Memory driving simulation test method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of driving simulation testing, in particular to a memory driving simulation testing method and device, electronic equipment and a storage medium. Acquiring original driving data corresponding to the acquired and memorized driving routes of the plurality of target vehicles; the original driving data comprises at least one of original map data, a navigation path map, V2 navigation path data, driving track data and driving route data; building a simulation map based on the original driving data; performing optimization processing on the simulation map to generate a target optimization map; and performing virtual simulation of the vehicle memory driving route based on the target optimization map. And a large number of real vehicles are not needed for field test. Therefore, manpower, material resources and time cost required by real vehicle testing are greatly reduced. Besides, multiple tests can be rapidly carried out through virtual simulation, and the situation that real vehicle tests are limited by factors such as time, sites and weather is avoided, so that the test period is greatly shortened, and the product research and development process is accelerated. And the simulation test of the memory driving route is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of driving simulation testing, and particularly to a method, device, electronic device and storage medium for memory driving simulation testing. Background Art

[0002] With the rapid development of intelligent and automated technologies in the automotive industry, assisted driving functions have gradually become an important configuration of vehicles. As an advanced assisted driving function, memory driving can realize assisted driving on the commuting route according to the remembered starting and ending routes without the user setting the navigation. However, in the process of research and development and optimization of the memory driving function, the testing link faces many challenges.

[0003] Currently, on-road testing is still the main means to verify the quality of memory driving routes. However, with the continuous increase in the number of high-end vehicle users and vehicle models across the country, the number of memory driving routes has increased explosively. If only relying on on-road testing to evaluate the quality of a large number of routes across the country, the cost will rise sharply. On-road testing requires a large amount of manpower and material resources, including professional test drivers, the purchase and maintenance of test vehicles, fuel or electricity consumption, and the rental of test sites. At the same time, the on-road testing cycle is long. From test plan formulation, vehicle preparation, on-site testing to data collection and analysis, each link is time-consuming and laborious, which seriously delays the research and development progress of the memory driving function and cannot quickly respond to market demands.

[0004] Therefore, how to conduct simulation testing of memory driving routes has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, the present invention provides a method, device, electronic device and storage medium for memory driving simulation testing, so as to provide a feasible solution for how to conduct simulation testing of memory driving routes.

[0006] In a first aspect, the present invention provides a method for memory driving simulation testing, the method comprising:

[0007] Obtaining original driving data corresponding to memory driving routes collected by a plurality of target vehicles; the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data;

[0008] Based on the original driving data, building a simulation map;

[0009] Performing optimization processing on the simulation map to generate a target optimized map;

[0010] Performing virtual simulation of the vehicle memory driving route based on the target optimized map.

[0011] The memory driving simulation test method provided by the embodiments of the present application can collect vehicle driving information comprehensively and from multiple perspectives by obtaining various original driving data collected by multiple target vehicles. These multi-source data complement each other, providing rich and accurate basic materials for subsequent simulation and analysis. Based on the original driving data, a simulation map is built, and the simulation map is optimized to generate a target optimized map, which can improve the quality and practicality of the map. Then, vehicle virtual simulation is carried out based on the target optimized map, without the need to use a large number of real vehicles for on-site testing. This greatly reduces the labor, material and time costs required for real vehicle testing. In addition, virtual simulation can quickly conduct multiple tests, unlike real vehicle testing which is limited by factors such as time, site and weather, thus significantly shortening the test cycle and accelerating the product R & D process. The simulation test of the memory driving route is realized.

[0012] In an alternative embodiment, building a simulation map based on the original driving data includes:

[0013] Preprocess the original driving data to obtain the preprocessed original driving data;

[0014] Identify the preprocessed original map data, and based on the identification result, construct an initial road network topology;

[0015] Identify the preprocessed navigation path map and determine the key detail elements in the navigation path map;

[0016] According to the position information corresponding to each key detail element, supplement the key detail elements to the initial road network topology to generate a candidate road network topology;

[0017] Match the V2 navigation path data and the driving trajectory data with the candidate road network topology respectively to generate a navigation trajectory;

[0018] Based on the original map data and the navigation path map, add semantic information to generate a simulation map.

[0019] The memory driving simulation test method provided by the embodiments of this application preprocesses the original driving data to obtain the preprocessed original driving data, ensuring the accuracy of the preprocessed original driving data. Then, it identifies the preprocessed original map data and constructs an initial road network topology based on the identification results. Then, it identifies the preprocessed navigation path map, determines the key detail elements in the navigation path map, and supplements the key detail elements to the initial road network topology according to the position information corresponding to each key detail element to generate a candidate road network topology. Thus, it can accurately restore the actual layout of the road, making the generated candidate road network topology closer to the reality, providing accurate geographical information for subsequent applications. For example, the positioning accuracy of autonomous driving vehicles is improved accordingly. Then, it matches the V2 navigation path data and the driving trajectory data with the candidate road network topology to generate a navigation trajectory, integrating the advantages of different data sources. The V2 navigation path provides accurate navigation information, and the driving trajectory reflects the actual driving path. When both are matched with the road network, the map is corrected and improved, enhancing the accuracy and reliability of the map and ensuring the accuracy of navigation and path planning. Based on the original map data and the navigation path map, semantic information is added to generate a simulation map, increasing the semantic understanding ability of the map. The semantic information covers road types, traffic rules, etc., providing a decision-making basis for the intelligent transportation system, enabling the auxiliary driving system to make reasonable decisions and improving driving safety and efficiency.

[0020] In an alternative embodiment, the simulation map is optimized to generate a target optimized map, including:

[0021] Perform topological optimization on the simulation map to obtain a first optimized map;

[0022] Perform geometric optimization on the first optimized map to obtain a second optimized map;

[0023] Perform semantic optimization on the second optimized map to generate a target optimized map.

[0024] The memory driving simulation test method provided by the embodiment of the present application obtains a first optimized map by topologically optimizing the simulation map, which can ensure that the connection and traffic relationship between elements such as roads and intersections conform to the actual situation. For example, the connectivity of the road is accurately set, and the turning rules of the vehicle at the intersection are clarified, which provides a reliable infrastructure for the navigation system and the automatic driving algorithm, making the path planning more reasonable and efficient, and reducing navigation errors and driving conflicts. The geometric shape and position deviation of the map elements can be corrected by geometrically optimizing the first optimized map. This not only improves the visualization effect of the map, allowing users to obtain geographic information more intuitively and accurately when viewing the map, but also enhances the accuracy of the map in measurement, geographic analysis, etc. The second optimized map is semantically optimized, and rich semantic information is added to the second optimized map to generate a target optimized map. The target optimized map after multiple optimizations can play an important role in multiple fields such as intelligent transportation, automatic driving, geographic information system, and urban planning. It provides a unified and high-quality map data foundation for different fields, and promotes data sharing and collaborative cooperation between various fields. Reduce costs and improve efficiency, high-quality target optimized maps can reduce resource waste and inefficiency caused by inaccurate maps.

[0025] In an optional implementation, topology optimization is performed on the simulation map to obtain a first optimized map, including:

[0026] Identify the simulated map and detect lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulated map;

[0027] According to the search results, the simulation map is topologically optimized to obtain a first optimized map.

[0028] The memory driving simulation test method provided by the embodiments of this application recognizes the simulation map, detects the lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulation map, and can timely discover connection problems existing in the map. Such as lane disconnection, poor intersection traffic flow, or missing road network, etc. According to the retrieval results, the simulation map is topologically optimized to obtain the first optimized map. After solving these problems through topological optimization, the generated first optimized map can provide more accurate road connection information for the path planning algorithm. For example, in an autonomous driving or navigation system, a vehicle can plan a more reasonable driving route based on the optimized map, avoiding route errors or unreasonable planning caused by map connectivity problems, and improving travel efficiency. Accurate lane and intersection connectivity and a complete road network are crucial for traffic simulation. In scenarios such as traffic flow simulation and urban traffic planning, simulation based on the optimized map can more realistically reflect the actual traffic conditions. For example, simulating the traffic flow of vehicles at complex intersections and the traffic congestion degree of different sections, providing more reliable data support for traffic planners, so as to formulate more effective traffic management strategies and improve the urban traffic congestion situation. In addition, during the process of map development and related technology research and development, early discovery and solution of map connectivity and integrity problems can avoid a large amount of rework and cost increase caused by map errors in the later stage. For example, in the memory driving simulation test, if a map with connectivity problems is used, it may lead to inaccurate test results and require retesting, wasting a lot of time and resources. By optimizing the map, the accuracy and reliability of the test can be improved, and the development and test costs can be reduced.

[0029] In an alternative embodiment, vehicle memory driving route virtual simulation based on the target optimized map includes:

[0030] Perform basic static inspection on the target optimized map to obtain the detection score corresponding to the target optimized map;

[0031] If the detection score is higher than or equal to the preset score threshold, perform vehicle memory driving route virtual simulation based on the target optimized map to generate a virtual simulation result;

[0032] Evaluate the virtual simulation result based on preset evaluation indicators; the preset evaluation indicators include at least one of a safety indicator, a positioning accuracy rate indicator, and a path planning rationality indicator.

[0033] The memory driving simulation test method provided by the embodiments of this application performs basic static inspections on the target optimized map and obtains inspection scores, which can discover potential problems in the map before virtual simulation, such as incorrect map elements and unreasonable topological structures. If these problems are not discovered, they may cause navigation errors, abnormal driving, etc. for the vehicle during virtual simulation or actual driving. Through the inspection, the map quality is ensured, providing a reliable basis for subsequent virtual simulation and indirectly guaranteeing the safety and stability of vehicle driving. Then, the virtual simulation results are evaluated based on preset evaluation indicators, including safety indicators (such as the number of user takeovers per 100 kilometers, the number of emergency brakes and hard brakes), positioning accuracy indicators, path planning rationality indicators, etc., which quantitatively evaluate the driving performance of the vehicle in the virtual environment from multiple dimensions. For example, by analyzing the positioning accuracy indicator, the positioning accuracy of the vehicle during the simulation process can be judged. Inaccurate positioning may cause the vehicle to deviate from the route, and this evaluation process can timely discover such problems, providing a basis for optimizing the vehicle positioning system and improving the safety and reliability of the vehicle during actual driving. In addition, a comprehensive evaluation of the vehicle performance in the virtual simulation link can quickly discover problems existing in the vehicle, such as unreasonable path planning. R & D personnel can optimize the vehicle algorithm and system targeted according to the virtual simulation results, avoiding discovering problems only in the real vehicle test stage, thus saving a large amount of time and cost and accelerating the vehicle R & D cycle. For example, if the virtual simulation shows that a certain route planning causes the vehicle to detour frequently, the R & D personnel can adjust the path planning algorithm in time to improve the driving efficiency of the vehicle. Virtual simulation can, to a certain extent, replace part of the real vehicle test. By evaluating the virtual simulation results, solutions with better performance can be initially screened out, reducing the number of unnecessary real vehicle tests. Real vehicle tests require a large amount of manpower, material resources and financial resources, such as the purchase and maintenance of test vehicles, the lease of test sites, etc., while virtual simulation only needs to be carried out in a computer environment, reducing the test cost.

