Method and apparatus for generating lane level map data

Abstract

The invention provides a method for generating lane-level map data, and the method comprises the following steps: S1, obtaining standard precise map data which comprises the geometric topology and road attributes of road lines; s2, generating a geometric topology of a road surface based on the geometric topology of the road line and the road attribute, wherein the geometric topology of the road surface at least comprises a road boundary line; s3, acquiring a predefined road element configuration rule, and generating a corresponding road element for geometric topology simulation of the road surface based on the predefined road element configuration rule; and S4, generating lane-level map data based on the geometric topology of the road surface and the corresponding road elements. The application also proposes a method for performing lane-level navigation, an apparatus for generating lane-level map data, an apparatus for performing lane-level navigation and a computer program product.

Description

Technical Field

[0001] This application relates to a method for generating lane-level map data, and also to a method for lane-level navigation, an apparatus for generating lane-level maps, an apparatus for lane-level navigation, and a computer program product. Background Technology

[0002] Due to the high cost of acquiring and the complexity of updating and maintaining high-precision map data, most map providers currently struggle to offer full-area high-precision map coverage across large geographical regions. In traditional navigation solutions, without high-precision map data, the system can only rely on lower-precision standard-precision map data to provide link-level navigation guidance. This relatively simplified navigation mode not only leads to a degraded user experience but may also cause vehicle errors due to overly generalized link-level guidance instructions. Furthermore, link-level navigation guidance severely restricts the development of autonomous driving functions, hindering their large-scale deployment.

[0003] Therefore, existing lane-level map generation solutions still have significant shortcomings. Summary of the Invention

[0004] The purpose of this application is to provide a method for generating lane-level map data, the method comprising the following steps: Step S1: Obtain high-precision map data, which includes the geometric topology of road lines and road attributes; Step S2: Generate the geometric topology of the road surface based on the geometric topology of the road lines and the road attributes. The geometric topology of the road surface includes at least the road boundary lines. Step S3: Obtain predefined road element configuration rules, and based on these rules, generate corresponding road elements for the geometric topology simulation of the road surface; and Step S4: Generate lane-level map data based on the geometric topology of the road surface and the corresponding road elements.

[0005] This application specifically includes the following technical concept: by fully utilizing extensive high-precision map data and combining it with predefined road element configuration rules, efficient and low-cost generation of lane-level map data is achieved. This method overcomes the technical bottleneck of limited coverage of high-precision maps and accurately reconstructs the geometric features and element details of the road surface without relying on vehicle perception data or deploying dedicated mapping vehicles to obtain road scene information. This significantly improves the accuracy and visualization effect of lane-level navigation guidance, providing a feasible low-cost solution for achieving full-domain lane-level navigation, enabling vehicles to obtain continuously available lane-level navigation services in any area, fundamentally improving the user's navigation experience.

[0006] In an exemplary embodiment, step S3 includes: determining a vectorization mode for road lines based on road attributes, wherein the vectorization mode includes unidirectional vectorized road lines and bidirectional vectorized road lines; and generating corresponding road elements based on road element configuration rules applicable under the determined vectorization mode. Thus, by distinguishing between unidirectional and bidirectional vectorized roads and adapting differentiated rules, accurate digital modeling of structural elements of roads at different functional levels in the real world is achieved, ensuring the standardization and accuracy of road element generation.

[0007] In an exemplary embodiment, step S3 includes: identifying the road scenario involved in the road surface based on the high-precision map data; and generating corresponding road elements based on the road element configuration rules applicable to the identified road scenario. This scenario-based rule dynamic invocation mechanism ensures that the generated road elements conform to both the physical structural characteristics of the road and the regulatory requirements of the specific traffic scenario without relying on the actual environmental perception data collection.

[0008] In an exemplary embodiment, identifying road scenarios based on the geometric topology and road attributes of road lines includes: identifying road scenarios involving the road surface based on high-precision map data, including: identifying adjacent intersection scenarios based on the spatial proximity between the road represented by the road line and the intersection represented by the intersection node it connects to; identifying consecutive intersection scenarios based on the length of the area between two intersections represented by the road line; identifying long curve scenarios and / or consecutive multi-curve scenarios based on the length and curvature information of the road line; identifying no-overtaking scenarios based on traffic sign information; identifying tunnel and / or bridge section scenarios based on tunnel and bridge sign information; identifying expressway / urban expressway or non-expressway / urban expressway scenarios based on the road function level of the road line; and / or, identifying ramp scenarios based on the road morphology attributes and road function level of the road line. This achieves accurate identification of various typical road scenarios, providing a key basis for the subsequent differentiated generation of road elements that conform to traffic rules.

[0009] In one exemplary embodiment, generating road elements includes determining the geometric position and style attributes of traffic markings and / or traffic signs; specifically, the traffic markings include lane lines, road center lines, stop lines, ground directional arrows, and / or guide strips, and the style attributes of the traffic markings include shape, line type, color, and / or size; the traffic signs include traffic signs, road boundary elements, and road dividers, and the style attributes of the traffic signs include type, material, and / or size. This constructs a complete road element attribute system, ensuring the uniformity of the generated traffic markings and signs in terms of geometric accuracy and style specifications.

[0010] In an exemplary embodiment, step S3 includes: for bidirectional vectorized straight road lines: if an adjacent intersection scenario, a continuous intersection scenario, or a tunnel and / or bridge segment scenario is identified, the road center line is set to a solid yellow line, and / or the lane lines are set to solid white lines; otherwise, the road center line is set to a dashed yellow line, and / or the lane lines are set to dashed white lines; and / or, for unidirectional vectorized straight road lines: if an adjacent intersection scenario, a continuous intersection scenario, or a tunnel and / or bridge segment scenario is identified, the lane lines are set to solid white lines; otherwise, the lane lines are set to dashed white lines. This achieves intelligent configuration of straight road markings under different vectorization modes and road scenarios, ensuring the standardization and safety of marking settings in intersection areas and ordinary road sections.

[0011] In an exemplary embodiment, step S3 includes: for bidirectional vectorized curved road lines: if a long curve scenario, a continuous multi-curve scenario, or a no-overtaking scenario is identified, the road center line is set to a solid yellow line, and / or the lane lines are set to solid white lines; otherwise, the road center line is set to a dashed yellow line, and / or the lane lines are set to dashed white lines; and / or, for unidirectional vectorized curved road lines: if a long curve scenario, a continuous multi-curve scenario, or a no-overtaking scenario is identified, the lane lines are set to solid white lines; otherwise, the lane lines are set to dashed white lines. This improves the accuracy of traffic marking settings on curved road sections.

[0012] In an exemplary embodiment, step S3 includes: for adjacent road surfaces generated based on unidirectional vectorized road lines, generating double yellow lines, green belts, or no separators between the adjacent road surfaces according to the distance between them and the continuous length of that distance; if a non-highway / urban expressway scenario is identified, simulating the generation of curb stones at a preset distance offset outward from the road boundary line; and / or, if a ramp scenario is identified, generating guide strips in the corresponding road surface bifurcation area. This achieves automatic generation of complex elements such as road separation facilities, curb stones, and guide strips, improving the coverage of diverse road facilities by lane-level maps.

[0013] In an exemplary embodiment, step S2 includes: determining the number of lanes and the road function level based on road attributes; and generating road boundary lines by extending laterally a predetermined distance from the geometric location of the road lines based on the number of lanes and the road function level. Specifically, the geometric topology of the road surface also includes lane lines, wherein lane lines are generated based on the number of lanes within the area defined by the road boundary lines on both sides. This establishes a precise geometric transformation mechanism from road lines to road surfaces, providing a basic road spatial representation for lane-level maps.

[0014] In an exemplary embodiment, step S2 includes adding relative height coordinates to points on the geometric topology of the road surface based on the height hierarchy and slope attributes contained in the road attributes of the road lines. This expands the two-dimensional road data into a three-dimensional representation, enabling the map to realistically reflect the slope changes and three-dimensional intersections of the road.

[0015] In an exemplary embodiment, step S2 includes: translating adjacent road surfaces according to the distance between them; and / or performing curve fitting on the road lines represented by polylines, and generating the geometric topology of the road surfaces based on the curve-fitted road lines. This effectively overcomes the problem of abnormal road surface spatial relationships caused by insufficient accuracy of the original data in the high-precision map, ensuring the geometric accuracy and rationality of the lane-level map data.

