-precision map data for a preset area. For example, if the preset area includes 10 intersections, high-precision map data can be collected for 7 of the intersections, while high-precision map data is not collected for the remaining 3 intersections.

[0120] In step S702, a light map is created for a preset area, wherein the light map includes high-precision map data of at least one intersection in the preset area and does not include high-precision map data of at least another intersection in the preset area.

[0121] In this application, "light map" is defined as an optimization solution for the scale of high-precision map data. Light map achieves lightweight map data by selectively including high-precision map data of key intersections while omitting high-precision map data of non-key intersections. For a preset area, the fewer intersections where high-precision map data is collected, the lighter the autonomous driving map of the entire preset area will be, and the "lighter" the map will be.

[0122] For example, if HD map data is collected for 7 of the 10 intersections in the preset area, the generated light map will only contain the HD map data of these 7 intersections, and for the remaining 3 intersections, it will not contain any map data, or may only contain basic standard map data. The light map solution enables the autonomous driving system to use the minimum HD map, thereby reducing or getting rid of the autonomous driving vehicle's dependence on HD maps.

[0123] Figure 8 A flow chart of a method for automatic driving of a vehicle according to an exemplary embodiment of the present application is shown. The method exemplarily includes steps S801 to S803.

[0124] In step S801, the high-precision map data requirements of the preset area are obtained. The high-precision map data requirements are based on Figure 1As described above, the high-precision map data requirements can be stored in the form of a trigger list together with the corresponding intersection number or location in the vehicle and / or the original equipment manufacturer's local or cloud server.

[0125] In step S802, when the vehicle travels to a preset distance from at least one intersection in a preset area that requires high-precision map data, the download of high-precision map data for the at least one intersection is automatically triggered. For example, the vehicle will continuously compare its current position with the intersection position in the trigger list, and determine in real time whether the distance between the two is less than the set threshold. For example, a circular area with the vehicle as the center and a predetermined distance threshold (such as 3 kilometers) as the radius can be set. Once the position of any intersection in the trigger list enters this circular area, the system will automatically start downloading the high-precision map data related to the intersection. This distance threshold is adjustable, which affects the amount of data preloaded in the map data, and is related to the specific functional requirements of the autonomous driving vehicle.

[0126] In step S803, the vehicle is controlled to automatically drive through the at least one intersection based on the downloaded high-precision map data. For intersections where high-precision map data is not collected, the vehicle will rely on its own perception system to navigate, or, when necessary, use standard high-precision map data to assist in automatic driving to ensure safe passage through these intersections.

[0127] Fig. 9 A flow chart of a method for automatic driving of a vehicle according to another exemplary embodiment of the present application is shown. The method exemplarily includes step S901 and step S902.

[0128] In step S901, obtain Figure 7 The light map provided by the method shown. The light map can be pre-stored in the local storage device of the vehicle, so that the vehicle can access the local storage device at any time during driving to view the light map. In addition, the light map can also be stored in a cloud server so that the vehicle can download the light map in slices or in its entirety when needed.

[0129] In step S902, the automatic driving of the vehicle in the preset area is controlled based on the light map. For example, for at least one intersection with high-precision map data in the light map, the vehicle can be controlled to automatically drive through the at least one intersection based on the high-precision map data, and for at least another intersection without high-precision map data in the light map, the vehicle can be controlled to automatically drive through the at least another intersection based on the vehicle perception system.

[0130] Fig.10 A block diagram of an apparatus 100 for evaluating high-precision map data requirements of an autonomous driving vehicle according to an exemplary embodiment of the present application is shown.

[0131] The apparatus 100 may include a memory 101 and a processor 102. The memory 101 may store executable instructions. The processor 102 may execute the executable instructions stored or encoded in the memory 101, thereby implementing the above-mentioned Figure 1 and Figure 5 Although not described in Fig.10 However, those skilled in the art will appreciate that the device 100 may also include various other components, such as various communication modules, buses, and possible user interfaces.

[0132] The specific operation and function implementation process of the above-mentioned device 100 has been combined with Figure 1 and Figure 5 The method shown has been described in detail and will not be repeated here.

[0133] Fig.11 A block diagram of an apparatus 110 for collecting high-precision map data according to an exemplary embodiment of the present application is shown.

