Reference data generation apparatus and method for in-vehicle navigation test

Abstract

This invention relates to a reference data generation apparatus and method for vehicle navigation testing. The reference data generation apparatus includes: a navigation engine for performing a first simulated navigation based on the start and end positions of navigation in a simulated map, obtaining path planning data and simulated navigation data for a first simulated vehicle; a differentiator, an interpolation alignment module, and a first delay unit cooperate to provide the necessary data to a simulator, which then performs a second simulated navigation in a simulated scenario based on radio frequency signals, position and attitude vector synthesis data, and a first delay sequence, obtaining first reference data for vehicle navigation testing. This application, by processing the results obtained from the first simulated navigation, can obtain the data used by the simulator in the second simulated navigation. Therefore, the simulator can automatically construct the actual required simulation scenario, obtaining the first reference data for vehicle navigation testing, thus improving the efficiency of vehicle navigation testing.

Description

Technical Field

[0001] This invention relates to the field of vehicle navigation testing, and more specifically to a reference data generation device and method for vehicle navigation testing. Background Technology

[0002] With the development of technologies such as ADAS and autonomous driving, vehicle positioning systems are evolving from traditional GPS positioning towards multi-sensor fusion. Their applications are expanding beyond traditional navigation functions in in-vehicle infotainment systems to include supporting autonomous driving decision-making and control, and providing the vehicle's precise location. Consequently, their application scenarios are no longer simply navigation-related, but increasingly involve interactions with the environment and other vehicles.

[0003] As one of the fundamental technologies for intelligent connected vehicles, in-vehicle positioning systems play a crucial role in areas such as cockpit navigation and autonomous driving. Current technologies for testing in-vehicle navigation often require setting up various scenarios and configuring different sensor parameters. Recreating real-world scenarios demands significant manpower and resources to adapt various scenario data and configured virtual sensor data one by one. However, the recreated scenarios are extremely limited, making the testing process cumbersome and inefficient. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, or at least partially solve the above-mentioned technical problems, this application provides a reference data generation device and method for vehicle navigation testing.

[0005] In a first aspect, this application provides a reference data generation device for vehicle navigation testing, including: a navigation engine, a differentiator, an interpolation alignment module, a first delay unit, and a simulator;

[0006] The navigation engine is used to import map data, construct a simulated map based on the map data, obtain the input navigation start and end positions, perform a first simulated navigation based on the navigation start and end positions in the simulated map, and obtain the path planning data and simulated navigation data of the first simulated vehicle.

[0007] The differentiator is used to differentiate the velocity information and curvature information in the simulated navigation data to obtain acceleration information and angular velocity information.

[0008] The interpolation alignment module is used to interpolate the path planning data and simulated navigation data to obtain interpolated data, match the interpolated data, the acceleration information and the angular velocity information, and perform alignment processing according to a preset period to obtain interpolated navigation data.

[0009] The first delayer is used to delay the interpolated navigation data to obtain a first delay sequence for time synchronization with the second simulated navigation;

[0010] The simulator is used to synthesize the acceleration information and the angular velocity information to obtain position and attitude vector synthesis data; construct a simulated scene based on preset scene parameters, obtain real-time scene data corresponding to the second simulated navigation, generate radio frequency signals from the interpolated navigation data and the real-time scene data, and perform second simulated navigation in the simulated scene based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence, and simulate the signals of the on-board sensors of the first simulated vehicle to obtain first reference data for testing the on-board navigation.

[0011] Optionally, it also includes: a navigation start and end location input module and a map module;

[0012] The navigation start and end position input module is used to receive the input navigation start and end positions and send them to the navigation engine;

[0013] The map module is used to provide the map data and send the map data to the navigation engine.

[0014] Optionally, it may also include: a timing module and a memory;

[0015] The timing module is used to synchronize the navigation engine, the differentiator, the interpolation alignment module, the first delay unit, and the simulator to obtain time synchronization information.

[0016] The memory is used to store the navigation start and end positions, the first delay sequence, the path planning data, the acceleration information, the angular velocity information, and the time synchronization information.

[0017] Optionally, it also includes: a first position offsetter, at least one second delayer, at least one second position offsetter, a first modeling comparator, and a first stitching smoothing processing module;

[0018] The first position offset is used to offset the first delay sequence to obtain the first offset sequence of the first simulated vehicle in the lane change scenario. The first position offset is connected to multiple serial branches composed of a second delay and a second position offset.

[0019] Each of the second delayers is used to time delay the first offset sequence to obtain a second delayed sequence for time synchronization with the third simulated navigation; each of the second position offsetters is used to offset the second delayed sequence to obtain a second offset sequence for at least one second simulated vehicle.

[0020] The first splicing and smoothing processing module is used to splice and smooth the first delayed sequence to obtain a first smoothed delayed sequence.

[0021] The first modeling comparator is used to model vehicles based on the first offset sequence, the first smooth delay sequence and each of the second offset sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results;

[0022] The simulator is also used to perform a third simulated navigation based on the simulated scenario and the conflict determination result, and to simulate the signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles to obtain second reference data for testing the on-board navigation.