[0016] In an exemplary embodiment, the method further includes the following steps: the refined map data further includes the geometric topology of intersection nodes; based on the geometric topology of road lines, the geometric topology of intersection nodes, and road attributes, a geometric topology of the intersection surface is generated; according to predefined road element configuration rules, corresponding road elements are generated for the geometric topology of the intersection surface; and lane-level map data is generated based on the geometric topology of the intersection surface and the corresponding road elements. This achieves complete modeling of the intersection area, represents the intersection as an independent geographic feature with refined detail, and improves the road network structure of the lane-level map.

[0017] In one exemplary embodiment, the geometric topology of an intersection surface is generated by: generating the geometric topology of all road lines connected to the intersection node; and determining the boundary range of the intersection surface based on the intersection area formed by the geometric topology of each road surface at the intersection node. This establishes an automatic intersection surface generation method based on road surface intersections, ensuring a natural connection between the intersection geometry and the connected roads.

[0018] In one exemplary embodiment, the method further includes the following steps: determining the style attributes and starting point positions of lane lines and / or lane centerlines of the road surface connected to the intersection surface according to the road element configuration rules applicable to the intersection scenario; generating stop lines and / or guide arrows at the entrance lanes of the road surface connected to the intersection surface; and / or determining the geometric position of zebra crossings within the boundary range of the intersection surface. This achieves systematic configuration of traffic elements in the intersection area, improving the accuracy and safety of lane-level navigation at intersections.

[0019] In one exemplary embodiment, the method further includes the step of sending the generated lane-level map data to at least one vehicle for lane-level navigation guidance. This enables the lane-level map data to effectively serve the high-precision navigation needs of the vehicle.

[0020] In an exemplary embodiment, the method further includes the following steps: generating lane-level map data based on standard-precision map data only for road areas not covered by high-precision map data; directly providing high-precision map data as lane-level map data for road areas already covered by high-precision map data; and / or, stitching together a first map block generated based on the standard-precision map data and a second map block generated based on the high-precision map data, wherein the first map block and the second map block are adjacent, and the stitching process includes at least ensuring a smooth connection between the lane directions on the first and second map blocks. This establishes a collaborative working mechanism between standard-precision and high-precision maps, utilizing the high-precision advantage of high-precision maps while filling coverage gaps with standard-precision maps, achieving a balance between cost and accuracy. It solves the problem of connection between map blocks of different precision, ensuring the continuity of lane directions through smooth connection technology.

[0021] According to a second aspect of this application, a method for lane-level navigation is provided, the method comprising the following steps: acquiring lane-level map data generated by the method described in the first aspect of this application; acquiring real-time location information of a vehicle and navigation route information; performing lane-level navigation guidance rendering based on the lane-level map data, the real-time location information of the vehicle, and the navigation route information; and visualizing the rendering result.

[0022] According to a third aspect of this application, an apparatus for generating lane-level map data is provided, the apparatus including a memory and a processor, the memory storing computer program instructions, which, when executed by the processor, enable the processor to perform the method according to the first aspect of this application.

[0023] According to a fourth aspect of this application, an apparatus for performing lane-level navigation is provided, the apparatus including a memory and a processor, the memory storing computer program instructions that, when executed by the processor, enable the processor to perform the method according to a second aspect of this application.

[0024] According to a fifth aspect of this application, a computer program product includes computer program instructions, wherein, when executed by one or more processors, the computer program instructions enable the one or more processors to perform the method according to the first and / or second aspects of this application. Attached Figure Description

[0025] The principles, features, and advantages of this application will be better understood below with reference to the accompanying drawings. The drawings include: Figure 1A flowchart is shown of a method for generating lane-level map data according to an exemplary embodiment of this application; Figure 2 It shows Figure 1 A flowchart of one method step of the method shown; Figure 3 A flowchart illustrating a method for performing lane-level navigation according to an exemplary embodiment of this application is shown; Figure 4A and Figure 4B A schematic diagram illustrating the geometric topology of road surfaces generated from high-precision map data according to an exemplary embodiment of this application is shown. Figure 5A and 5B A schematic diagram illustrating the translation process performed on the geometric topology of a generated road surface according to an exemplary embodiment of this application is shown; Figure 6 A schematic diagram illustrating the smoothing of the geometric topology of a generated road surface according to an exemplary embodiment of this application is shown; Figure 7 A schematic diagram illustrating the generation of the geometric topology of a road surface with continuous height information according to an exemplary embodiment of this application is shown; Figures 8A to 8G A schematic diagram is shown illustrating the generation of road elements based on the geometric topology of a road surface according to an exemplary embodiment of this application; Figures 9A to 9C A schematic diagram illustrating the generation of road dividers based on a vectorized pattern and the geometric topology of the road surface according to an exemplary embodiment of this application is shown. Figures 10A to 10D A schematic diagram illustrating the geometric topology of an intersection surface generated from refined map data according to an exemplary embodiment of this application is shown; and Figure 11 Block diagrams of an apparatus for generating lane-level map data and an apparatus for performing lane-level navigation, according to exemplary embodiments of the present application, are shown. Detailed Implementation

[0026] To make the technical problems to be solved, the technical solutions, and the beneficial technical effects of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and several exemplary embodiments. It should be understood that the specific embodiments described herein are only for explaining this application and are not intended to limit the scope of protection of this application.

[0027] Figure 1 A flowchart of a method for generating lane-level map data according to an exemplary embodiment of this application is shown. The method includes steps S1 to S4.

[0028] In step S1, high-precision map data is obtained, which includes the geometric topology of road lines and road attributes.

[0029] In this context, a high-precision / high-definition map (HD map) refers to precise map data acquired through professional surveying methods, with an absolute accuracy down to the centimeter level. This type of map uses lanes as the basic recording unit and fully includes precise geometric information, topological connections, and rich semantic attributes (such as lane type, traffic signs, and road markings) of elements such as lane lines and road boundaries. It can provide accurate data support for lane-level navigation and autonomous driving functions. However, HD maps suffer from technical bottlenecks such as high acquisition costs and complex update and maintenance, making it difficult to achieve rapid coverage of large-scale geographical areas.

[0030] Standard Definition Map (SD Map) is a type of vector map data designed for traditional road-level navigation applications, typically with meter-level accuracy. This type of map primarily describes the overall spatial orientation and macroscopic morphological features of roads, lacking detailed lane geometry, road boundary lines, and other refined elements. Therefore, it can only support road-level navigation guidance and cannot meet the requirements for lane-level precision positioning and route planning.

[0031] In a high-resolution map data model, roads are abstracted using links, and intersections are abstracted using nodes. These elements are typically recorded as an ordered sequence of coordinates in a high-resolution map. By connecting these coordinate points in sequence, road lines can be formed. Each coordinate point on a road line corresponds to a specific geographical location (usually using the WGS84 latitude and longitude coordinate system), and the entire coordinate sequence fully depicts the geometric shape of the road on the map.

[0032] The geometric topology information of a road line mainly includes vector lines composed of coordinate sequences. In addition, it may include the connection relationships between the road line and other road lines and road nodes. Furthermore, this geometric topology may also include the relative positional relationship between the road line and the actual road it represents. In practical applications, different map providers may select different methods for the road geometric reference in their high-resolution maps. In some cases, the centerline of each actual road is selected as the geometric reference, and the road line is drawn based on the location of the centerline. In other cases, the right boundary line of the leftmost lane is selected as the geometric reference, and the road line is constructed based on the location of this boundary line.

[0033] Road attributes include, for example, the number of lanes on the actual road as represented by lane lines, the road function level (e.g., expressway, urban expressway, national highway, provincial highway, urban arterial road, urban secondary arterial road), speed limit information, tunnel and bridge signage information, curvature information, and traffic direction information. In addition, road attributes may also include the type of road node and the number of lanes it connects to. Furthermore, this may include the vectorization mode of the road lines, i.e., whether it involves unidirectional or bidirectional vectorized road lines.

[0034] According to the road function level, different countries have regulations on lane width and number of lanes that comply with national standards. Table 1 lists the relevant standards of several countries.