[0034] In an alternative embodiment, the method further includes:

[0035] If the inspection score is lower than the preset score threshold, continue to obtain other original driving data, and re-optimize the target optimized map based on the other original driving data until the inspection score of the target optimized map is higher than the preset score threshold; the other original driving data is the original driving data collected again.

[0036] In the memory driving simulation test method provided in the embodiment of the present application, if the detection score is lower than the preset score threshold, it indicates that the map is insufficient, and other original driving data are continuously obtained and the map is optimized again, so that the accuracy and completeness of the map can be continuously improved. The newly collected data can supplement the missing information and correct the erroneous elements, such as improving the topological structure of the road network, updating the location of traffic signs, etc., so that the map is more in line with the actual situation and provides a reliable basis for vehicle driving and navigation. In addition, as time goes by and the road environment changes, the map needs to be continuously updated. This optimization mechanism based on new data can enable the target optimized map to keep up with the actual changes and maintain a good quality state. Whether it is the construction of new urban roads, the adjustment of traffic rules, or the accuracy improvement requirements of the map itself, it can be met by continuously obtaining new data to optimize the map, ensuring the reliability of the map in long-term use. In addition. Accurate maps are the key to vehicle virtual simulation. The re-optimized map can provide a more accurate environmental model for virtual simulation, so that the behavior of the vehicle in the simulation is closer to the real situation. The optimized map is used for virtual simulation, which can more accurately evaluate the performance indicators of the vehicle, which helps R&D personnel to more accurately find problems with the vehicle and improve the vehicle system and algorithm in a targeted manner. High-quality maps can reduce the number and time of vehicle real vehicle testing. Through virtual simulation, full testing based on optimized maps can be carried out to discover and solve potential problems in advance, reducing invalid tests and repeated tests caused by map problems in real vehicle testing. Real vehicle testing is costly, including vehicle wear and tear, test site fees, and labor costs. Reducing real vehicle testing can effectively reduce overall R&D costs.

[0037] In an optional implementation, a basic static check is performed on the target optimization map to obtain a detection score corresponding to the target optimization map, including:

[0038] Performing static element detection on the target optimized map to obtain an element score; the element score includes at least one of a basic element score, a full navigation score, and a semantic information score;

[0039] Performing a legitimacy check on the target optimized map to obtain a legitimacy score; the legitimacy score includes at least one of a boundary legitimacy score, a road legitimacy score, a lane legitimacy score, and a semantic legitimacy score;

[0040] Based on the element score and the legitimacy score, the detection score corresponding to the target optimization map is obtained.

[0041] The memory driving simulation test method provided by the embodiments of the present application can deeply evaluate the accuracy and integrity of each element of the map by performing element static detection on the target optimized map to obtain basic element scores, full-navigation scores, semantic information scores, etc. For example, in the basic element score, it is possible to check whether there are errors in elements such as anchor points, roads, intersections, etc., and timely discover problems such as "the anchor in the map pb is 0,0,0" and "the number of intersection vertices <= 3"; the full-navigation score can detect situations such as V2 deviating from the trajectory and a large difference between full-navigation and V2. This ensures that the map elements conform to the actual situation and provides accurate data support for vehicle navigation, path planning, etc. The boundary legality score, road legality score, lane legality score, semantic legality score, etc. obtained from the legality detection ensure that the map follows the established rules and standards from different perspectives. For example, the boundary legality score can detect whether the boundary ID is unique and whether the layer_type is legal; the road legality score can determine whether the roadID is legal and whether there is a missing boundary or lane, etc. By ensuring the legality of the map, it is possible to avoid abnormal vehicle driving caused by map data errors and improve the reliability of the map. Based on the element score and the legality score, a detection score is obtained, realizing a comprehensive evaluation of the map. This multi-dimensional scoring method can more accurately reflect the overall quality of the map. R & D personnel can optimize the map targeted according to the score, timely discover and solve potential problems, and ensure the accuracy and stability of the map in various application scenarios.

[0042] In a second aspect, the present invention provides a memory driving simulation test device, which includes:

[0043] An acquisition module, configured to acquire original driving data corresponding to the memory driving routes collected by multiple target vehicles; the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data;

[0044] A construction module, configured to construct a simulation map based on the original driving data;

[0045] An optimization module, configured to perform optimization processing on the simulation map to generate a target optimized map;

[0046] A simulation module, configured to perform virtual simulation of the vehicle memory driving route based on the target optimized map.

[0047] The memory driving simulation test device provided by the embodiments of this application can collect vehicle driving information comprehensively and from multiple perspectives by obtaining various original driving data collected by multiple target vehicles. These multi-source data complement each other, providing rich and accurate basic materials for subsequent simulation and analysis. Based on the original driving data, a simulation map is built, and the simulation map is optimized to generate a target optimized map, which can improve the quality and practicality of the map. Then, vehicle virtual simulation is carried out based on the target optimized map, eliminating the need to use a large number of real vehicles for on-site testing. This greatly reduces the human, material, and time costs required for real vehicle testing. In addition, virtual simulation can quickly conduct multiple tests, unlike real vehicle testing which is limited by factors such as time, site, and weather, thus significantly shortening the test cycle and accelerating the product R & D process.

[0048] In a third aspect, the present invention provides a computer device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the memory driving simulation test method according to the first aspect or any corresponding embodiment thereof.

[0049] In a fourth aspect, the present invention provides a computer-readable storage medium, on which computer instructions are stored. The computer instructions are used to cause a computer to perform the memory driving simulation test method according to the first aspect or any corresponding embodiment thereof.

[0050] In a fifth aspect, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to perform the memory driving simulation test method according to the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0052] Figure 1 is a flowchart of the memory driving simulation test method according to an embodiment of the present invention;

[0053] Figure 2 is a flowchart of another memory driving simulation test method according to an embodiment of the present invention;

[0054] Figure 3 is a structural block diagram of the memory driving simulation test device according to an embodiment of the present invention;

[0055] Figure 4 It is a schematic diagram of the hardware structure of the computer device according to an embodiment of the present invention. Specific implementation manners

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0057] With the rapid development of intelligent and automated technologies in the automotive industry, assisted driving functions have gradually become an important configuration of vehicles. As an advanced assisted driving function, memory driving can achieve assisted driving on the commuting route according to the memorized start and end routes without the need for users to set navigation. However, during the research and development and optimization of the memory driving function, the testing process faces many challenges.

[0058] Currently, on-road vehicle testing is still the main means to verify the quality of memory driving routes. However, with the continuous increase in the number of high-end vehicle users and vehicle models nationwide, the number of memory driving routes has increased explosively. If only on-road vehicle testing is relied on to evaluate the quality of a huge number of routes nationwide, the cost will increase sharply. On-road vehicle testing requires a large amount of manpower and material resources, including professional test drivers, the purchase and maintenance of test vehicles, fuel or electricity consumption, and the rental of test sites. At the same time, the on-road vehicle testing cycle is long. From test plan formulation, vehicle preparation, on-site testing to data collection and analysis, each link is time-consuming and laborious, which seriously delays the research and development progress of the memory driving function and cannot quickly respond to market demands.

[0059] Therefore, how to conduct simulation testing of memory driving routes has become an urgent problem to be solved.

[0060] It should be noted that for the method for memory driving simulation testing provided in the embodiments of the present application, the execution subject may be a device for memory driving simulation testing. The device for memory driving simulation testing can be implemented as part or all of a computer device through software, hardware, or a combination of software and hardware. Among them, the computer device can be a server or a terminal. Among them, the server in the embodiments of the present application can be a single server or a server cluster composed of multiple servers. The terminal in the embodiments of the present application can be other intelligent hardware devices such as a smart phone, a personal computer, a tablet computer, a wearable device, and a smart robot. In the following method embodiments, the execution subject is taken as an electronic device for illustration.

[0061] According to an embodiment of the present invention, an embodiment of a memory driving simulation test method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0062] In this embodiment, a memory driving simulation test method is provided, which can be used in the above-mentioned electronic device. Figure 1 It is a flowchart of the memory driving simulation test method according to an embodiment of the present invention, as Figure 1 shown, this process includes the following steps:

[0063] Step S101, obtain the original driving data corresponding to the memory driving route collected by multiple target vehicles.

[0064] Among them, the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data.

[0065] Specifically, each target vehicle actively turns on the memory driving button before collecting data, so as to save the collected original driving data during the vehicle driving process. Then, the target vehicle transmits the collected original driving data to the electronic device.

[0066] Among them, the original map data is the basic geographical information data about the vehicle driving area. It covers the basic shape, position, and orientation of the road, as well as the surrounding terrain, landforms, landmark buildings, etc. information. For example, it is collected by devices such as satellite remote sensing and vehicle-mounted lidar, and records the curvature and slope of the road, as well as the relative position relationship with the surrounding natural and man-made environments. The original map data is the cornerstone for constructing and optimizing the map. In vehicle navigation, it provides a geographical framework for path planning; in the field of autonomous driving, it helps the vehicle perceive the surrounding environment, perform positioning and decision-making. That is to say, the original map data includes the input files required for building the simulation map. For example, the smooth pose, global pose, traffic lights, boundaries, intersections, crossroads, road markings, dotted lines of the target vehicle.