[0023] Optionally, it also includes: a first traffic sign modeler, a third delayer, a visual simulator, and a radar simulator;

[0024] The third delayer is used to perform a time delay based on the first delay sequence to obtain a third delay sequence;

[0025] A first traffic sign modeler is used to simulate traffic signs based on the third delay sequence;

[0026] The visual simulator is used to simulate the movement of the first simulated vehicle and each of the second simulated vehicles according to the indication signals of the traffic signs;

[0027] The radar simulator is used to simulate the distance and time between the first simulated vehicle and each of the second simulated vehicles.

[0028] Optionally, it also includes: a first cache;

[0029] The first cache is used to temporarily store the first delay sequence and the interpolated navigation data;

[0030] The first splicing and smoothing processing module is further configured to perform splicing and smoothing processing on the interpolated navigation data to obtain smooth navigation data;

[0031] The memory is also used to store the first smoothing delay sequence and smoothing navigation data.

[0032] Optionally, the number of the first delayers is at least two, and the reference data generation device for vehicle navigation testing further includes: at least two third position offsetters, at least two fourth delayers, a second modeling comparator, and at least two fifth delayers, wherein the first delayers, the third position offsetters, and the fourth delayers constitute at least two branches connected in parallel between the interpolation alignment module and the second modeling comparator;

[0033] Each of the third position offsetters is used to offset each of the first delay sequences to obtain a third offset sequence of multiple third simulated vehicles;

[0034] Each of the fourth delayers is used to time delay each of the third offset sequences to obtain a fourth delayed sequence for time synchronization with the fourth simulated navigation;

[0035] The second modeling comparator is used to model vehicles based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results;

[0036] The simulator is also used to perform a fourth simulated navigation based on the simulated scenario and the conflict determination result, and to simulate the signals of the on-board sensors of multiple third simulated vehicles to obtain third reference data for testing on-board navigation.

[0037] Optionally, it also includes: a second traffic sign modeler, at least two fifth delayers, and a time recovery module;

[0038] The time recovery module is used to synchronize simulated time with real time;

[0039] The fifth delayer is used to delay the initial time to obtain the delay time;

[0040] A second traffic sign modeler is used to simulate traffic signs based on the third delay sequence and the delay time.

[0041] Optionally, it also includes: at least two second caches and at least two second splicing smoothing processing modules;

[0042] Each of the second caches is used to temporarily store each of the first delay sequences and the interpolated navigation data;

[0043] The second splicing and smoothing processing module is used to splice and smooth each of the first delay sequences to obtain a second smooth delay sequence;

[0044] The memory is also used to store the determination result, the second smooth delay sequence, the indication signal of the second traffic sign modeler, and the delay time.

[0045] Secondly, this application provides a method for generating reference data for vehicle navigation testing, including:

[0046] Import map data and construct a simulated map based on the map data to obtain the input navigation start and end positions;

[0047] Based on the navigation start and end positions, a first simulated navigation is performed on the simulated map to obtain the path planning data and simulated navigation data of the first simulated vehicle;

[0048] The velocity and curvature information in the simulated navigation data are differentiated to obtain acceleration and angular velocity information.

[0049] The path planning data and simulated navigation data are interpolated to obtain interpolated data. The interpolated data, the acceleration information and the angular velocity information are matched and aligned according to a preset period to obtain interpolated navigation data.

[0050] The interpolated navigation data is time-delayed to obtain a first delay sequence, which is then synchronized with the second simulated navigation.

[0051] The acceleration information and the angular velocity information are synthesized to obtain position and attitude vector synthesis data; a simulated scene is constructed based on preset scene parameters, and real-time scene data corresponding to the second simulated navigation is obtained. The interpolated navigation data and the real-time scene data are used to generate radio frequency signals. In the simulated scene, the second simulated navigation is performed based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence. The signals of the on-board sensors of the first simulated vehicle are simulated to obtain first reference data for testing on-board navigation.

[0052] Optionally, it also includes:

[0053] The first delay sequence is offset to obtain the first offset sequence of the first simulated vehicle in the lane change scenario;

[0054] The first offset sequence is time-delayed to obtain a second delayed sequence for time synchronization with the third simulated navigation. The second delayed sequence is then offset to obtain a second offset sequence for at least one second simulated vehicle.

[0055] The first delayed sequence is spliced ​​and smoothed to obtain the first smoothed delayed sequence;

[0056] Vehicle modeling is performed based on the first offset sequence, the first smooth delay sequence, and each of the second offset sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results.

[0057] Based on the simulated scenario and the conflict determination result, a third simulated navigation is performed. The signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles are simulated to obtain second reference data for testing of the on-board navigation.

[0058] Optionally, it also includes:

[0059] The first delay sequence is offset to obtain a third offset sequence of multiple third simulated vehicles;

[0060] Each of the third offset sequences is time-delayed to obtain a fourth delayed sequence for time synchronization with the fourth simulated navigation;

[0061] Vehicle modeling is performed based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results.