[0035] Table 1

[0036] In step S2, the geometric topology of the road surface is generated based on the geometric topology of the road lines and road attributes. The geometric topology of the road surface includes at least the road boundary lines, and optionally also lane lines and road centerlines. In another embodiment, the geometric topology of the road surface may further encompass the closed surface region enclosed by the road boundary lines, thus forming a complete regional representation. Furthermore, the geometric topology of the road surface may also include the relative height value (Z coordinate) of any point within the road surface to construct a three-dimensional representation of the road surface. This geometric representation including relative elevation data not only accurately describes the road plan layout but also reflects longitudinal height changes, providing richer spatial environmental features for lane-level navigation and autonomous driving decisions. In lane-level maps, the specific representation of the road surface geometric topology can be diverse, and can be dynamically adjusted based on factors such as actual storage capacity and transmission bandwidth.

[0037] In one embodiment, the number of lanes and the road function level are first determined based on road attributes. The standard lane width is determined according to the road function level (e.g., 3.75 meters for highways or urban expressways, and 3.5 meters for ordinary urban roads). Combined with the number of lanes, the total road width can be calculated. The road boundary line is then generated by extending laterally a predetermined distance from the geometric position of the road line. The specific direction and distance of the extension depend on how the geometric reference of the road line is selected in the refined map. For example, if the road line is selected based on the road centerline, it needs to be extended symmetrically to both sides. If the road line is selected based on the road boundary, it may only need to be extended to one side.

[0038] In one embodiment, after generating the road boundary lines, the geometric positions of the lane lines and the road centerline can be generated further. These two, together with the road boundary lines, constitute the geometric topology of the road surface. Specifically, within the area defined by the road boundary lines on both sides, the positions of each dividing line can be determined by equal division or by dividing according to preset rules based on the known number of lanes or the calculated lane width. These dividing lines constitute the geometric topology of the lane lines. For the road centerline, the position of the original road line can be directly used as the centerline. Alternatively, the centerline within the area defined by the road boundary lines can be extracted as the road centerline.

[0039] In another embodiment, the vectorization mode of the road line can be determined in advance based on road attributes, and then the corresponding geometric expansion rules can be used to generate the geometric topology of the road surface according to different vectorization modes. For example, for a unidirectional vectorized road line, the lateral expansion can be directly performed based on the number of lanes in the road attributes, starting from the geometric position of the road line, to generate the complete road surface geometric topology. For a bidirectional vectorized road line, if the lane number configuration is not explicitly recorded in the refined map data, it can be assumed that each direction contains one lane, and the lane width can be expanded by 3.5 meters to both sides from the geometric position of the road line to generate the basic road surface geometric topology.

[0040] In another embodiment, step S2 may further include optimizing the geometric topology of the road lines and road surfaces. For example, the relative positional relationship of adjacent road surfaces can be adjusted through a geometric translation algorithm, and the original polyline road can be smoothed using curve fitting technology. This allows the generated road surface geometry to better fit the real road shape, improving the stability and visual effect of lane-level navigation rendering. This will be further explained below with reference to specific embodiments.

[0041] In step S3, predefined road element configuration rules are obtained, and corresponding road elements are generated for the geometric topology simulation of the road surface based on the predefined road element configuration rules.

[0042] Predefined road element configuration rules may include mandatory standards and specifications such as the "Road Traffic Signs and Markings" promulgated by specific countries or regions (e.g., China's GB 5768 series standards), which can be obtained through government public channels or industry databases. In addition, road element configuration rules may also include design guidelines summarized or generalized by automobile manufacturers (OEMs) based on actual road environments. These configuration rule systems define, for example, the geometric positions and style attributes (including shape, color, pattern, size, etc.) of traffic signs and markings, as well as the setting requirements under different road environments and scenarios. For example, GB 5768.2 stipulates that double solid yellow lines must be set on road sections where overtaking is prohibited, while yellow dashed and solid lines are set on road sections where temporary overtaking is permitted.

[0043] Predefined road element configuration rules are typically formulated for the entire road network. However, during the actual generation process, the specific rules applicable must be determined based on the characteristics of the specific road surface or road segment. The specific methods for rule selection will be discussed below. Figure 2 To elaborate further.

[0044] Road elements include, for example, the geometric positions and corresponding style attributes of traffic markings and traffic signs. Traffic markings include, for example, lane lines, road center lines, stop lines, ground directional arrows, and / or guide strips. The style attributes of these traffic markings include shape, pattern, line type (solid line / dashed line / double solid line, etc.), color (white / yellow, etc.), and / or size. Traffic signs include traffic signs, road boundary markers, and road dividers. The style attributes of these traffic signs include type (printed ground markings, solid dividers), material (green belts, guardrails, sound barriers, curbs, etc.), and / or size.

[0045] In this context, "simulation generation" can be understood as, for example, the technical process of creating virtual road elements using only existing high-precision map data and in accordance with established standards and specifications, without the aid of additional field surveying, sensor sensing, or any on-site data collection. This method does not use actual collected data such as images or point clouds; instead, it uses a digital rule engine to transform standard clauses into a computable model, outputting virtual road elements that conform to legal style specifications and are precisely registered with road geometry. This process supports fully automatic or semi-automatic execution.

[0046] In addition Figure 1 In the illustrated embodiment, the method may optionally include step S2' in parallel or alternate with step S2. In step S2', the refined map data may also include the geometric topology of intersection nodes. Based on the geometric topology of road lines, the geometric topology of intersection nodes, and road attributes, the geometric topology of intersection surfaces may also be generated.

[0047] Specifically, for example, nodes labeled "intersection" are first selected from the road nodes, and then corresponding road surface geometry is generated for all road lines connected to these nodes. By calculating the geometric intersection area of ​​these road surfaces at the intersection nodes, the boundary range of the intersection surface can be preliminarily determined. In one embodiment, this intersection area can be directly used as the effective range of the intersection surface, and its outline can be defined as the intersection boundary line. Alternatively, the intersection area can be trimmed, spliced, and smoothed to obtain an intersection surface geometry that conforms to the characteristics of a real road. This will be further explained below with reference to specific embodiments.

[0048] Accordingly, in step S3, road elements can be generated for the geometric topology of the intersection surface and the road surface according to the road element configuration rules applicable to the intersection scenario. Specifically, for example, the style attributes and starting point positions of the lane lines and / or lane center lines of the road surface connected to the intersection surface can be determined according to the road element configuration rules applicable to the intersection scenario (such as traffic regulations, national standards, etc.), stop lines and / or guide arrows can be generated at the entrance lanes of the road surface connected to the intersection surface, and the geometric position of the zebra crossing can be determined within the boundary range of the intersection surface.

[0049] In step S4, lane-level map data is generated based on the geometric topology of the road surface and the corresponding road elements. Additionally, lane-level map data can also be generated in this step based on the geometric topology of the intersection surface and the corresponding road elements.

[0050] In this step, for example, the road surface geometry topology (including the coordinate information of lane lines, road boundary lines and other elements) generated in step S2 can be integrated with the road elements (including the geometric position and style attributes of traffic markings and signs) generated in step S3 to construct a complete lane-level road network description.

[0051] Furthermore, to achieve continuous map coverage, different map blocks can be stitched together in step S4. For example, a first map block generated from the standard-precision map data and a second map block generated from the high-precision map data can be stitched together, with the first and second map blocks being adjacent to each other. The stitching process includes, for example, smoothing the transition and geometric continuity of key elements such as lane direction and road boundaries between blocks to construct a globally consistent lane-level map covering a specific geographical area.

[0052] In one embodiment, the generation of lane-level map data can be performed by a remote server (such as a map provider's cloud platform or a qualified vehicle manufacturer's backend server), and the generated lane-level map data is sent to the vehicle terminal. This map data distribution process, for example, supports an on-demand response mode, which dynamically extracts map data blocks within the relevant geographical area based on the vehicle's reported real-time location and planned navigation route, enabling targeted fragmented data delivery. Alternatively, the remote server can proactively distribute the data. In another embodiment, this data generation process can also be performed at least partially on a local device, such as a high-performance computer or other terminal device.