[0067] The navigation path map is the intersection map saved in the memory route, which comes from OLM and UMapNet. Currently, it saves four elements: Boundary, StopLine, Crosswalk, and Arrow. The coordinates are in the Map Frame coordinate system based on the Anchor point. The fngp map outputs two formats: protobuffer and flatbuff. Among them, fngp_map.bin is the default flatbuff format, and fngp_map.pb.bin is the protobuffer format.

[0068] The V2 navigation path data is the concatenation result of v2 in the saved memory route, sourced from RefinedV2Map of V2 Printing, and the final result is saved in the paths field of V2NaviPath, where routing_path_index is the index of the main path.

[0069] The driving trajectory data is the driving trajectory on the saved memory route, sourced from LocalPoseMsg, mainly including smooth pose and global pose.

[0070] The driving route data is for the route information management of fngp, mainly saving the route id, the start and end points of the route, and the route mileage, etc.

[0071] Step S102: Based on the original driving data, build a simulation map.

[0072] Specifically, the electronic device can identify the original driving data, and then build a simulation map according to the identification result.

[0073] This step will be introduced in detail below.

[0074] Step S103: Optimize the simulation map to generate a target optimized map.

[0075] Specifically, the electronic device can detect the simulation map, and then, according to the detection result, process the simulation map.

[0076] This step will be introduced in detail below.

[0077] Step S104: Based on the target optimized map, conduct virtual simulation of the vehicle's memory driving route.

[0078] Specifically, after generating the target optimized map, the electronic device can conduct vehicle virtual simulation based on the target optimized map.

[0079] This step will be introduced in detail below.

[0080] The memory driving simulation test method provided by the embodiments of the present application can collect vehicle driving information comprehensively and from multiple perspectives by obtaining various original driving data collected by multiple target vehicles. These multi-source data complement each other, providing rich and accurate basic materials for subsequent simulation and analysis. Based on the original driving data, a simulation map is built, and the simulation map is optimized to generate a target optimized map, which can improve the quality and practicality of the map. Then, vehicle virtual simulation is performed based on the target optimized map, eliminating the need for a large number of real vehicles for on-site testing. This greatly reduces the human, material, and time costs required for real vehicle testing. In addition, virtual simulation can quickly conduct multiple tests, unlike real vehicle testing which is limited by factors such as time, site, and weather, thus significantly shortening the test cycle and accelerating the product R & D process. Simulation test of memory driving route

[0081] In this embodiment, a memory driving simulation test method is provided, which can be used in the above-mentioned electronic device, Figure 2 is a flowchart of the memory driving simulation test method according to an embodiment of the present invention, as Figure 2 shown, this process includes the following steps:

[0082] Step S201, obtain the original driving data corresponding to the memory driving route collected by multiple target vehicles.

[0083] Among them, the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data.

[0084] For this step, please refer to the introduction of step S101 above and will not be elaborated here.

[0085] Step S202, build a simulation map based on the original driving data.

[0086] Specifically, the above step S202 may include the following steps:

[0087] Step S2021, preprocess the original driving data to obtain the preprocessed original driving data.

[0088] Specifically, the electronic device can identify the original driving data, determine the error data, duplicate data, and missing data included in the original driving data. Then, through data cleaning, identify and eliminate the error data caused by reasons such as sensor failures and communication interferences, such as vehicle speeds and position information that deviate significantly from the normal range. Use a duplicate checking algorithm to find and delete duplicate records to avoid the interference of data redundancy on subsequent analysis. For missing data, use interpolation methods, fill it according to historical data or similar data to ensure the integrity of the data. For example, if the vehicle position data at a certain moment is missing, the missing position can be estimated by linear interpolation based on the position and speed information at the previous and subsequent moments.

[0089] Next, since the data formats collected by different sensors and devices may be different, it is necessary to convert the original driving data into a unified format. Unify the position data collected by GPS devices from different manufacturers into the standard latitude and longitude format; unify the timestamps output by various sensors into the same time standard for the convenience of data integration and analysis. This can ensure the compatibility and consistency of the data in the subsequent processing and analysis process.

[0090] In addition, the original data collected often has noise, which affects the accuracy of the data. The electronic device can use filtering algorithms such as Kalman filtering and Gaussian filtering to denoise the data. When processing vehicle driving trajectory data, filter out the small fluctuations caused by sensor accuracy problems through filtering to make the trajectory smoother and more accurate, providing more reliable data support for subsequent path planning and map construction.

[0091] Finally, the electronic device normalizes the data with different magnitudes and distribution ranges, maps the vehicle speed data uniformly to the interval [0,1] to make the data comparable. This helps to improve the performance of machine learning algorithms and data analysis models, and improve the efficiency and accuracy of data processing. Especially in the multi-source data fusion analysis, it can avoid analysis deviations caused by data magnitude differences.

[0092] Step S2022: Identify the preprocessed original map data, and based on the identification results, construct an initial road network topology structure.

[0093] Specifically, the electronic device can use technologies such as image recognition and pattern recognition to extract key features from the preprocessed original map data. For map image data containing road information, use an edge detection algorithm (such as Canny edge detection) to extract the edge contours of the roads, and detect the straight line features through the Hough transform to identify the directions of the roads; directly obtain the attribute information such as the coordinate point sequence and road type of the roads from the vector map data. These features are the basic elements for constructing the road network topology structure.

[0094] Then, based on the extracted key features, the electronic device determines the nodes and edges in the road network. Among them, the intersections, endpoints, etc. of the roads are used as nodes, and the road segments connecting these nodes are defined as edges. During the recognition process, by combining the geographical coordinate information and topological relationship information in the map data, the positions of the nodes and the connection relationships of the edges are accurately determined. For example, at an intersection, the meeting point of four roads is determined as a node, and the four road segments are respectively used as the edges connecting this node to other nodes.

[0095] Finally, based on the determined nodes and edges, the electronic device constructs an initial road network topological structure. Specifically, the electronic device uses a graph data structure to represent the topological structure, where the nodes are the vertices of the graph and the edges are the edges of the graph. Corresponding attributes are assigned to each node and edge, such as the geographical location of the node, traffic flow information, the road length, number of lanes, speed limit, etc. of the edge. In this way, an initial topological structure that can describe the connection relationships and basic attributes of the road network is formed.

[0096] Step S2023: Identify the preprocessed navigation path map to determine the key detail elements in the navigation path map.

[0097] Among them, the key detail elements include but are not limited to road boundaries, stop lines, crosswalks, arrow markings, traffic lights, intersections, and lane lines.

[0098] Specifically, the electronic device can use a target recognition algorithm to identify the preprocessed navigation path map to determine the key detail elements in the navigation path map.

[0099] Among them, the key detail elements include but are not limited to road boundaries, stop lines, crosswalks, arrow markings, traffic lights, intersections, and lane lines.

[0100] Exemplarily, the electronic device can use an edge detection algorithm, such as the Canny algorithm, to identify the road boundaries in the navigation path map. This algorithm is based on the gray-scale change of the image, determines the possible edge points by calculating the gradient magnitude and direction, and then uses non-maximum suppression and double-threshold processing to accurately extract the edge contour of the road. At the same time, by combining the geographical coordinate information and vector data in the map, the specific position and orientation of the road boundary are determined, the scope of the road is clarified, providing a boundary constraint for vehicle driving to prevent the vehicle from deviating from the road.

[0101] An electronic device can separate the stop line from the map background through a color threshold segmentation method. Then, morphological operations (such as dilation and erosion) are used to process the segmented image to remove noise and enhance the features of the stop line. Finally, line detection algorithms such as the Hough transform are adopted to determine the position and direction of the stop line. The accurate identification of the stop line is crucial for the safe stop and start of vehicles at intersections, and can effectively prevent vehicles from illegally crossing the line.

[0102] An electronic device can use feature extraction methods based on color and texture, such as Local Binary Pattern (LBP), to extract the texture information of the crosswalk. Then, combined with machine learning algorithms (such as Support Vector Machine (SVM)), the extracted features are classified and identified to determine whether there is a crosswalk in the map. In addition, the position of the crosswalk can be further confirmed by analyzing the relationship between the roads and traffic elements in adjacent areas. The identification of the crosswalk helps to remind drivers to pay attention to pedestrian safety and protect the passing rights of pedestrians.

[0103] Arrow marking recognition: Arrow markings are used to indicate the driving direction of vehicles. An electronic device can first preprocess the map image through image recognition technology to enhance the contrast of the arrow. Then, using the method of template matching, the pre-set arrow template is matched with the target in the map image. According to the matching result and similarity score, the position, direction, and type (such as straight arrow, turning arrow, etc.) of the arrow marking are determined. The accurate identification of arrow markings can provide clear driving direction guidance for vehicles, guide vehicles to drive correctly, and reduce traffic chaos.

[0104] For traffic signal recognition, traffic signals usually exist in the form of specific icons on the map. An electronic device can adopt object detection algorithms, such as the YOLO (You Only Look Once) algorithm or the Faster R-CNN algorithm based on Convolutional Neural Network (CNN), to detect the navigation path map. These algorithms learn the features of traffic signals through training on a large number of images containing traffic signals, so as to accurately identify the position, color (red light, green light, yellow light), and status (whether it is flashing) of traffic signals in the map. The recognition result of traffic signals is very crucial for the driving decision of vehicles, and can help vehicles reasonably plan the driving speed and parking time, improving the road traffic efficiency and safety.

[0105] For intersection recognition: Intersections are important nodes in the road network with complex topological structures. First, the electronic device can identify possible intersection areas by analyzing the connection relationships and geometric shapes of roads. Then, using graph theory algorithms such as Dijkstra's algorithm or A* algorithm, analyze the road connection conditions at the intersections to determine the types of intersections (such as cross-shaped, T-shaped, circular, etc.). At the same time, combining other information in the map, such as traffic signs and lane lines, further improve the recognition results of intersections. The accurate recognition of intersections is crucial for route planning and traffic flow management, and can help vehicles choose appropriate driving routes at complex intersections.