[0134] The apparatus 110 may include a memory 111 and a processor 112. The memory 111 may store executable instructions. The processor 112 may execute the executable instructions stored or encoded in the memory 111, thereby implementing the above-mentioned combination Figure 6 Although not described in Fig.11 However, those skilled in the art will appreciate that the device 110 may also include various other components, such as various communication modules, buses, and possible user interfaces.

[0135] The specific operation and function implementation process of the above-mentioned device 110 has been combined with Figure 6 The method shown has been described in detail and will not be repeated here.

[0136] Fig.12 A block diagram of an apparatus for providing a light image according to an exemplary embodiment of the present application is shown.

[0137] The apparatus 120 may include a memory 121 and a processor 122. The memory 121 may store executable instructions. The processor 122 may execute the executable instructions stored or encoded in the memory 121, thereby implementing the above-mentioned combination Figure 7 Although not described in Fig.12 However, those skilled in the art will appreciate that the device 120 may also include various other components, such as various communication modules, buses, and possible user interfaces.

[0138] The specific operation and function implementation process of the above-mentioned device 120 has been combined with Figure 7 The method shown has been described in detail and will not be repeated here.

[0139] Fig.13 A block diagram of an automatic driving device for a vehicle according to an exemplary embodiment of the present application is shown.

[0140] The automatic driving device 130 may include a memory 131 and a processor 132. The memory 131 may store executable instructions. The processor 132 may execute the executable instructions stored or encoded in the memory 131, thereby realizing the above combination Figure 8 and Fig. 9 Although not described in Fig.13 However, those skilled in the art will appreciate that the automatic driving device 130 may also include various other components, such as various communication modules, buses, and possible user interfaces.

[0141] The specific operation and function implementation process of the above-mentioned automatic driving device 130 has been combined with Figure 8 and Fig. 9 The method shown has been described in detail and will not be repeated here.

[0142] It is understood that the methods of the various embodiments of the present application can be implemented by computer programs / software. These software can be loaded into the working memory of the processor and used to execute the methods according to the various embodiments of the present application when running.

[0143] According to another embodiment of the present application, a computer program product is provided, such as a machine (e.g., computer) readable medium, such as a CD-ROM, which includes a computer program code, which, when executed, causes a computer or a processor to perform a method according to each embodiment of the present application. The machine readable medium is, for example, an optical storage medium or a solid-state medium supplied with other hardware or as part of other hardware.

[0144] Although the specific embodiments of the present application are described in detail herein, they are provided for the purpose of explanation only and should not be considered to limit the scope of the present application. Various substitutions, changes and modifications may be conceived without departing from the spirit and scope of the present application.

Claims

1. A method for evaluating the high-precision map data requirements of an autonomous driving vehicle, the method comprising the following steps: Step S1: obtaining a sensing range (10) of an onboard sensing system of a vehicle (1); Step S2: Obtaining the intersection range (20) of the intersection (2) and the road and / or lane information connected to the intersection (2); Step S3: considering the road and / or lane information, comparing the perception range (10) of the vehicle (1) with the intersection range (20) to check whether the perception range (10) of the vehicle (1) meets a preset condition: before the vehicle (1) enters the intersection (2), the perception range (10) of the vehicle (1) can extend to the road and / or lane to be entered after the vehicle (1) passes through the intersection (2); and Step S4: If the perception range (10) of the vehicle (1) satisfies the preset condition, it is determined that high-precision map data is not required at the intersection (2); if the perception range (10) of the vehicle (1) does not satisfy the preset condition, it is determined that high-precision map data is required at the intersection (2).

2. The method according to claim 1, wherein: The preset conditions further include: The sensing range (10) of the vehicle (1) covers all lanes of the road to be entered in width; and / or For all potential passage plans of the vehicle (1) at the intersection (20), before the vehicle (1) enters the intersection (2), the perception range (10) of the vehicle (1) can be extended to the road and / or lane to be entered after the vehicle (1) passes through the intersection (2); and / or While the vehicle (1) is traveling within the intersection range (20), the perception range (10) of the vehicle (1) is always able to extend to the road and / or lane to be entered after the vehicle (1) passes through the intersection (2).