[0053] In one embodiment, lane-level map data is transmitted in its raw, unrendered format, containing only geometric topological information (such as the 3D coordinate sequences of various lane lines and road elements) and corresponding style attribute descriptions. After this data is transmitted to the in-vehicle terminal, the local navigation client is responsible for the final visualization rendering and graphical presentation.

[0054] In another embodiment, the server can pre-process the visualization rendering of the map data, thereby directly sending the rendered map data to the vehicle terminal. The vehicle can then use the data directly without any additional graphics processing.

[0055] Figure 2 It shows Figure 1 A flowchart of one method step of the method shown is provided. In this embodiment, Figure 1 Step S3 of the method shown is illustrated as including sub-steps S31 to S34.

[0056] In step S31, the vectorization mode of the road line is determined based on the road attributes of the road line. The vectorization mode includes unidirectional vectorized road lines and bidirectional vectorized road lines. For example, this vectorization mode is stored in the road traffic direction field (unidirectional / bidirectional) in the high-resolution map data.

[0057] In step S32, the road scenarios involved in the road surface are identified based on the geometric topology and road attributes of the road lines.

[0058] In this context, "road scenario" refers to a classification of typical traffic scenarios defined based on traffic regulations and related technical standards. These scenarios include, but are not limited to: near intersection scenarios, consecutive intersection scenarios, long curve scenarios, consecutive multi-curve scenarios, no-overtaking scenarios, tunnel and / or bridge scenarios, highway / urban expressway scenarios, and ramp scenarios. Each road scenario is associated with corresponding traffic rule constraints and road element configuration specifications. These predefined scenario rules provide crucial decision-making basis and semantic context for generating lane-level map data that meets regulatory requirements and functional needs.

[0059] For example, adjacent intersection scenarios can be identified based on the spatial proximity between the road represented by the roadline and the intersection represented by the intersection node it connects to. Furthermore, consecutive intersection scenarios can be identified based on the length of the area between two intersections represented by the roadline. Additionally, long curve scenarios and / or consecutive multi-curve scenarios can be identified based on the length and curvature information of the roadline. Furthermore, no-overtaking scenarios can be identified based on traffic sign information. Furthermore, tunnel and / or bridge segment scenarios can be identified based on tunnel and bridge signage information. Furthermore, expressway / urban expressway or non-expressway / urban expressway scenarios can be identified based on the road function level of the roadline. Furthermore, ramp scenarios can be identified based on the road morphology attributes and road function level of the roadline. This will be discussed in conjunction with... Figures 8A to 8G To elaborate further.

[0060] For road scenarios involving changes in the number of lanes, a smooth transition in geometric topology is required in sections where lanes are added or removed. Specifically, when a change in lane configuration is detected (such as the expansion from two lanes to three lanes), a continuous and smooth transition curve can be constructed using a curve fitting algorithm (such as a B-spline curve) with the original lane end as the reference point and the target geometric position of the new lane as the endpoint, thereby achieving a natural transition of the newly added lane.

[0061] In one embodiment, steps S31 and S32 can be performed independently of each other, and step S32 can also be performed based on the result of step S31. For example, the road scenario can be determined first based on the vectorization mode of the road lines, and then applicable road element configuration rules can be selected based on the road scenario. The settings for road scenario determination parameters may differ under different vectorization modes. For example, when determining a continuous intersection scenario, different distance threshold parameters may be required for different road line vectorization modes.

[0062] In step S33, based on the determined vectorization mode and the identified road scenario, configuration rules that match the current vectorization mode and road scenario are called from the predefined road element configuration rule set.

[0063] Specifically, map providers' cloud platforms, for example, pre-acquire and store a complete set of road element configuration rules, covering all possible combinations of road scenarios. For instance, differentiated configuration rules are stored for non-intersection scenarios and intersection scenarios. After identifying the specific scenario to which each road line (or the road it represents) belongs based on the high-precision map data, the corresponding configuration rules can be dynamically invoked.

[0064] For example, when a non-adjacent intersection scenario is detected, the corresponding configuration rule can be invoked to generate lane lines in the form of white dashed lines. When an adjacent intersection scenario is detected, the configuration rule is automatically switched to generate lane lines in the form of white solid lines. This scenario-based rule dynamic invocation mechanism ensures that the generated road elements conform to both the physical structural characteristics of the road and the regulatory requirements of the specific traffic scenario, without relying on the collection of actual environmental perception data.

[0065] In step S34, road elements are generated for the geometric topology of the road surface according to the applicable road element configuration rules that are invoked.

[0066] For example, when a "two-way vectorized road line" is identified and it is in an "adjacent intersection scenario," the corresponding configuration rule will be invoked. This rule may stipulate that in this scenario, lane lines should be set in the form of solid white lines, while the road center line should be set in the form of solid yellow lines. Therefore, following this rule, lane lines that meet the above scenario range can be set as solid white lines, and the road center line can be set as solid yellow lines.

[0067] Road elements in lane-level map data can be stored in a structured vector format, including geometric location and style attributes. Geometric location records the spatial shape of the road element using a three-dimensional coordinate sequence, while style attributes define element type, line type, and color using standardized coding. Road elements can be stored in association with corresponding lane surfaces or intersection surfaces and can be retrieved by geographic block.

[0068] The specific details regarding the generation of road elements based on road scenarios and road line vectorization patterns will be elaborated below.

[0069] It should be noted that, although Figure 2 In the illustrated embodiment, the selection of road element configuration rules considers both the vectorization mode and the road scenario. However, in actual implementation, only one factor may be used to select applicable road element configuration rules. Furthermore, the execution order of steps S31 and S32 is not fixed; they can be swapped or executed in parallel according to actual needs.

[0070] Figure 3 A flowchart of a method for performing lane-level navigation according to an exemplary embodiment of this application is shown. In this embodiment, the method includes steps 301 to 304.

[0071] In step 301, obtain through Figure 1 or Figure 2 The lane-level map data generated by the method shown means that the map data for at least some areas is generated solely based on the standard-precision map data.

[0072] In one embodiment, for example, while the vehicle is in motion, a real-time request can be initiated directly or via the vehicle manufacturer's backend server to the map provider's server. Upon response, the map provider's server distributes the corresponding lane-level map data. Alternatively, during the pre-installation phase, the generated full-area lane-level map data provided by the map provider can be pre-stored locally in the vehicle to support offline navigation in environments without a network connection.

[0073] In step 302, the vehicle's real-time location information and navigation route information are acquired. Specifically, real-time location parameters such as the vehicle's current geographic coordinates, driving direction, and speed are continuously collected using sensor fusion technologies such as the onboard GPS positioning system and inertial navigation unit. In addition, pre-calculated navigation route information can also be obtained from the navigation planning module simultaneously.

[0074] In step 303, lane-level navigation guidance rendering is performed based on lane-level map data, real-time location information, and navigation route information.

[0075] Vehicle navigation software may offer lane-level and road-level navigation modes. When lane-level navigation is selected, the vehicle requests lane-level map data from a map provider server (cloud platform). When road-level navigation is selected, it requests road-level map data from the map provider server. The two types of map data are rendered differently on the vehicle.

[0076] In lane-level navigation mode, refined rendering is achieved based on the acquired lane-level map data. Specifically, the original road surface geometry and topology data and road element style data are first transformed into geometric patterns on a complete visual plane, establishing an accurate lane-level map base. Subsequently, based on the vehicle's real-time location and navigation route information, the vehicle's current lane position is accurately marked on the rendered map, forming a continuous lane-level guidance route.

[0077] In addition, it can fully render elements such as lane line styles (e.g., solid and dashed lines, colors), ground directional arrows, and intersection guidance areas, and dynamically display traffic signs related to the current lane (e.g., speed limits, no overtaking, etc.).

[0078] Furthermore, by utilizing the relative height coordinates of road surface geometry contained in lane-level map data, it is possible to render the road surfaces of interchanges, the slope variations at the entrances of elevated roads, and complex road structures such as underpasses and auxiliary roads located at the lower levels. This 3D visualization process makes the relative hierarchical relationships and slope trends between roads readily apparent, providing users with a more realistic and three-dimensional understanding of the road environment, further enhancing the navigation experience and driving safety in complex road conditions.

[0079] This rendering method not only clearly shows the vehicle's precise lane position on the current road, but also provides advance notice of lane change points and intersection guidance information, offering drivers a more accurate and intuitive navigation experience.