[0106] For lane line recognition: Lane lines are used to divide different lanes and guide vehicles to drive orderly. In the navigation path map, lane lines usually appear as continuous or discontinuous line segments. The electronic device can adopt methods based on edge detection and tracking. First, use edge detection algorithms to extract the edges of lane lines, and then through tracking algorithms (such as Kalman filter tracking algorithm), track the position and shape changes of lane lines in consecutive map image frames. In addition, deep learning algorithms such as semantic segmentation networks (such as U-Net) can also be combined to perform pixel-level segmentation recognition of lane lines, improving the accuracy and robustness of recognition. The recognition results of lane lines provide important bases for functions such as lane keeping and lane change assistance of vehicles, ensuring the safe driving of vehicles within the lanes.

[0107] The electronic device can also recognize other key detail elements, which will not be specifically introduced here.

[0108] Step S2024, according to the position information corresponding to each key detail element, supplement the key detail elements to the initial road network topological structure to generate a candidate road network topological structure.

[0109] Specifically, in the initial road network topological structure, the electronic device makes precise matches based on the position information of the key detail elements and the initial road network topological structure.

[0110] Specifically, for road boundaries, the electronic device can determine their corresponding positions in the topological structure by comparing their coordinate information with the positions of road edges in the topological structure. If the road boundary highly coincides with the orientation and position of a certain road edge, the road boundary can be associated with this road edge. For stop lines, using the intersection or section where they are located as a clue, find the corresponding nodes or edges in the topological structure. If the stop line is on a specific entrance road of a certain intersection, determine the position of the stop line near the edge where the intersection is connected to the corresponding road in the topological structure. Elements such as crosswalks, arrow markings, traffic lights, intersections, and lane lines also use similar methods to find the matching nodes, edges, or regions in the topological structure using their position information.

[0111] After determining the positions of all key detail elements in the initial road network topology, the electronic device can incorporate these elements into the topology. Taking intersections as an example, if in the initial topology, a certain area is identified as an intersection but does not contain detailed internal structure information, at this time, according to the intersection information in the key detail elements, supplement the internal lane connection relationships, turning rules, etc. Connect the different entrance and exit lanes of the intersection to the corresponding road edges in the topology, clarify the passing relationships between the lanes, and make the topology more accurately reflect the actual situation of the intersection. For lane lines, based on the existing road edges, subdivide the lane information of the road edges according to the position and number of lane lines. If a road edge was originally represented as a single lane in the topology but the actual lane lines show a double lane, modify the lane attribute of this road edge and add information related to lane lines, such as the type and solidity of lane dividers.

[0112] While supplementing the key detail elements, the electronic device can update the attributes of relevant nodes and edges in the topology. For nodes or edges containing traffic lights, add attributes such as the status and cycle of the traffic lights. If there is a traffic light at a certain intersection, record information such as the color change cycle and current status (red light, green light, or yellow light) of the traffic light in the node attributes of this intersection in the topology, so as to consider the traffic light factor during subsequent traffic flow simulation and route planning. For road edges with crosswalks, update their pedestrian passage-related attributes, such as whether pedestrians are allowed to pass and the width of the crosswalk. For road edges with arrow markings, update the driving direction attribute according to the arrow direction to clarify the driving direction restrictions of vehicles on this section of the road. Through these attribute updates, the candidate road network topology not only contains richer road details but also better reflects the actual traffic rules and conditions.

[0113] In step S2025, match the V2 navigation path data and the driving trajectory data with the candidate road network topology respectively to generate a navigation trajectory.

[0114] Specifically, the electronic device can first parse the V2 navigation path data to extract key information such as the coordinates, driving direction, and section length of path nodes. This data is usually stored in a specific format and needs to be parsed according to its data structure to convert it into a form convenient for processing. Using the parsed data, extract features that can be used for matching. For example, the coordinates of path nodes can be directly used as position features, and the driving direction can be used as a direction feature.

[0115] Then, the electronic device matches the extracted V2 navigation path data features with the candidate road network topology. By calculating the distances between the path node coordinates and the road edges and nodes in the topology, it determines whether the path node is located on a certain road. If the distance between the coordinates of a path node and a certain road edge is within a certain threshold range, it is considered that the node matches this road edge. At the same time, combined with the driving direction feature, the accuracy of the match is further confirmed. If the driving direction in the V2 navigation path data is consistent with the allowed driving direction of the corresponding road edge in the topology, the credibility of the match is enhanced. During the matching process, the connection relationship of the roads also needs to be considered to ensure that the path in the V2 navigation path data can continuously find the corresponding road in the topology.

[0116] Due to data acquisition errors or the complexity of the actual road conditions, inaccurate matching may occur. Therefore, the electronic device needs to optimize and adjust the matching results. The dynamic programming algorithm can be used to search for the optimal matching path within a certain range to solve the problem of poor local matching. If a deviation is found in the matching between a certain path node and the road edge in the topology during the matching process, the matching strategy can be adjusted, such as expanding the distance threshold, considering the connectivity of the surrounding roads, etc., and rematching to make the matching between the V2 navigation path data and the candidate road network topology more accurate.

[0117] For the driving trajectory data, the electronic device can preprocess the driving trajectory data, including operations such as data cleaning, denoising, and interpolation. Data cleaning is used to remove outliers, such as points that deviate significantly from the normal driving trajectory due to sensor failures; denoising can use filtering algorithms, such as Kalman filtering, to remove the noise interference in the data and make the trajectory smoother; interpolation is used to supplement the missing trajectory points to ensure the continuity of the trajectory. Through these preprocessing operations, the quality of the driving trajectory data is improved, providing a reliable data basis for subsequent matching.

[0118] Similar to the matching of V2 navigation path data, the driving trajectory data is first matched with the candidate road network topology according to the location information. By calculating the distances between the coordinates of the trajectory points and the road edges and nodes in the topology, the road where the trajectory points are located is determined. At the same time, considering the timestamp information of the trajectory points and combining with the driving speed of the vehicle, it is judged whether the driving order and position change of the trajectory points on the road are reasonable. If the time interval between the timestamps of a certain trajectory point and the previous trajectory point is short, but the position has changed significantly, exceeding the range that the normal driving speed of the vehicle can reach, then there may be a matching error and it needs to be corrected.

[0119] Since the driving trajectory is dynamically changing, a vehicle may be affected by various factors during driving, such as traffic congestion and temporary route changes. Therefore, during the matching process, the electronic device needs to track the driving state of the vehicle in real time and continuously adjust the matching result according to the new trajectory points. A real-time updated matching algorithm can be adopted, such as a matching algorithm based on particle filtering. According to the current matching result and the information of the new trajectory points, the next position of the vehicle is predicted and matched in the topological structure, so that the matching result can reflect the actual driving situation of the vehicle in a timely manner.

[0120] Finally, the electronic device fuses the matching results of the V2 navigation path data and the driving trajectory data. The V2 navigation path data provides global navigation information, while the driving trajectory data reflects the actual driving situation of the vehicle. By fusing these two types of data, their advantages can be fully utilized to improve the accuracy and reliability of the navigation trajectory. During the fusion process, weighted processing is performed according to the credibility and real-time nature of the data. If the update frequency of the V2 navigation path data is high and its accuracy is reliable, a higher weight is given to it during fusion; for the driving trajectory data, although it can better reflect the real-time position of the vehicle, there may be certain errors, so an appropriate weight is given according to its error range. Then, the electronic device generates a navigation trajectory based on the fused data. The navigation trajectory should not only include the driving path of the vehicle but also consider information such as driving direction and speed changes. When generating the navigation trajectory, the trajectory is optimized to make it smoother and more reasonable. Methods such as spline curve fitting can be used to fit the discrete trajectory points to generate a continuous trajectory curve. At the same time, in combination with traffic rules and road conditions, the trajectory is adjusted. For example, at intersections, the driving direction is adjusted according to traffic lights and turning rules, and the driving speed is adjusted in speed-limited sections, so that the generated navigation trajectory conforms to the actual driving scenario.

[0121] Step S2026: Based on the original map data and the navigation path map, add semantic information to generate a simulation map.

[0122] Specifically, the electronic device can extract various semantic features from the original map data and the navigation path map. Among them, the semantic information can include key node semantics, such as intersections, toll stations, service areas, etc., and their functions are marked. From the perspective of traffic rules, rule information such as speed limits, bans, and one-way streets is extracted, and this information is crucial for simulating the driving behavior of vehicles on the road.

[0123] Then, the electronic device correlates and integrates the semantic information extracted from the original map data and the navigation path map. Based on geographical coordinates, the geographical elements in the original map are matched with the corresponding positions in the navigation path. The position information of a specific intersection in the original map is combined with the navigation semantic information of this intersection in the navigation path map to ensure the consistency of the semantic information in the map space.

[0124] The electronic device performs semantic annotation on each element in the map to generate a simulation map. For roads, annotate their types (such as highways, urban arterials, rural paths), the number of lanes, road surface conditions, etc.; for intersections, annotate intersection types (cross-shaped, T-shaped, roundabout, etc.), turning rules, signal control methods, etc. Using a vector data structure or a raster data structure, embed the semantic information into the map data so that the map data not only contains geometric information but also has rich semantic descriptions.

[0125] Optionally, the electronic device can perform visual design based on the simulation map after adding semantic information. Set different map symbols, colors, and annotation styles according to the semantic information to improve the readability of the map. Represent parks in green, water bodies in blue, and use icons of specific shapes to represent different types of buildings; for navigation paths, use lines of different colors to represent unobstructed and congested sections according to the driving direction and road condition information. At the same time, reasonably arrange the annotation positions to avoid information overlap and make the map more visually clear and intuitive.

[0126] Step S203, perform optimization processing on the simulation map to generate a target optimized map.

[0127] Specifically, the above step S203 may include the following steps:

[0128] Step S2031, perform topological optimization on the simulation map to obtain a first optimized map.

[0129] Specifically, the above step S2031 may include the following steps:

[0130] Step a1, identify the simulation map and detect the lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulation map.