3. The method according to claim 1 or 2, wherein: In step S3, it is checked whether the sensing range (10) of the vehicle (1) satisfies a preset condition by: Obtaining an overlapping area (11) between a perception range (10) of the vehicle (1) and an intersection range (20) when the vehicle (1) has not yet entered the intersection (2); Checking whether the overlapping area (11) extends continuously between the current position of the vehicle (1) and a local geometric boundary of the intersection range (20), wherein the local geometric boundary is directly connected to the road and / or lane to be entered by the vehicle (1) after passing through the intersection (2); and If the overlapping area (11) extends continuously between the current position of the vehicle (1) and the local geometric boundary of the intersection range (20), it is confirmed that the perception range (10) of the vehicle (1) meets the preset conditions.

4. The method according to any one of claims 1 to 3, wherein: Step S2 includes determining an intersection range (20) of the intersection (2) based on the high-precision map, wherein: Extracting a stop line (21) and / or a zebra crossing (27) associated with the intersection (2) from a high-precision map; Extracting road boundary information related to the intersection (2) from the high-precision map, the road boundary information at least including a road boundary line (23) of a road connected to the intersection (2); Generating feature points related to the intersection (2) based on the stop line (21) and / or the zebra crossing (27) and the road boundary information; Connecting the characteristic points in sequence according to a determined connection order to form a closed figure surrounding the intersection (2); and The closed figure is determined as the intersection range (20) of the intersection (2).

5. The method according to claim 4, wherein: The characteristic points associated with the intersection (2) are determined in the following manner: Extending the stop line (21) so that its extension line intersects with the road boundary line (23), and determining the intersection point of the stop line (21) and its extension line with the road boundary line (23) as a characteristic point related to the intersection (2); and / or The end point of the zebra crossing (27) or the intersection point of the zebra crossing (27) and the road boundary line (23) is determined as a feature point related to the intersection (2).

6. The method according to any one of claims 1 to 5, wherein: Step S2 includes determining the intersection range (20) of the intersection (2) based on the precise map, wherein: Extracting all nodes related to the intersection (2) from the precise map, wherein the nodes represent the starting point and the end point of the road segment in the precise map at the intersection (20); Connecting the nodes in sequence according to the determined connection order to form a closed graph surrounding the intersection (2); and The closed figure is determined as the intersection range (20) of the intersection (2).

7. The method according to any one of claims 1 to 6, wherein: In step S2, only the precise map that meets the following mapping specifications is used to determine the intersection range (20) of the intersection (2): Under the mapping specification, the nodes associated with the intersection (2) are located outside the intersection (2), and / or the nodes associated with the intersection (2) are located at the position where the corresponding stop line (21) or zebra crossing (27) is located.

8. The method according to any one of claims 1 to 7, wherein: Step S2 further comprises: After forming a closed figure surrounding the intersection (2), determining the correspondence between the geometric boundary or the local geometric boundary of the closed figure and each exit road at the intersection (2); marking the corresponding relationship for the geometric boundary or the local geometric boundary of the closed figure; The closed figure with the marked correspondence is determined as the intersection range (20) of the intersection (2).

9. The method according to any one of claims 1 to 8, wherein: Step S2 further comprises: obtaining static obstacle information, slope and / or road coverage relationship at the intersection (2) from the high-precision map and / or the standard-precision map; and In step S3, taking into account the occlusion of the perception range (10) of the vehicle (1) caused by static obstacle information, slope and / or road coverage, the perception range (10) of the vehicle (1) and the intersection range (20) are compared, and it is checked whether the perception range (10) of the vehicle (1) meets the preset conditions.

10. The method according to any one of claims 1 to 9, wherein: Step S3 further comprises: Importing the perception range (10) of the vehicle (1), the intersection range (20) of the intersection (2), and information of roads and / or lanes connected to the intersection (2) into the simulation platform; Simulating different postures of a vehicle (1) at a plurality of intersections by means of a simulation platform, wherein the different postures include different lane positions and / or vehicle head directions; and For each intersection, the perception range (10) of the vehicle (1) is compared with the intersection range (20) at each simulated posture.

11. The method according to any one of claims 1 to 10, wherein: Step S3 further comprises: The vehicle (1) is stopped in front of the intersection in different postures, and sensor data collected by the vehicle-mounted sensing system of the vehicle (1) about a determined point at the intersection (20) is acquired in real time, wherein the determined point is outside the intersection range (20) and is located on the road and / or lane to be entered after the vehicle (1) passes through the intersection (2); analyzing the sensor data for signal strength, reflectivity, and / or other measurable attributes; Based on the result of the analysis, it is checked whether the determined point is located within the perception range (10) of the vehicle (1).