[0080] In step 304, the rendering result is visualized. For example, it can be displayed using the vehicle's display unit (such as a central control screen, head-up display, instrument panel display, etc.), or it can be displayed on the vehicle user's smartphone, tablet, or other smart mobile terminals. Simultaneously, corresponding voice prompts can be generated to provide clear voice guidance to the driver at key navigation nodes (such as when changing lanes or about to enter a ramp).

[0081] In one embodiment, in road-level navigation mode, the navigation interface only presents abstract road lines and does not provide lane-level details. Navigation guidance remains at the road level, providing turn instructions at key nodes such as intersections and roundabouts, but it cannot display specific lane divisions, nor can it provide the vehicle's current lane position or lane-level driving suggestions.

[0082] In one embodiment, under lane-level navigation mode, detailed rendering of road scenes is achieved based on lane-level map data. This not only displays detailed road elements such as lane lines, ground guide arrows, and road boundary lines in the interface, but also shows the vehicle's specific lane position in the current road in real time. Furthermore, the navigation interface can present continuous and intuitive lane-level guidance routes to provide driving suggestions accurate to the specific lane.

[0083] Figure 4A and Figure 4B A schematic diagram of the geometric topology of road surfaces generated from high-precision map data according to an exemplary embodiment of this application is shown.

[0084] Figure 4A The left side shows the representation of a unidirectional vectorized road line 51 in a high-resolution map. In reality, such roads are typically high-grade roads (e.g., expressways, urban expressways, national highways, provincial highways, urban arterial roads, urban secondary arterial roads), with physical or legal barriers separating two-way traffic. In high-resolution map data modeling, each unidirectional road is modeled as an independent road line 51, and the direction of travel is marked as unidirectional in the road attributes.

[0085] For example, in this embodiment, the geometric reference of road line 51 is set at the road centerline, the road function level is highway, the corresponding standard lane width is set to 3.75 meters, and the number of lanes is 2. Based on this, as Figure 4A As shown on the right, starting from the geometric position of road line 51, extending 3.75 meters to the left and right respectively, the left boundary line 511 and the right boundary line 512 of the road can be reconstructed. The area bounded by these two boundary lines 511 and 512 constitutes the road surface 510.

[0086] Based on the restoration of road boundary lines 511 and 512, the area enclosed by the boundary lines can be divided according to the number of lanes to determine the specific geometric position of each lane line 513, thereby improving the geometric topology of the road surface.

[0087] Figure 4B The left side shows the representation of bidirectional vectorized road lines 53 in a refined map. In reality, such roads are typically low-grade roads such as county roads, township roads, village roads, or ordinary urban roads. These roads have no physical or legal barriers between them, only a single dashed / solid yellow line, or no markings at all, allowing two-way traffic. In the refined map data modeling, such roads are abstracted as a single road line 53, and the traffic direction is marked as bidirectional in the road attributes.

[0088] In this embodiment, the geometric reference of road line 53 is, for example, set at the center of the two-way road, the road function level is ordinary urban road, the corresponding standard lane width is set to 3.5 meters, and by default, one lane is configured in each direction. Based on this, as Figure 4B As shown on the right, starting from the geometric position of road line 53, extend 3.5 meters to the left and right sides respectively to generate the left boundary line 531 and the right boundary line 532, which together enclose and form a complete road surface 530.

[0089] Based on the restoration of the road boundary lines, the specific location of the road centerline can be further determined within the road surface area according to the configuration of one lane in each direction, thereby improving the geometric topology of the road surface.

[0090] Figure 5A and 5B A schematic diagram illustrating the translation process performed on the geometric topology of a generated road surface according to an exemplary embodiment of this application is shown.

[0091] Because road lines 51 and 52 in the standard map data themselves have an accuracy error of approximately 10-30 meters, road surfaces 510 and 520 directly generated from them may have positional deviations in actual space. For example... Figure 5A As shown, the road surfaces 510 and 520 generated by the two unidirectional vectorized road lines 51 and 52 with opposite directions partially overlap. This overlap can be identified, for example, by calculating the distance vector from the right boundary line of the left road surface 510 to the left boundary line of the right road surface 520 (in this embodiment, for example, pointing to the right is defined as positive direction).

[0092] When an overlap is detected, for example, the geometry of road surfaces 510 and 520 can be translated to achieve a reasonable spatial layout. Specifically, if the geometry of road surfaces 510 and 520 only contains road boundary lines, they can be translated as a whole until the absolute value D of the distance vector between the right boundary line of the left road surface 510 and the left boundary line of the right road surface 520 is less than 1 meter and the direction remains positive.

[0093] Figure 5B This shows the case where no translation is required, where the road surfaces 510 and 520 generated by the two unidirectional vectorized road lines 51 and 52 do not overlap.

[0094] Figure 6 A schematic diagram illustrating the smoothing process performed on the geometric topology of a generated road surface according to an exemplary embodiment of this application is shown.

[0095] Figure 6The left side shows a typical storage data structure for road line 51 in a high-resolution map, which is, for example, stored as a sequence of polylines in the WGS84 coordinate system. When the road it represents is curved, this representation results in adjacent line segments being connected by polylines, creating unnatural turns. The road surface geometry generated directly from such polylines (including road boundary lines parallel to the road lines) also inherits the non-smoothness, exhibiting a tortuous geometry.

[0096] Therefore, such as Figure 6 As shown on the right, curve fitting technology can be used to smooth the original road line 51 to obtain the curve-fitted road line 61. In specific implementation, algorithms such as B-spline curves, Bézier curves, or polynomial fitting can be used to generate a continuous and smooth road line 61 through interpolation or approximation. Then, the road boundary line is extended based on the curve-fitted road line 61, and the resulting road surface 610 will present a natural and smooth state.

[0097] It should be noted that, although in Figure 6 The diagram shows a smoothing process where curve fitting is first performed on road line 51 before generating road surface 610. However, in actual implementation, road boundary lines can also be generated directly from the original road line 51, and then curve fitting is performed on the road boundary lines.

[0098] Figure 7 A schematic diagram of generating the geometric topology of a road surface with continuous height information according to an exemplary embodiment of this application is shown.

[0099] Figure 7 The left side shows the two-dimensional abstract representation of roads at three different levels in the refined map. These roads are geometrically represented by road lines 71, 72, and 73. This two-dimensional representation cannot directly reflect the actual spatial height relationships. Therefore, in the refined map data, only the vertical dimension attributes of z-level (height level attribute) and slope (slope attribute) are recorded for them.

[0100] Specifically, road line 73 has a z-level value of 0, serving as a relative height reference. Road line 71 has a z-level value of 1, indicating that the spatial entity it represents is above the reference layer. Road line 72 has a z-level value of -1, indicating that it is below the reference layer. Regarding slope attributes, road line 73 has a positive slope value, representing an uphill section, while roads 71 ​​and 72 both have zero slope values, representing flat road sections. Although these attribute values ​​allow for spatial relationship reasoning, in practical navigation applications, this abstract representation does not meet the spatial intuitiveness requirements of lane-level navigation.

[0101] Therefore, as Figure 7As shown on the right, during the generation of lane-level map data, continuous relative height coordinates can be added to each point on the generated road surface's geometric topology. First, road boundary lines are formed based on road line extensions, thus establishing the planar geometric region of the road surface. Then, based on the vertical dimension attributes recorded in the refined map data, height coordinates (z-coordinates) are generated for points on the road boundary lines, points on lane lines, points on the road centerline, and / or arbitrary sampling points within the closed planar region enclosed by the road boundary lines, using interpolation algorithms. This forms a continuous relative height coordinate representation of the entire road surface. Here, the height coordinates are not, for example, actual elevation values, but rather relative heights in a simulated three-dimensional coordinate system.

[0102] In this embodiment, the road surface 730 of the uphill road achieves an accurate simulation of the inclined plane through continuous height coordinates, clearly showing its spatial relationship of being connected to the road surface 720 of the auxiliary road at its lower part and to the elevated road surface 710 at its upper part. This geometric topology containing continuous height information provides a more realistic three-dimensional environment representation for lane-level navigation rendering.