[0131] Specifically, the electronic device can track along the lane lines through image recognition technology. Starting from a starting point of a certain lane in the simulation map, the continuity of the lane lines is detected section by section according to the direction of the lane lines. If during the tracking process, it is found that the lane lines are interrupted, missing, or have an obviously incorrect connection, mark this position as a possible connectivity problem point. For example, on a normal straight lane, if a section of the lane line suddenly disappears for a certain distance and then reappears not far away, this may mean that there is a lane connectivity problem here. Then, the electronic device identifies the lane change areas in the simulation map, such as the dashed line areas or specific lane change signs. Check whether the lane lines in these areas are clear and standardized, and whether the connection with the adjacent lanes is correct. For multi-lane roads, ensure that in the lane change areas, vehicles can legally and smoothly change from one lane to another. For example, at the ramp where the expressway merges into the main road, it is necessary to detect whether the connection between the ramp lane and the main road lane is reasonable and whether there is a situation where vehicles cannot merge normally. In addition, for the entrances and exits of places such as parking lots, communities, and service areas, detect their connectivity with the lanes of the external roads. Ensure that the lanes at the entrances and exits match the lanes of the external roads in terms of position, width, slope, etc., so that vehicles can enter and exit safely and conveniently. For example, the connection between the lane at the community entrance and the urban road should comply with traffic regulations to avoid problems such as sharp turns or too narrow lanes that affect vehicle passage. Thus, the electronic device completes the detection of the lane connectivity of the simulation map.

[0132] For the detection of intersection connectivity, the electronic device can first identify and analyze the geometric shape of the intersections in the simulation map to determine their types, such as cross-shaped, T-shaped, circular, etc. Then check whether the road connections in all directions of the intersection conform to the specifications of this type of intersection. For example, the four roads at a cross-shaped intersection should be perpendicular or nearly perpendicular to each other, and the turning radius at the corners of the intersection should meet the requirements of vehicle turning. If it is found that the shape of a certain intersection is irregular or the road connection angle is abnormal, it is necessary to further check its connectivity. In addition, the electronic device can analyze the traffic flow information at the intersection, including the driving directions of the lanes in different directions, the setting of turning lanes, etc., and match it with the setting of traffic lights. Ensure that the control of the traffic lights can reasonably guide the passage of vehicles at the intersection and avoid traffic conflicts. For example, at an intersection with a left-turn traffic light, check whether the setting of the left-turn lane matches the duration and cycle of the traffic light to ensure that left-turn vehicles can pass through the intersection safely. Finally, the electronic device can detect the positional relationship between the crosswalk and the lane, ensure that there is a reasonable safety distance between the crosswalk and the lane, and the setting of the crosswalk does not affect the normal passage of vehicles. At the same time, check whether vehicles have sufficient visibility to observe pedestrians at the crosswalk to ensure the safety of pedestrians. For example, at an intersection near a school, the crosswalk should be set at a position where vehicles can easily decelerate and stop, and the isolation facilities between the lane and the crosswalk should be perfect.

[0133] For the integrity detection of the road network, the electronic device can construct a topological structure model based on the road network data of the simulation map and check whether the connection relationships of the road nodes and edges are complete. Ensure that there are no isolated road segments or unconnected nodes. For example, in the road network of a city, there should be no dead-end road that is not connected to any other road. Through topological analysis, potential vulnerabilities and incompleteness in the road network can be discovered. In addition, the electronic device can compare the simulation map with the actual geographical area to evaluate the coverage of the road network in this area. Check whether there are certain areas not covered by the road network or areas with low coverage. For some newly developed areas or remote areas, special attention should be paid to the extension and coverage of the road network to ensure that the map can accurately reflect the actual traffic accessibility. Finally, the electronic device checks the connection of the road network with other transportation facilities, such as railways, airports, ports, etc. Ensure that there are convenient and efficient connection channels between these important transportation hubs and the urban road network to enable seamless transfer between different transportation modes. For example, the connection between the airport and the urban expressway or arterial road should be smooth, with clear indication signs to facilitate passengers to reach the airport quickly.

[0134] Step a2, according to the retrieval results, perform topological optimization on the simulation map to obtain the first optimized map.

[0135] Specifically, after obtaining the retrieval results of lane connectivity, intersection connectivity, and road network integrity, the electronic device can analyze the problems therein in detail. For lane connectivity problems, clarify the specific locations and reasons for lane line interruptions, incorrect connections, or unreasonable lane change areas; for intersection connectivity problems, determine the locations of problems such as abnormal intersection geometries, mismatches between traffic flows and signal lights, and improper crosswalk settings; in terms of road network integrity, find the locations of isolated road segments, unconnected nodes, and poor connections with other transportation facilities. For example, it is found that the lane of a certain road suddenly interrupts in the middle, or the turning lane setting at an intersection does not match the duration of the traffic signal light, or there is a lack of connection between some areas of the urban road network and the surrounding roads.

[0136] For lane connectivity problems, the electronic device can take corresponding adjustment measures. If the lane line is interrupted, extend or repair the lane line reasonably according to the surrounding road conditions and traffic demands to ensure the continuity of the lane. In the lane change area, optimize the lane line setting to make it more in line with the actual needs of vehicle driving, such as adjusting the length and position of the dotted line to clarify the lane change range. For multi-lane roads, ensure the reasonable connection between each lane to avoid situations where vehicles cannot change lanes normally.

[0137] Regarding the problem of intersection connectivity, the electronic device can optimize the topological structure of the intersection. If the intersection geometry is irregular, adjust the connection angles and lengths of the roads to conform to standard intersection types (such as cross-shaped, T-shaped, etc.). For the situation where the traffic flow does not match the traffic lights, re-plan the traffic flow, adjust the settings of turning lanes, and optimize the duration and cycle of traffic lights according to the actual traffic volume. At the same time, reasonably set the location and width of crosswalks to ensure that pedestrians can cross the street safely without affecting vehicle traffic. For example, transform an irregular intersection into a standard cross-shaped intersection, re-set the lanes and turning lanes in all directions, and adjust the signal timing plan to improve the traffic efficiency and safety of the intersection.

[0138] Regarding the optimization of the road network topology, for the problem of road network integrity, the electronic device can repair isolated road segments, reasonably connect unconnected nodes with surrounding roads, and improve the topological structure of the road network. For the situation where the connection with other traffic facilities is not smooth, plan new connecting roads or optimize existing connection channels to ensure efficient transfer between different traffic modes. For example, in a newly built area on the edge of the city, connect the originally isolated road with the urban arterial road, or build a dedicated expressway between the airport and the urban road to improve traffic convenience.

[0139] Optionally, after completing the topological structure adjustment, the electronic device can re-detect the optimized map to verify whether the lane connectivity, intersection connectivity, and road network integrity have been improved. By simulating the driving of vehicles on the map, check whether the lanes are unobstructed, whether the intersections can pass normally, and whether the road network is intact. At the same time, compare the retrieval results before and after optimization to evaluate the optimization effect. If problems are still found, continue to adjust and optimize until the requirements are met. In addition, the actual traffic data or user feedback can be collected to further verify the effect of the optimized map in actual applications and ensure its accuracy and practicality. For example, through actual vehicle tests, observe the driving conditions of vehicles on the optimized road network, whether they can drive along the expected path, whether there are new problems or hidden dangers, and further optimize and improve the map according to the test results.

[0140] Step S2032, perform geometric optimization on the first optimized map to obtain the second optimized map.

[0141] Specifically, the road curves in the first optimized map may be non-smooth due to data collection or processing reasons. The electronic device can use curve fitting algorithms, such as Bezier curve fitting or spline curve fitting, to smooth the road curves. By adjusting the control points and parameters of the curve, the road curves can be made more natural and smooth, conforming to the actual road shape. For the curves of highways, after fitting and optimization, the curves are smoother and can more accurately reflect the trajectory of vehicles when driving on the curves.

[0142] The electronic device can also correct the length and width of the roads in the first optimized map according to actual measurement data or more accurate geographical information sources. For some roads with length deviations, by comparing with high-precision satellite images or field measurement data, adjust their lengths on the map. At the same time, according to the actual number of lanes and design standards of the roads, accurately set the width of the roads, so that the geometric dimensions of the roads on the map are closer to the real situation. For example, the main roads in the city may actually have a width of six lanes in both directions, but the width is inaccurately marked in the first optimized map and is corrected through geometric optimization. For roads with slope changes, the electronic device can use terrain data and elevation models to calculate the slopes of the roads and mark them in a suitable way on the first optimized map. The electronic device can use methods such as color gradients or contour lines to visually display the changes in road slopes, helping users better understand the terrain characteristics of the roads. For example, on the map of mountain roads, clearly mark the sections with large slopes to provide reference for drivers. To precisely process the geometric shapes of intersections, the electronic device can adjust the shapes of intersections according to actual intersection designs and field measurement data to make them more in line with the actual situation. For irregularly shaped intersections, by refining the information of nodes and edges, accurately depict their outlines. For example, for some complex multi-road intersections, after geometric optimization, they can more accurately show the intersection angles and ranges of each road.

[0143] In addition, the electronic device can optimize the turning radii of intersections according to the driving characteristics of vehicles and traffic regulations. Ensure that the settings of the turning radii can meet the turning needs of different types of vehicles and avoid situations where vehicles have difficulty turning or are unsafe. For intersections with a large number of large vehicles, appropriately increase the turning radii and accurately mark them on the map to provide more accurate navigation information for drivers. Then, the electronic device can adjust the lane layouts of intersections according to the actual traffic flow and traffic organization plans. Reasonably set the number and positions of straight lanes, turning lanes, and U-turn lanes to improve the traffic efficiency of intersections.

[0144] Finally, the electronic device can perform position calibration on geographical features in the first optimized map, such as buildings, bridges, rivers, etc. The electronic device uses high-precision geolocation data to ensure that the positions of these features on the map are accurate. By comparing with satellite images or field measurement data, the coordinates of geographical features are adjusted to match the actual positions. For example, a building with a position deviation on the map is moved to the correct position to improve the accuracy of the map. Then, the electronic device adjusts the relative positional relationship between geographical features and roads to make it more in line with the actual situation. For buildings adjacent to roads, ensure that the distance and relative position between them and the roads are accurate. At the same time, for bridges or rivers spanning roads, accurately represent the crossing method and positional relationship with the roads.

[0145] Step S2033, perform semantic optimization on the second optimized map to generate the target optimized map.