12. The method according to any one of claims 1 to 11, wherein: The method further comprises step S5: At least for the portion of intersections determined in step S4 as not requiring high-precision map data, deploying a test vehicle with the perception range (10) to attempt autonomous driving; Checking whether the test vehicle can successfully drive autonomously through the portion of the intersection; as well as If the test vehicle is able to successfully drive automatically through the portion of the intersection, it is verified that high-precision map data is not required at the portion of the intersection (20).

13. The method according to claim 12, wherein: Step S5 further comprises: If the verification result is inconsistent with the result determined in step S4, the modeling parameters or measurement parameters of the intersection range (20) of the partial intersection (2) and / or the perception range (10) of the on-board perception system are adjusted, and steps S3 and S4 are re-executed until the re-determined result is consistent with the verification result.

14. The method according to claim 13, wherein: Step S4 further comprises: On an autonomous driving map covering multiple intersections (2), marking the high-precision map data requirements for each intersection; and / or The results of the determined high-precision map data requirements are stored in a local storage device and / or a cloud server of the vehicle (1) and / or the original equipment manufacturer.

15. A method for collecting high-precision map data, wherein: The method comprises the following steps: Obtaining high-precision map data requirements for a preset area, wherein the high-precision map data requirements are determined according to any one of the methods of claims 1 to 14; and For at least one intersection in the preset area that requires high-precision map data, high-precision map data is collected; for at least another intersection in the preset area that does not require high-precision map data, high-precision map data is not collected.

16. A method for providing a light image, wherein: The method The following steps are involved: The method according to claim 15 collects high-precision map data for a preset area; A light map is created for a preset area, the light map including high-precision map data of at least one intersection (2) in the preset area and excluding high-precision map data of at least another intersection (2) in the preset area.

17. A method for autonomous driving of a vehicle (1), the method comprising the following steps: Obtaining high-precision map data requirements for a preset area, wherein the high-precision map data requirements are determined according to any one of the methods of claims 1 to 14; When the vehicle (1) travels to a preset distance from at least one intersection in a preset area requiring high-precision map data, the downloading of high-precision map data of the at least one intersection (2) is automatically triggered; as well as Based on the downloaded high-precision map data, the vehicle (1) is controlled to automatically pass through the at least one intersection.

18. A method for autonomous driving of a vehicle (1), the method comprising the following steps: Obtaining a light map provided by the method according to claim 16; as well as Based on the light map, the vehicle (1) is controlled to automatically drive within a preset area, wherein, for at least one intersection in the light map with high-precision map data, the vehicle (1) is controlled to automatically drive through the at least one intersection based on the high-precision map data, and for at least another intersection in the light map without high-precision map data, the vehicle (1) is controlled to automatically drive through the at least another intersection based on the vehicle-mounted perception system.

19. A device (100) for evaluating the high-precision map data requirements of an autonomous driving vehicle, the device (100) comprising a memory (101) and a processor (102), the memory (101) storing computer program instructions, and when the computer program instructions are executed by the processor (102), the processor (102) is capable of executing the method according to any one of claims 1 to 14.

20. A device (110) for collecting high-precision map data, the device (110) comprising a memory (111) and a processor (112), the memory (111) storing computer program instructions, and when the computer program instructions are executed by the processor (112), the processor (112) is capable of executing the method according to claim 15.

21. A device (120) for providing a light map, the device (120) comprising a memory (121) and a processor (122), the memory (121) storing computer program instructions, when the computer program instructions are executed by the processor (122), the processor (122) is capable of executing the method according to claim 16.

22. An automatic driving device (130) for a vehicle (1), the automatic driving device (130) comprising a memory (131) and a processor (132), the memory (131) storing computer program instructions, and when the computer program instructions are executed by the processor (132), the processor (132) is capable of executing the method according to claim 17 or 18.

23. A computer program product comprising computer program instructions, wherein: The computer program instructions, when executed by one or more processors, enable the one or more processors to perform a method according to any one of claims 1 to 18.

Images