[0103] Figures 8A to 8G A schematic diagram is shown illustrating the generation of road elements based on the geometric topology of a road surface according to an exemplary embodiment of this application.

[0104] exist Figure 8A The process of generating road elements under scenarios A1 (near an intersection) and A2 (non-near an intersection) is illustrated. In this embodiment, firstly, based on the road attributes recorded in the refined map data, a specific road line is determined to be of a bidirectional vectorized type, and its straight-ahead nature is confirmed by analyzing the curvature parameters in the road attributes (e.g., curvature radius less than 1300 meters). Based on this, the spatial proximity between the road represented by the road line and the intersection represented by the intersection node it connects to can be used to identify whether an adjacent intersection scenario is involved.

[0105] For example, when the distance between a defined area of ​​the road surface 530 generated by the road line and the boundary lines 811 and 821 of the connected intersection surfaces 810 and 820 is less than a preset threshold of 30 meters, the area can be determined to belong to the adjacent intersection scenario A1. Areas exceeding the preset threshold are identified as non-adjacent intersection scenario A2. It should be noted that other distance evaluation methods can also be used in actual implementation. For example, the distance between feature points on the road line and intersection nodes can be calculated, or the spatial relationship between sampling points on the road surface and the geometric center of the intersection surface can be analyzed.

[0106] According to the road element configuration rules stipulated in the national standard, for scenario A1 (within 30 meters of the intersection), lane line 22 should be set as a solid white line and road center line 21 should be set as a solid yellow line, as required by the standard. For scenario A2 (more than 30 meters from the intersection), lane line 22 should be set as a dashed white line and road center line 21 should be set as a dashed yellow line, as required by the standard.

[0107] In another embodiment (not shown), when identifying near-intersection scenarios for unidirectional vectorized straight roads, a different judgment threshold than that used in bidirectional vectorization modes can be employed. For example, when the distance between a sampling point or specific area on the road surface and the boundary line of the intersection surface is less than 50 meters, it is determined to be a near-intersection scenario, and all lane lines should be set to solid white lines. When the distance between an area on the road surface and the intersection boundary exceeds 50 meters, it should be set to dashed white lines.

[0108] Figure 8B The process of generating road elements under continuous intersection scenario A3 is illustrated. For road surface 530 generated from bidirectional vectorized straight road lines, continuous intersection scenario A3 can be identified, for example, by calculating the length of the segment of road surface 530 between the two intersection surfaces 810 and 820. For example, if the segment length is less than or equal to 100 meters, it is continuous intersection scenario A3. If it exceeds 100 meters, it is a discontinuous intersection scenario.

[0109] According to the applicable road element configuration rules, in the case of consecutive intersections, lane line 22 is set as a solid white line and road center line 21 is set as a solid yellow line. In the case of non-consecutive intersections and road sections that are not adjacent intersections, lane line 22 is set as a dashed white line and road center line 21 is set as a dashed yellow line.

[0110] In an embodiment not shown, for a unidirectional vectorized straight road, if a continuous intersection scenario is identified, the lane line is set to a solid white line; otherwise, the lane line is set to a dashed white line.

[0111] Figure 8C The process of generating road elements under the long curve scenario A4 is illustrated. For the road surface 540 generated by bidirectional vectorized road lines, if it is a curved road (e.g., with a radius of curvature less than 1300 meters), and the continuous extension distance of the curved road segment reaches or exceeds a preset threshold D1 (D1 ≥ 200 meters), then the long curve scenario A4 is identified. Therefore, according to the applicable road element configuration rules, lane line 22 can be set as a solid white line, and road center line 21 as a solid yellow line. Otherwise, it is a normal curve scenario A5, where lane line 22 can be a dashed white line, and road center line 21 can be a dashed yellow line.

[0112] In an embodiment not shown, for a unidirectional vectorized curved road, if a long curve scenario is identified (continuously exceeding 200 meters (WS_D>=200m)), the lane line is set to a solid white line.

[0113] Figure 8D The process of generating road elements under the scenario A6 of continuous multiple curves is illustrated. For a road surface 540 generated from bidirectional vectorized road lines, if it is a curved road (e.g., with a radius of curvature less than 1300 meters), and the road includes at least two consecutive curves with a section containing consecutive curves of less than 20 meters, then the scenario A6 of continuous multiple curves is identified. Under this scenario, according to the applicable road element configuration rules, the road centerline 21 is set as a solid yellow line. If lane lines are included, they can also be set as solid white lines (not shown in detail for simplicity).

[0114] In an embodiment not shown, for a unidirectional vectorized curved road line, if a continuous multi-curve scenario A6 is identified, the lane lines are set to solid white lines, for example.

[0115] exist Figure 8E The process of generating road elements under the no-overtaking scenario A7 is illustrated. For the road surface 540 generated by bidirectional vectorized road lines, if it is a curved road (e.g., with a radius of curvature less than 1300 meters), the location information of the no-overtaking and overtaking-free traffic signs 31 and 32 can be determined based on the road attributes stored in the refined map data. The road segment between these two traffic signs 31 and 32 belongs to the no-overtaking scenario A7. Under this scenario, according to the applicable road element configuration rules, lane line 22 is set as a solid white line, and road centerline 21 is set as a solid yellow line.

[0116] Accordingly, in embodiments not shown, for unidirectional vectorized curved road lines, if a no-overtaking scenario is identified, the road lane lines can be set as solid white lines, for example.

[0117] In embodiments not shown, for bidirectional vectorized road lines, if road attributes include tunnel and bridge identification information, a tunnel and / or bridge segment scenario is identified. In this scenario, according to applicable road element configuration rules, lane lines can be set to solid white lines, and the road centerline can be set to solid yellow lines. Correspondingly, for unidirectional vectorized road lines, if a tunnel and / or bridge segment scenario is identified, lane lines can be set to solid white lines.

[0118] Figure 8FThe process of generating road elements in a non-highway urban expressway scenario is illustrated. If a non-highway urban expressway scenario is identified, curb stones 24 and 25 are simulated and generated at a preset distance offset outward from road boundary lines 511 and 512. In practical implementation, for example, curb stones 24 and 25 can be set by offsetting them outward by 0.5 meters from the left and rightmost boundary lines respectively (other suitable distance values ​​can also be used depending on the navigation rendering effect) for all non-adjacent intersection sections of the road, or the positions of the curb stones can be directly set to coincide with the corresponding boundary lines 24 and 25.

[0119] Figure 8G The process of generating road elements in a ramp intersection scenario is illustrated. Two unidirectional vectorized road lines, 51 and 52, exist in the refined map data. Both are high-level roads (including expressways, urban expressways, urban arterial roads, national highways, and provincial highways). When road line 52 is detected to have bifurcation or entrance / exit features recorded in its road attributes, and the road morphology attribute of the bifurcation road line 52 is marked as a dedicated left / right turn lane or ramp, the location is identified as a ramp intersection scenario.

[0120] In this scenario, a guide strip 27 will be installed in the bifurcation area between road surfaces 510 and 520 generated by road lines 51 and 52, respectively. In specific implementation, as follows... Figure 8G As shown, within the legally defined triangular area formed by the bifurcation zone, a solid white line 27 must be set along the entire length from the rightmost boundary line of the main road surface 510 to the leftmost boundary line of the ramp road surface 520 as a guide strip.

[0121] Figures 9A to 9C A schematic diagram illustrating the generation of road dividers based on the geometric topology of road surfaces according to a vectorization pattern and distance, as per an exemplary embodiment of this application, is shown. In the following embodiments, it is described how different types of road dividers are selected by comparing the distance between adjacent road surfaces with a preset threshold (a wide range of thresholds).

[0122] like Figure 9A As shown, two unidirectional vectorized road lines with opposite traffic directions generate road surfaces 510 and 520, respectively. When the distance d between these two road surfaces is detected to be less than 1 meter, and this distance condition continuously exceeds 10 meters, a double yellow line 28 will be automatically generated between road surfaces 510 and 520 as a road divider according to the road element configuration rules corresponding to this road scenario.

[0123] exist Figure 9B The same example shows road surfaces 510 and 520 generated by two unidirectional vectorized road lines with opposite traffic directions, but... Figure 9BIn this scenario, the distance d between two road surfaces 510 and 520 is greater than 1 meter and less than or equal to 12 meters, and this distance condition continuously exceeds 10 meters. According to the applicable road element configuration rules for this road scenario, a green belt 29 is generated between road surfaces 510 and 520 as a road divider.