[0146] Specifically, the electronic device can perform more detailed semantic annotation on geographical features in the second optimized map. For example, for roads, annotate their types (such as highways, urban streets), and also add functional information about the roads, such as whether it is a bus lane, a tidal lane, etc. For buildings, annotate their uses. For example, label a certain road as "Urban arterial road (bus lane, 7:00 - 9:00, 17:00 - 19:00)". For traffic facilities, improve their semantic information. For traffic lights, annotate their control methods, such as timed control, induction control, and the signal timing plan for different time periods. For intersections, annotate their complexity, such as simple intersections, multi-road intersection complex intersections, and whether there are special structures such as roundabouts and overpasses.

[0147] Then, the electronic device optimizes the semantic association between geographical features, clarifying the connection relationships between roads and surrounding buildings, bus stops, parking lots, etc. For example, clearly show on the map which bus stops are adjacent to a certain road, how to enter a nearby parking lot from the road, and what the main buildings are along the road. For the relationship between rivers and bridges, annotate the position and name where the bridge crosses the river, and the connectivity between the two banks of the river and the surrounding roads.

[0148] Finally, the electronic device integrates the semantic information of traffic rules to closely associate it with geographical features. Clearly label on the map traffic rules such as speed limits, traffic restrictions, one-way streets, etc., and the changes of these rules at different time periods. For example, label on a certain road "Speed limit 60 km / h, no left turn allowed from 7:00 - 9:00 and 17:00 - 19:00 on weekdays", and combine these rules with the geometric shape of the road and intersection information to provide users with comprehensive traffic guidance.

[0149] Step S204, perform virtual simulation of the vehicle's memorized driving route based on the target optimized map.

[0150] Specifically, the above-mentioned step S204 may include the following steps:

[0151] Step S2041: Perform a basic static check on the target optimized map to obtain a detection score corresponding to the target optimized map.

[0152] Specifically, the above-mentioned step S2041 may include the following steps:

[0153] Step b1: Perform an element static detection on the target optimized map to obtain an element score.

[0154] Among them, the element score includes at least one of a basic element score, a full-scale navigation score, and a semantic information score.

[0155] Specifically, the electronic device can comprehensively detect the road elements in the target optimized map to check whether the road is missing, whether the lane is interrupted, and whether the lane crosses the road boundary. In the topological structure of the target optimized map, the electronic device detects nodes (such as intersections, road junctions, etc.) and edges (road segments connecting nodes). For example, it checks for abnormal intersection shapes and intersection edge crossings.

[0156] Then, the electronic device detects the V2 navigation or full-scale navigation path in the target optimized map and evaluates its accuracy. Specifically, the electronic device can check for V2 deviation from the trajectory, wrinkles in the full-scale navigation or V2, V2 Turn Type errors, large differences between the full-scale navigation and V2, excessive missing linkIDs in the full-scale navigation, failed association between V2 and the trajectory, whether the V2 seam segment navigation card is correct, no navigation route check, loop route check, V2 check, etc., and will not list them one by one here.

[0157] Finally, the electronic device detects the semantic annotation information in the target optimized map and evaluates its accuracy. Optionally, the electronic device can check for overlapping adjacent idle intervals, missing road width information, overly large adjacent idle intervals, and scene division checks, etc., and will not list them one by one here.

[0158] As shown in Table 1, it is an evaluation form for the electronic device to perform element score, full-scale navigation score, and semantic information score on the target optimized map.

[0159] Table 1 Evaluation form for element score, full-scale navigation score, and semantic information score

[0160]

[0161]

[0162] Step b2: Perform a legality check on the target optimized map to obtain a legality score.

[0163] Among them, the legality score includes at least one of the boundary legality score, the road legality score, the lane legality score, and the semantic legality score.

[0164] Specifically, the electronic device can perform boundary legality scoring, road legality scoring, lane legality scoring, and semantic legality scoring on the target optimized map, so as to obtain the legality score. Exemplarily, as shown in Table 2, it is an evaluation form for the electronic device to perform boundary legality scoring, road legality scoring, lane legality scoring, and semantic legality scoring on the target optimized map.

[0165] Table 2 Evaluation form for boundary legality scoring, road legality scoring, lane legality scoring, and semantic legality scoring

[0166]

[0167]

[0168]

[0169] Step b3, based on the element score and the legality score, obtain the detection score corresponding to the target optimized map.

[0170] Specifically, the electronic device can perform weighted summation again on the calculated element score and legality score according to the pre-determined weight, and finally obtain the detection score corresponding to the target optimized map. This detection score comprehensively reflects the overall performance of the map in terms of element quality and legality. If the detection score is relatively high, it indicates that the map has good quality in all aspects and can meet the requirements of practical applications; if the detection score is relatively low, it is necessary to analyze the problems existing in the map according to the specific element score and legality score, and make targeted improvements and optimizations.

[0171] Step S2042, if the detection score is higher than or equal to the preset score threshold, then perform virtual simulation of the vehicle's memorized driving route based on the target optimized map to generate a virtual simulation result.

[0172] Specifically, the electronic device accurately imports the target optimized map into the virtual simulation platform. During the import process, ensure that the geometric information (such as road shape, lane layout, intersection structure), semantic information (such as traffic rules, point-of-interest annotations), and topological information (road connection relationship, node attributes) of the map can be presented completely and correctly. According to the characteristics and requirements of the simulation platform, perform adaptation processing on the map data. For example, adjust the coordinate system and resolution of the map to match the requirements of the simulation platform, and ensure that the driving simulation of the vehicle on the map can proceed smoothly.

[0173] Then, build a vehicle model in the virtual simulation platform, and set environmental parameters for the virtual simulation to create a driving scenario close to the real world. Set weather conditions, such as sunny, rainy, snowy, foggy days, etc. Different weather conditions will affect the driving performance of the vehicle and the driver's vision. Set lighting conditions, including day, night, and the intensity and direction of light at different times, which have an important impact on the performance of sensors and the vehicle's visual perception. In addition, parameters such as traffic flow and pedestrian density can also be set to simulate different traffic conditions. Then, in the set virtual environment, according to the preset memory driving route, control the vehicle model to perform memory driving simulation on the target optimization map. The vehicle travels according to the road network and traffic rules on the map, and its driving trajectory is jointly affected by the vehicle dynamics model and environmental factors. When encountering intersections, the vehicle performs corresponding operations according to traffic lights and turning rules; on different road types, the vehicle adjusts its driving speed according to the speed limit requirements.

[0174] During the virtual simulation process, the electronic device can record information such as the driving trajectory, speed, acceleration, sensor data, and decision-making process of the vehicle. Analyze these data to evaluate the performance and behavior of the vehicle. Analyze the driving stability, safety of the vehicle under different road conditions and environmental conditions, as well as the decision-making accuracy and reliability of the autonomous driving system. Through data statistics and visualization techniques, display the driving situation and performance indicators of the vehicle to provide a basis for subsequent improvement and optimization. Then, evaluate the performance of the vehicle according to the recorded and analyzed data. Compare the simulation results with the preset performance indicators to determine whether the vehicle meets the expected performance requirements. If performance problems are found in the vehicle, such as insufficient driving stability, decision-making errors, etc., optimize and adjust the vehicle model, parameters, or decision-making algorithm. Through repeated simulation and optimization, gradually improve the performance and reliability of the vehicle.

[0175] In an alternative embodiment of the present application, the electronic device may associate the virtual simulation results generated by performing virtual simulation of the vehicle's memorized driving route based on the target optimized map with the real test results obtained by testing the target vehicle on the actual driving route. Thus, the virtual simulation results can be compared with the real test results, and then, based on the comparison results, the accuracy of the virtual simulation results and the potential problems existing in the target optimized map can be determined. If all the error indicators in the quantitative comparison are within the acceptable range, and the vehicle behavior and event simulation in the qualitative comparison highly conform to the real situation, then it can be considered that the virtual simulation results have high accuracy. On the contrary, if there are large trajectory deviations, inaccurate simulation of dynamic parameters, or incorrect simulation of behavior events, the virtual simulation model needs to be optimized. This may involve adjusting the simulation algorithm, correcting the map input data, or improving the vehicle model parameter settings, etc. In addition, through the comparison, potential problems existing in the target optimized map can also be discovered. If the driving trajectory of the vehicle in the virtual simulation significantly deviates from the real trajectory in some sections, and the problem of the virtual simulation algorithm is excluded, then it is very likely that there are errors in the road shape, position, or topological structure of this section in the map. If the vehicle can pass through a certain intersection smoothly in the real test, but there are path planning errors or abnormal driving in the virtual simulation, this may imply that the semantic information (such as turning rules, lane settings) of the intersection in the map is inaccurate. By analyzing the comparison results in detail, the problem areas and problem types in the map can be accurately located, providing a strong basis for further optimizing the map.

[0176] Step S2043, evaluate the virtual simulation results based on a preset evaluation index.

[0177] Among them, the preset evaluation index includes at least one of a safety index, a positioning accuracy rate index, and a path planning rationality index.

[0178] Specifically, for the evaluation of safety indicators, in virtual simulation, the electronic device can closely monitor the relative positions and motion states of the vehicle model and surrounding environmental elements (such as other vehicles, pedestrians, obstacles), and calculate the probability of collision risk. By analyzing the driving trajectory of the vehicle, it is judged whether there is a potential collision possibility. If the predicted driving trajectories of the vehicle and other vehicles intersect at a certain moment and the time interval is less than the safety threshold, it indicates a high collision risk, which will have a negative impact on the safety indicator score. Using methods such as Monte Carlo simulation, the driving conditions of the vehicle in the same scenario are simulated multiple times, and the frequency of collisions is statistically counted to more accurately evaluate the collision risk. In addition, the electronic device can check the degree of compliance of the vehicle model with traffic rules during the entire virtual simulation process. This includes whether it correctly responds to traffic lights, such as being able to stop in time when the red light is on and start driving according to regulations when the green light is on; whether it complies with the speed limit regulations and maintains a legal driving speed on different road types; whether it correctly uses the turn signal and turns on the corresponding turn signal in advance when turning or changing lanes. Quantify and record violations. For example, running a red light will deduct a certain number of points, and speeding will deduct different scores according to the speeding ratio. The cumulative violation deduction situation is directly reflected in the safety indicator score. Finally, the electronic device can set specific emergency scenarios, such as a suddenly appearing pedestrian or a sudden brake of the vehicle in front, to test the emergency braking and avoidance capabilities of the vehicle model. Evaluate whether the vehicle model can make the correct braking reaction in the shortest time, so that the vehicle can stop safely or take reasonable avoidance measures to avoid collisions. By measuring parameters such as the braking distance of the vehicle, the rationality of the avoidance path, and the stability during the avoidance process, the emergency response ability of the vehicle is scored. If the vehicle can complete the emergency braking or avoidance operation quickly and smoothly, the safety indicator score will increase accordingly.