[0124] exist Figure 9C The diagram shows a road scenario where the distance d between two adjacent road surfaces 510 and 520 is greater than 12 meters. In this case, according to the applicable road element configuration rules, no road dividers need to be set between the two road surfaces 510 and 520.

[0125] Figures 10A to 10D A schematic diagram is shown illustrating the generation of the geometric topology of an intersection surface from high-resolution map data and the generation of road elements thereon, according to an exemplary embodiment of this application.

[0126] exist Figure 10A In the high-precision map data, two perpendicular bidirectional vectorized road lines 53 and 54 intersect at the same intersection node 58. Taking a typical urban road as an example, using a standard lane width of 3.5 meters, extending 3.5 meters to the left and right from the geometric positions of road lines 53 and 54 generates the corresponding road boundary lines, thus obtaining the road surface geometry. When these road surface areas 530 and 540, defined by the road boundary lines, intersect at the intersection node, they form an intersection area (overlapping area). This intersection area can be determined as the effective range of the intersection surface 580, and the outline of the overlapping area constitutes the boundary line of the intersection surface.

[0127] Regarding road element configuration, based on the simulated intersection area 580 and its boundary positions, the starting positions and patterns of lane lines outside the intersection area can be calculated, and stop lines can be set. For example, according to the configuration rules applicable to adjacent intersection scenarios, the center lines of each road start from the intersection boundary, using solid yellow lines within 30 meters, and dashed yellow lines for the remaining sections. Furthermore, stop lines and / or directional arrows can be generated at the entrance lanes of each road surface connected to the intersection. Additionally, zebra crossings can be placed at appropriate locations within the intersection area.

[0128] exist Figure 10B In the high-precision map data, four perpendicular, unidirectional vectorized road lines intersect to form four intersection nodes, which collectively represent a real-world crossroads area. First, based on the number of lanes and road function level corresponding to each road line, and using the corresponding lane width (e.g., a standard 3.5-meter lane width), road boundary lines are generated by extending outwards from the reference position of each road line to both sides, forming the road surface geometry. During the road surface generation process, for example, green belts are generated based on the distance between adjacent road surfaces to serve as physical separators for adjacent road surfaces in opposite directions.

[0129] Based on the intersection areas formed by each road surface at the intersection node, the basic range of the intersection surface can be roughly determined. To further optimize the geometry, for example, curve fitting algorithms can be used to smooth the corners of the intersection surface, making the generated intersection surface geometry more consistent with the actual road characteristics.

[0130] Regarding road element configuration, stop lines and / or directional arrows can be generated for entrance roads (lanes) connected to intersections, based on the applicable configuration rules for scenarios near intersections. For multi-lane road structures, differentiated directional signs can be generated for different lanes, taking into account road topology and traffic regulations (such as restrictions prohibiting left turns and U-turns): straight lanes are marked with straight arrows, while turning lanes are marked with turning arrows in the corresponding direction. This ensures that traffic guidance is fully aligned with actual traffic rules.

[0131] exist Figure 10C In the high-precision map data, a horizontally extending bidirectional vectorized road line intersects perpendicularly with two horizontally arranged unidirectional vectorized road lines, converging at two intersection nodes. First, based on the road attributes of each road line (such as the number of lanes and road function level), the corresponding road surface geometry can be generated by expanding the standard lane width. Then, by calculating the intersection area of ​​these road surfaces at the intersection nodes, the spatial extent of the intersection surface is accurately defined. The configuration process of road elements is similar to... Figure 10A and Figure 10B Similarly, I will not go into details here.

[0132] exist Figure 10D In the illustrated embodiment, the high-precision map data records multiple vectorized road lines (two-way and / or one-way) that intersect at the same intersection node at irregular angles to represent complex at-grade intersections in reality. First, the geometric topology of the corresponding road surface can still be generated based on the attribute parameters of each road line.

[0133] Due to the irregularity of road orientation, the boundaries of the intersection areas formed by these road surfaces at intersection nodes cannot be directly used as intersection surface boundaries. Therefore, appropriate geometric algorithms can be used to process the intersection surface boundaries. For example, clipping algorithms can be used to eliminate overlapping phenomena between road surfaces, and smooth connection techniques can be used to optimize the boundaries, ultimately generating an intersection surface geometric topology that conforms to the actual road morphology.

[0134] For this intersection area, the generation process of road elements is similar to... Figure 10A and Figure 10B Similarly, I will not go into details here.

[0135] Figure 11Block diagrams of an apparatus for generating lane-level map data and an apparatus for performing lane-level navigation, according to exemplary embodiments of the present application, are shown.

[0136] exist Figure 11 The diagram illustrates a map provider cloud platform 2, a vehicle manufacturer backend server 5, and a vehicle 1. In this embodiment, a device 20 for generating lane-level map data is deployed on the map provider cloud platform 2. This device 20 includes a processor and a memory (not specifically shown for brevity). The memory stores computer program instructions, which may be stored in a computer-readable storage medium such as a hard disk, RAM, or flash memory card. The processor may be a central processing unit (CPU), microcontroller unit (MCU), graphics processing unit (GPU), neural network processing unit (NPU), digital signal processor (DSP), or other general-purpose processor. When the processor executes the computer program instructions in the memory, it can implement a method for generating lane-level map data.

[0137] The map provider cloud platform 2 includes, for example, a standard-precision map database 210 and a high-precision map database 220. Additionally, the cloud platform 2 may optionally include a fusion map database (not shown in the figure) for fusing standard-precision and high-precision maps (commonly known in the industry as the "OneMap" solution). These map data at different precision levels provide data support for the navigation map generation function of device 20.

[0138] Generally, road-level navigation maps can be generated entirely from high-precision map data. For lane-level map data, the source data can be selected. For example, in road areas covered by high-precision map data, the high-precision map data can be directly provided as lane-level map data. In road areas not covered by high-precision map data, lane-level map data can be generated based on high-precision map data. Furthermore, full-area lane-level map data can also be generated entirely from high-precision map data.

[0139] In device 20, for example, map blocks generated based on multi-source map data can also be fused and stitched together. For example, geometric correction can be performed on different map blocks to ensure a smooth transition of lane directions between blocks. Furthermore, when a conflict is found between a first map block generated based on standard-precision map data and an adjacent second map block generated based on high-precision map data, the road / lane geometry, road element style, and position in the first map block can be automatically adjusted, for example, based on the high-precision map data, to eliminate data inconsistencies.

[0140] Accordingly, a device 10 for lane-level navigation can be deployed in vehicle 1, including a memory and a processor. The memory stores computer program instructions, which, when executed by the processor, enable the processor to perform a method for lane-level navigation. During navigation, the user can flexibly switch between road-level and lane-level navigation modes as needed. Upon receiving a request, the vehicle manufacturer's backend server 5 initiates a corresponding data request to the real-time map service interface of the map data provider's cloud platform 2. Based on the received parameters, the cloud platform 2 generates corresponding lane-level map data and returns it to the backend server 5. This data typically includes map tiles centered on the current location of vehicle 1, or may include map tiles covering the planned route.

[0141] After successfully receiving lane-level map data, vehicle 1 can render the lane-level navigation screen locally through device 10, combining the vehicle's real-time location information and navigation route information, and finally provide visual navigation guidance to the user through the in-vehicle display unit.

[0142] It should be noted that, Figure 11 The system architecture and deployment methods of devices 10 and 20 shown are merely examples. In actual implementation, device 20, used to generate lane-level map data, can also be deployed on the map provider's local server, vehicle terminal, or vehicle manufacturer's backend. Data interaction between vehicles and map providers can also bypass the vehicle manufacturer, and map data distribution can also adopt offline mode or other distribution modes.

[0143] It should be understood that the methods of the various embodiments of this disclosure can be implemented by computer program products / software. This software can be loaded into the processor's working memory and, when run, is used to perform the methods according to the various embodiments of this disclosure.

[0144] It should be understood that the same or similar parts between the various embodiments in this specification can be referred to each other, and each embodiment focuses on describing the differences from other embodiments. In particular, for the apparatus embodiments, since their control logic basically corresponds to that of the method embodiments, the description is relatively brief, and relevant parts can be referred to the description of the method embodiments.

[0145] According to another embodiment of this disclosure, a machine-readable storage medium, such as a CD-ROM, is provided, including a computer program that, when executed, causes a computer or processor to perform methods according to various embodiments of this disclosure. The machine-readable storage medium is, for example, an optical storage medium or a solid-state medium supplied together with or as part of other hardware.

[0146] Although specific embodiments of this application are described in detail herein, they are given for illustrative purposes only and should not be construed as limiting the scope of this application. Various substitutions, modifications, and alterations can be conceived without departing from the spirit and scope of this application.

Claims

1. A method for generating lane-level map data, the method comprising the following steps: Step S1: Obtain high-precision map data, which includes the geometric topology of road lines and road attributes; Step S2: Generate the geometric topology of the road surface based on the geometric topology of the road lines and the road attributes. The geometric topology of the road surface includes at least the road boundary lines. Step S3: Obtain predefined road element configuration rules, and generate corresponding road elements for the geometric topology simulation of the road surface based on the predefined road element configuration rules; as well as Step S4: Generate lane-level map data based on the geometric topology of the road surface and the corresponding road elements.

2. The method according to claim 1, wherein, Step S3 includes: The vectorization mode of road lines is determined based on road attributes, and the vectorization mode includes unidirectional vectorized road lines and bidirectional vectorized road lines; and Based on the road element configuration rules applicable under the determined vectorization mode, the corresponding road elements are generated.

3. The method according to claim 1 or 2, wherein, Step S3 includes: Identify the road scenarios involved in the road surface based on the high-precision map data; and Based on the road element configuration rules applicable to the identified road scenarios, corresponding road elements are generated.

4. The method according to claim 3, wherein, Based on the high-precision map data, the road scenarios involved in the road surface identification include: Based on the spatial proximity between the road represented by the road line and the intersection represented by the intersection node it connects to, the near intersection scenario is identified. Identify consecutive intersection scenarios based on the length of the road between two intersections as represented by the road lines; Based on road length and curvature information, identify long curve scenarios and / or continuous multi-curve scenarios; Based on traffic sign information, identify scenarios where overtaking is prohibited; Based on tunnel and bridge signage information, identify tunnel and / or bridge road segment scenarios; Based on the road function level of the road alignment, identify whether it is a high-speed urban expressway or a non-high-speed urban expressway scenario; and / or Based on the road morphology attributes and road function level of the road line, identify ramp intersection scenarios.

5. The method according to any one of claims 1 to 4, wherein, Generating road elements includes determining the geometric position and style attributes of traffic markings and / or traffic signs; Specifically, the traffic markings include lane lines, road center lines, stop lines, ground directional arrows and / or guide strips, and the style attributes of the traffic markings include shape, line type, color and / or size; Specifically, the traffic signs include traffic signs, road boundary markers, and road dividers, and the style attributes of the traffic signs include type, material, and / or size.

6. The method according to any one of claims 1 to 5, wherein, Step S3 includes: For bidirectional vectorized straight roads: if adjacent intersections, consecutive intersections, or tunnel and / or bridge sections are identified, the road centerline is set to a solid yellow line, and / or the lane lines are set to solid white lines; otherwise, the road centerline is set to a dashed yellow line, and / or the lane lines are set to dashed white lines; and / or For single-direction vectorized straight road lines, if adjacent intersection scenarios, consecutive intersection scenarios, or tunnel and / or bridge sections are identified, the lane lines are set to solid white lines; otherwise, the lane lines are set to dashed white lines.

7. The method according to any one of claims 1 to 6, wherein, Step S3 includes: For bidirectional vectorized curved roads: if a long curve scenario, a continuous multi-curve scenario, or a no-overtaking scenario is identified, set the road centerline to a solid yellow line and / or set the lane lines to solid white lines; otherwise, set the road centerline to a dashed yellow line and / or set the lane lines to dashed white lines; and / or For single-direction vectorized curved road lines: if a long curve scenario, a continuous multi-curve scenario, or a no-overtaking scenario is identified, the lane line is set to a solid white line; otherwise, the lane line is set to a dashed white line.

8. The method according to any one of claims 1 to 7, wherein, Step S3 includes: For adjacent road surfaces generated based on unidirectional vectorized road lines, double yellow lines, green belts, or no separators are generated between the adjacent road surfaces according to the distance between them and the continuous length of that distance. If a non-highway urban expressway scenario is identified, a curbstone is simulated and generated at a preset distance offset outward from the road boundary line; and / or If a ramp intersection scenario is identified, a guide strip is generated in the corresponding road surface bifurcation area.

9. The method according to any one of claims 1 to 8, wherein, Step S2 includes: The number of lanes and the road function level are determined based on road attributes; Based on the number of lanes and the road function level, the road boundary line is generated by extending laterally a predetermined distance from the geometric position of the road line. In particular, the geometric topology of the road surface also includes lane lines, wherein lane lines are generated based on the number of lanes within the area defined by the road boundary lines on both sides.

10. The method according to any one of claims 1 to 9, wherein, Step S2 includes: Based on the height hierarchy and slope attributes contained in the road attributes of the road line, continuous relative height coordinates are added to each point on the geometric topology of the road surface.

11. The method according to any one of claims 1 to 10, wherein, Step S2 includes: Based on the distance between the generated adjacent road surfaces, the adjacent road surfaces are translated; and / or The road lines represented by polylines are subjected to curve fitting, and the geometric topology of the road surface is generated based on the curve-fitted road lines.

12. The method according to any one of claims 1 to 11, wherein, The method further includes the following steps: The high-precision map data also includes the geometric topology of intersection nodes. Based on the geometric topology of road lines, the geometric topology of intersection nodes, and road attributes, the geometric topology of intersection surfaces is generated. Based on predefined road element configuration rules, generate corresponding road elements for the geometric topology of the intersection face; and Based on the geometric topology of the intersection and the corresponding road elements, lane-level map data is generated.

13. The method of claim 12, wherein, The geometric topology of the intersection face is generated in the following way: Generate the geometric topology of road surfaces for all road lines connected to intersection nodes; and The boundary range of the intersection surface is determined based on the intersection area formed at the intersection node by the geometric topology of each road surface.

14. The method according to claim 12 or 13, wherein, The method further includes the following steps: According to the road element configuration rules applicable to the intersection scenario, determine the style attributes and starting point positions of the lane lines and / or lane center lines of the road surface connected to the intersection surface; Generate stop lines and / or directional arrows at the entrance lanes of the road surface connected to the intersection; and / or Determine the geometric location of the zebra crossing within the boundary of the intersection.

15. The method according to any one of claims 1 to 14, wherein, The method further includes the following steps: sending the generated lane-level map data to at least one vehicle for lane-level navigation guidance of the vehicle.

16. The method according to any one of claims 1 to 15, wherein, The method further includes the following steps: Lane-level map data is generated only from standard-precision map data for road areas not covered by high-precision map data; for road areas already covered by high-precision map data, the high-precision map data is directly provided as lane-level map data; and / or The first map block generated based on the standard precision map data in the lane-level map data is stitched together with the second map block generated based on the high precision map data. The first map block and the second map block are adjacent to each other. The stitching process includes at least making the lane direction smoothly connected on the first and second map blocks.

17. A method for performing lane-level navigation, the method comprising the following steps: Obtain lane-level map data generated by the method according to any one of claims 1 to 16; Obtain real-time vehicle location information and navigation route information; Based on the lane-level map data, the vehicle's real-time location information, and the navigation route information, lane-level navigation guidance rendering is performed; as well as The rendering results are then visualized and output.

18. An apparatus for generating lane-level map data, the apparatus comprising a memory and a processor, the memory storing computer program instructions that, when executed by the processor, enable the processor to perform the method according to any one of claims 1 to 16.

19. An apparatus for performing lane-level navigation, the apparatus comprising a memory and a processor, the memory storing computer program instructions that, when executed by the processor, enable the processor to perform the method according to claim 17.

20. A computer program product comprising computer program instructions, wherein, When executed by one or more processors, the computer program instructions enable the one or more processors to perform the method according to any one of claims 1 to 17.

Images