[0179] For the evaluation of the positioning accuracy index, in a virtual simulation environment, the electronic device can use the high-precision map coordinates as a reference to obtain the simulated positioning information of the vehicle model in real time. By calculating the difference between the simulated positioning coordinates of the vehicle model and the real map coordinates, the positioning deviation is obtained. For different types of positioning technologies (such as GPS, inertial navigation, visual positioning, etc.), their positioning deviations are calculated separately. In a complex urban environment, GPS positioning may be affected by factors such as high-rise building occlusion, resulting in a large positioning deviation. By statistically analyzing the positioning deviation data over a period of time, indicators such as the average positioning deviation and the maximum positioning deviation are calculated to measure the positioning accuracy. In addition to the magnitude of the positioning deviation, the stability of positioning is also crucial. The electronic device can observe the fluctuation of the positioning data of the vehicle model during driving. If the positioning data frequently shows large jumps, it indicates that the positioning is unstable. The electronic device uses statistical methods to calculate the standard deviation of the positioning data. The smaller the standard deviation, the more stable the positioning. For autonomous vehicles, stable positioning is the basis for accurate driving. Poor positioning stability may lead to abnormal driving trajectories of the vehicle, affecting driving safety. Therefore, the evaluation result of positioning stability will significantly affect the score of the positioning accuracy index. In virtual simulation, the effects of different positioning technology fusion algorithms are evaluated. The electronic device can compare the changes in positioning deviation and stability before and after fusion to determine whether the fusion algorithm effectively improves the positioning accuracy. If the positioning deviation after fusion is significantly reduced and the positioning stability is enhanced, it indicates that the fusion algorithm has achieved good results, and the score of the positioning accuracy index will increase accordingly. By simulating different scenarios and working conditions, the performance of positioning technology fusion in various situations is comprehensively evaluated.

[0180] For the rationality index of path planning, the electronic device can analyze the actual driving path length of the vehicle model in virtual simulation and compare it with the theoretically optimal path length. The optimal path can be calculated according to map information and driving goals (such as reaching the destination) through classic path planning algorithms such as Dijkstra algorithm and A* algorithm. If the actual driving path length is close to the optimal path length, it indicates that the path planning is relatively reasonable and the score is high; if the actual path length is significantly greater than the optimal path length, there may be an unreasonable path planning situation, such as detouring, and the score of the path planning rationality index will decrease at this time. In addition, the electronic device records the actual driving time of the vehicle model from the starting point to the end point, and evaluates the driving efficiency in combination with factors such as road speed limits and traffic flow. In the case of large traffic flow, reasonable path planning should be able to avoid congested sections and select a route with a shorter driving time. By comparing with the driving times of other possible paths under the same traffic conditions, it is judged whether the current path planning is efficient. If the vehicle can reach the destination in a shorter time, the score of the path planning rationality index will increase accordingly.

[0181] Step S2044, if the detection score is lower than the preset score threshold, continue to obtain other original driving data, and re-optimize the target optimized map based on the other original driving data until the detection score of the target optimized map is higher than the preset score threshold.

[0182] Among them, the other original driving data is the original driving data obtained by re-collection.

[0183] Specifically, if the detection score is lower than the preset score threshold, the electronic device can continue to obtain other original driving data, and re-optimize the target optimized map based on the other original driving data until the detection score of the target optimized map is higher than the preset score threshold.

[0184] The optimization process can be referred to the above, and will not be elaborated here.

[0185] The memory driving simulation test method provided by the embodiments of the present application preprocesses the original driving data to obtain the preprocessed original driving data, ensuring the accuracy of the preprocessed original driving data. Then, the preprocessed original map data is recognized, and based on the recognition result, an initial road network topology structure is constructed. Then, the preprocessed navigation path map is recognized to determine the key detail elements in the navigation path map; according to the position information corresponding to each key detail element, the key detail elements are supplemented to the initial road network topology structure to generate a candidate road network topology structure. Thus, the actual layout of the road can be accurately restored, making the generated candidate road network topology structure closer to the reality, providing accurate geographical information for subsequent applications. For example, the positioning accuracy of autonomous driving vehicles is improved. Then, the V2 navigation path data and the driving trajectory data are matched with the candidate road network topology structure to generate a navigation trajectory, integrating the advantages of different data sources. The V2 navigation path provides accurate navigation information, and the driving trajectory reflects the actual driving path. The two are matched with the road network to correct and improve the map, enhancing the accuracy and reliability of the map, and ensuring the accuracy of navigation and path planning. Based on the original map data and the navigation path map, semantic information is added to generate a simulation map, increasing the semantic understanding ability of the map. The semantic information covers road types, traffic rules, etc., providing a decision-making basis for the intelligent transportation system. The assisted driving system can make reasonable decisions based on this, improving driving safety and efficiency.

[0186] The simulation map is identified, and the lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulation map are detected, so that connection problems in the map can be discovered in time. Such as lane disconnection, poor traffic at intersections, or missing road networks. According to the search results, the simulation map is topologically optimized to obtain the first optimized map. After solving these problems through topological optimization, the generated first optimized map can provide more accurate road connection information for the path planning algorithm. Accurate lane and intersection connectivity and a complete road network are essential for traffic simulation. In scenarios such as traffic flow simulation and urban traffic planning, simulation based on optimized maps can more realistically reflect actual traffic conditions. For example, simulating the traffic conditions of vehicles at complex intersections and the degree of traffic congestion on different sections of roads can provide more reliable data support for traffic planners to formulate more effective traffic management strategies and improve urban traffic congestion. In addition, in the process of map development and related technology research and development, early discovery and resolution of map connectivity and integrity issues can avoid a lot of rework and cost increase caused by map errors in the later stage. For example, in the memory driving simulation test, if a map with connectivity problems is used, the test results may be inaccurate, and retesting is required, wasting a lot of time and resources. By optimizing the map, the accuracy and reliability of the test can be improved, and the development and testing costs can be reduced. For example, accurately setting the connectivity of the road and clarifying the turning rules of the vehicle at the intersection provide a reliable infrastructure for the navigation system and the autonomous driving algorithm, making the path planning more reasonable and efficient, and reducing navigation errors and driving conflicts. The geometric shape and position deviation of the map elements can be corrected by geometrically optimizing the first optimized map to obtain the second optimized map. For example, the road shape can be made more in line with the actual curvature, and geographical elements such as buildings and landmarks can be accurately located. This not only improves the visualization effect of the map, allowing users to obtain geographical information more intuitively and accurately when viewing the map, but also enhances the accuracy of the map in measurement, geographical analysis, etc. The second optimized map is semantically optimized, and rich semantic information is added to the second optimized map to generate a target optimized map. The target optimized map that has been optimized multiple times can play an important role in many fields such as intelligent transportation, autonomous driving, geographic information systems, and urban planning. It provides a unified, high-quality map data foundation for different fields, and promotes data sharing and collaborative cooperation among various fields. Reduce costs and improve efficiency: High-quality target optimization maps can reduce resource waste and inefficiency caused by inaccurate maps.

[0187] Then, by performing static detection on the target optimized map, basic element scores, full-navigation scores, semantic information scores, etc. are obtained, which can deeply evaluate the accuracy and integrity of each element of the map. This ensures that the map elements conform to the actual situation and provides accurate data support for vehicle navigation, path planning, etc. The boundary legality scores, road legality scores, lane legality scores, semantic legality scores, etc. obtained from the legality detection ensure that the map follows the established rules and standards from different perspectives. For example, the boundary legality score can detect whether the boundary ID is unique, whether the layer_type is legal, etc.; the road legality score can judge whether the road ID is legal, whether there is a missing boundary or lane, etc. By ensuring the legality of the map, abnormal vehicle driving caused by map data errors is avoided, and the reliability of the map is improved. Based on the element scores and legality scores, a detection score is obtained, realizing a comprehensive evaluation of the map. This multi-dimensional scoring method can more accurately reflect the overall quality of the map. R & D personnel can optimize the map targeted according to the scores, timely discover and solve potential problems, and ensure the accuracy and stability of the map in various application scenarios.

[0188] Then, based on preset evaluation indicators, the virtual simulation results are evaluated. Among them, safety indicators (such as the number of user takeovers per 100 kilometers, the number of emergency brakes and hard brakes), positioning accuracy indicators, path planning rationality indicators, etc. quantify the driving performance of the vehicle in the virtual environment from multiple dimensions. For example, by analyzing the positioning accuracy indicator, it is possible to judge the positioning accuracy of the vehicle during the simulation process. Inaccurate positioning may cause the vehicle to deviate from the route, and this evaluation process can timely discover such problems, providing a basis for optimizing the vehicle positioning system and improving the safety and reliability of the vehicle during actual driving. In addition, comprehensively evaluating the vehicle performance in the virtual simulation link can quickly discover problems existing in the vehicle, such as unreasonable path planning. R & D personnel can optimize the vehicle algorithm and system targeted according to the virtual simulation results, avoiding discovering problems only in the real vehicle test stage, thus saving a large amount of time and cost and accelerating the vehicle R & D cycle. By evaluating the virtual simulation results, it is possible to initially screen out better-performing solutions and reduce the number of unnecessary real vehicle tests. Real vehicle tests require a large amount of manpower, material resources and financial resources, such as the purchase and maintenance of test vehicles, the lease of test sites, etc., while virtual simulation only needs to be carried out in a computer environment, reducing the test cost.

[0189] In addition, if the detection score is lower than the preset score threshold, it indicates that the map has deficiencies. Continuing to obtain other original driving data and optimizing the map again can continuously improve the accuracy and integrity of the map. The newly collected data can supplement missing information and correct error elements. For example, improving the topological structure of the road network, updating the positions of traffic signs, etc., making the map more in line with the actual situation and providing a reliable basis for vehicle driving and navigation. In addition, over time and with changes in the road environment, the map needs to be continuously updated. This optimization mechanism based on new data can enable the target optimized map to closely follow the actual changes and maintain a good quality state. Whether it is the construction of new urban roads, the adjustment of traffic rules, or the need to improve the accuracy of the map itself, it can be met by continuously obtaining new data to optimize the map, ensuring the reliability of the map during long-term use. In addition, an accurate map is the key to vehicle virtual simulation. The map optimized again can provide a more accurate environmental model for virtual simulation, making the behavior of the vehicle in the simulation closer to the real situation. The optimized map is used for virtual simulation, which can more accurately evaluate the performance indicators of the vehicle, helping R & D personnel to more precisely discover problems existing in the vehicle and improve the vehicle system and algorithms targeted. A high-quality map can reduce the number and time of real vehicle tests of the vehicle. Through sufficient testing based on the optimized map in virtual simulation, potential problems can be discovered and solved in advance, reducing invalid and repeated tests caused by map problems in real vehicle tests. Real vehicle tests are costly, including vehicle wear, test site costs, and labor costs, etc. Reducing real vehicle tests can effectively reduce the overall R & D cost.

[0190] In this embodiment, a memory driving simulation test device is further provided. This device is used to implement the above-mentioned embodiments and preferred implementation manners, and those that have been described will not be repeated here. As used hereinafter, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0191] This embodiment provides a memory driving simulation test device, as Figure 3 shown, including:

[0192] An acquisition module 301, configured to acquire original driving data corresponding to the memory driving routes collected by multiple target vehicles; the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data;

[0193] A building module 302, configured to build a simulation map based on the original driving data;

[0194] An optimization module 303, configured to perform optimization processing on the simulation map to generate a target optimized map;

[0195] The simulation module 304 is used to perform virtual simulation of the vehicle's memorized driving route based on the target optimized map.

[0196] In some alternative embodiments, the building module 302 is specifically configured to preprocess the original driving data to obtain the preprocessed original driving data; identify the preprocessed original map data, and construct an initial road network topology based on the identification result; identify the preprocessed navigation path map, and determine the key detail elements in the navigation path map; according to the position information corresponding to each key detail element, supplement the key detail elements to the initial road network topology to generate a candidate road network topology; respectively match the V2 navigation path data and the driving trajectory data with the candidate road network topology to generate a navigation trajectory; based on the original map data and the navigation path map, add semantic information to generate a simulation map.

[0197] In some alternative embodiments, the optimization module 303 is specifically configured to perform topology optimization on the simulation map to obtain a first optimized map; perform geometric optimization on the first optimized map to obtain a second optimized map; perform semantic optimization on the second optimized map to generate a target optimized map.

[0198] In some alternative embodiments, the optimization module 303 is specifically configured to identify the simulation map, and detect the lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulation map; according to the retrieval result, perform topology optimization on the simulation map to obtain a first optimized map.

[0199] In some alternative embodiments, the simulation module 304 is specifically configured to perform basic static inspection on the target optimized map to obtain a detection score corresponding to the target optimized map; if the detection score is higher than or equal to a preset score threshold, perform virtual simulation of the vehicle's memorized driving route based on the target optimized map to generate a virtual simulation result; evaluate the virtual simulation result based on a preset evaluation index; the preset evaluation index includes at least one of a safety index, a positioning accuracy rate index, and a path planning rationality index.

[0200] In some alternative embodiments, the simulation module 304 is further configured to, if the detection score is lower than the preset score threshold, continue to obtain other original driving data, and perform secondary optimization on the target optimized map based on the other original driving data until the detection score of the target optimized map is higher than the preset score threshold; the other original driving data is the original driving data collected again.

[0201] In some alternative embodiments, the simulation module 304 is specifically configured to perform element static detection on the target optimized map to obtain an element score; the element score includes at least one of a basic element score, a full-scale navigation score, and a semantic information score; perform legality detection on the target optimized map to obtain a legality score; the legality score includes at least one of a boundary legality score, a road legality score, a lane legality score, and a semantic legality score; and obtain a detection score corresponding to the target optimized map based on the element score and the legality score.

[0202] The further function descriptions of the above-mentioned various modules and units are the same as those in the corresponding embodiments above, and will not be elaborated here.

[0203] The memory driving simulation test device in this embodiment is presented in the form of functional units. Here, the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and a memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0204] The embodiment of the present invention further provides a computer device having the above-mentioned Figure 3 shown memory driving simulation test device.

[0205] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of a computer device provided by an alternative embodiment of the present invention. As shown in Figure 4 , the computer device includes: one or more processors 10, a memory 20, and an interface for connecting various components, including a high-speed interface and a low-speed interface. Each component communicates with each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed within the computer device, including instructions stored in the memory or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some alternative embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a set of blade servers, or a multi-processor system). Figure 4 In

[0206] The processor 10 may be a central processing unit, a network processor, or a combination thereof. Among them, the processor 10 may further include a hardware chip. The above-mentioned hardware chip may be an application specific integrated circuit, a programmable logic device, or a combination thereof. The above-mentioned programmable logic device may be a complex programmable logic device, a field programmable gate array, a generic array logic, or any combination thereof.

[0207] Among them, the memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiments.

[0208] The memory 20 may include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some alternative embodiments, the memory 20 may optionally include a memory remotely provided with respect to the processor 10, and these remote memories may be connected to the computer device through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0209] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk, or a solid-state drive; the memory 20 may also include a combination of the above-mentioned types of memories.

[0210] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or communication networks.

[0211] The embodiments of the present invention also provide a computer-readable storage medium. The method according to the embodiments of the present invention can be implemented in hardware, firmware, or be implemented as computer code that can be recorded on a storage medium, or be implemented as computer code originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium may be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid-state drive, etc.; further, the storage medium may also include a combination of the above-mentioned types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code, and when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method shown in the above embodiments is implemented.

[0212] A part of the present invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can call or provide the methods and / or technical solutions according to the present invention through the operations of the computer. Those skilled in the art should understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executes the instructions, or the computer compiles the instructions and then executes the corresponding compiled program, or the computer reads and executes the instructions, or the computer reads and installs the instructions and then executes the corresponding installed program. Herein, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to the computer.

[0213] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A memory driving simulation test method, characterized in that The method includes: Obtaining original driving data corresponding to the memory driving routes collected by multiple target vehicles; the original driving data includes at least one of original map data, navigation path map, V2 navigation path data, driving trajectory data, and driving route data; Based on the original driving data, building a simulation map; Performing optimization processing on the simulation map to generate a target optimized map; Performing virtual simulation of the vehicle memory driving route based on the target optimized map.

2. The method according to claim 1, characterized in that The building of the simulation map based on the original driving data includes: Performing preprocessing on the original driving data to obtain preprocessed original driving data; Identifying the preprocessed original map data, and based on the identification result, constructing an initial road network topology structure; Identifying the preprocessed navigation path map to determine the key detail elements in the navigation path map; According to the position information corresponding to each key detail element, supplementing the key detail elements to the initial road network topology structure to generate a candidate road network topology structure; Matching the V2 navigation path data and the driving trajectory data with the candidate road network topology structure respectively to generate a navigation trajectory; Adding semantic information based on the original map data and the navigation path map to generate the simulation map.

3. The method according to claim 1, wherein The performing of optimization processing on the simulation map to generate a target optimized map includes: Performing topology optimization on the simulation map to obtain a first optimized map; Performing geometric optimization on the first optimized map to obtain a second optimized map; Performing semantic optimization on the second optimized map to generate the target optimized map.

4. The method according to claim 3, characterized in that, The performing of topology optimization on the simulation map to obtain a first optimized map includes: Identifying the simulation map to detect the lane connectivity, and / or intersection connectivity, and / or road network integrity of the simulation map; According to the retrieval result, performing topology optimization on the simulation map to obtain the first optimized map.

5. The method according to claim 1, wherein The performing of virtual simulation of the vehicle memory driving route based on the target optimized map includes: Performing basic static inspection on the target optimized map to obtain a detection score corresponding to the target optimized map; If the detection score is higher than or equal to a preset score threshold, then performing virtual simulation of the vehicle memory driving route based on the target optimized map to generate a virtual simulation result; Evaluating the virtual simulation result based on a preset evaluation index; the preset evaluation index includes at least one of a safety index, a positioning accuracy rate index, and a path planning rationality index.

6. The method according to claim 5, wherein The method further includes: If the detection score is lower than the preset score threshold, then continue to obtain other original driving data, and perform re-optimization on the target optimized map based on the other original driving data until the detection score of the target optimized map is higher than the preset score threshold; the other original driving data is the original driving data collected again.

7. The method according to claim 5, wherein The performing of basic static inspection on the target optimized map to obtain a detection score corresponding to the target optimized map includes: Perform element static detection on the target optimized map to obtain an element score; the element score includes at least one of a basic element score, a full-scale navigation score, and a semantic information score; Perform legality detection on the target optimized map to obtain a legality score; the legality score includes at least one of a boundary legality score, a road legality score, a lane legality score, and a semantic legality score; Based on the element score and the legality score, obtain the detection score corresponding to the target optimized map.

8. A memory driving simulation test device, characterized in that, The device includes: An acquisition module, configured to acquire original driving data corresponding to the memory driving routes collected by multiple target vehicles; the original driving data includes at least one of original map data, a navigation path map, V2 navigation path data, driving trajectory data, and driving route data; A construction module, configured to construct a simulation map based on the original driving data; An optimization module, configured to perform optimization processing on the simulation map to generate a target optimized map; A simulation module, configured to perform virtual simulation of the vehicle memory driving route based on the target optimized map.

9. A computer device, characterized in that, Including: A memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to execute the memory driving simulation test method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Computer instructions are stored on the computer-readable storage medium, and the computer instructions are used to cause a computer to execute the memory driving simulation test method according to any one of claims 1 to 7.

Images