[0062] A fourth simulated navigation is performed based on the simulated scenario and the conflict determination result. The signals of the on-board sensors of multiple third simulated vehicles are simulated to obtain third reference data for testing of on-board navigation.

[0063] The beneficial effects of this invention are:

[0064] This invention can automatically construct a simulated map based on imported map data, and perform a first simulated navigation based on the input navigation start and end positions in the simulated map to obtain the path planning data and simulated navigation data of the first simulated vehicle. Then, based on the results of the first simulated navigation, through the cooperation of a differentiator, an interpolation alignment module, and a first delay unit, the simple navigation scenario in the first simulated navigation is constructed into data supporting ADAS, autonomous driving, and V2X high-order navigation, preparing data for the simulator to perform complex second simulated navigation. Subsequently, the simulator can perform second simulated navigation based on the acquired real-time scene data, automatically constructing the complex high-order navigation simulation scenario required in reality, simulating the signals of the on-board sensors of the first simulated vehicle, and obtaining first reference data for testing on-board navigation, such as: location information, time information, collision information (radar recognition results), visual information (feature point simulation), and LiDAR point cloud (recognition results), etc. The first reference data can be used for testing positioning-related algorithms or as a data source for hardware-in-the-loop testing, without consuming a lot of manpower and resources to build scenarios and configure parameters, thus improving the efficiency of on-board navigation testing. Furthermore, the simulated navigation data is delayed by the first delayer to achieve time synchronization and avoid time asynchrony during simulator simulation. Attached Figure Description

[0065] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 A schematic diagram of a reference data generation device for vehicle navigation testing provided in an embodiment of this application;

[0068] Figure 2 A schematic diagram of the structure of another reference data generation device for vehicle navigation testing provided in an embodiment of this application;

[0069] Figure 3 A schematic diagram of the structure of another reference data generation device for vehicle navigation testing provided in an embodiment of this application;

[0070] Figure 4 A flowchart illustrating a method for generating reference data for vehicle navigation testing, provided in an embodiment of this application;

[0071] Figure 5 A flowchart illustrating a method for generating reference data for vehicle navigation testing in a practical application, as provided in this application embodiment;

[0072] Figure 6 A schematic diagram illustrating the principle of a reference data generation method for vehicle navigation testing provided in an embodiment of this application;

[0073] Figure 7 A schematic diagram illustrating the principle of another method for generating reference data for vehicle navigation testing provided in this application embodiment;

[0074] Figure 8 A schematic diagram illustrating the principle of another method for generating reference data for vehicle navigation testing provided in this application embodiment;

[0075] Figure 9 A schematic diagram illustrating the principle of another method for generating reference data for vehicle navigation testing provided in this application embodiment;

[0076] Figure 10 A schematic diagram illustrating the principle of another method for generating reference data for vehicle navigation testing provided in this application embodiment;

[0077] Figure 11 A schematic diagram illustrating the principle of another method for generating reference data for vehicle navigation testing provided in this application embodiment. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0079] Traditional simulation testing, when generating large-scale scenarios, often provides interconnected data sources due to their spatiotemporal differences. However, these sources lack actual geographical significance, making it difficult to simulate real-world scenarios, integrate directly with maps, and intuitively access HMI effects information of the device under test. Furthermore, equipment manufacturers typically integrate industry and national standard scenarios, resulting in low user personalization and significant adaptation difficulties. On the other hand, existing solutions follow a "scenario construction - scenario usage" process. Therefore, this application provides a reference data generation device and method for vehicle navigation testing. Through three steps—"virtual navigation usage - complex scenario construction - advanced navigation scenario application"—this solution combines the advantages of both simulation testing and simulator testing, while simplifying complex system adaptation and synchronization processes. Specifically, compared to simulation methods, this invention automatically generates scenario data without the need for scenario construction, achieving a high scenario generation rate. It also eliminates the need for matching and synchronizing various virtual sensors, simplifying the generation process. Furthermore, this application constructs a test data source that supports advanced navigation application scenarios such as autonomous driving, ADAS, and V2X by enabling various simulators to work together through network processing of simple simulated navigation data via delayers, offsetters, etc., without the need for complex system matching. This data source contains radar information, visual recognition information, real-time information, and satellite positioning radio frequency information.

[0080] This application provides a reference data generation device for vehicle navigation testing, such as... Figure 1 As shown, it includes: a navigation engine 11, a differentiator 12, an interpolation alignment module 13, a first delayer 14, and a simulator 15;

[0081] The navigation engine 11 is used to import map data, construct a simulated map based on the map data, obtain the input navigation start and end positions, perform first simulated navigation based on the navigation start and end positions in the simulated map, and obtain the path planning data and simulated navigation data of the first simulated vehicle; the navigation start and end positions include: navigation start position and navigation end position;

[0082] In this embodiment, the navigation engine 11 is mainly used to perform path planning from the navigation start position to the navigation end position (destination), obtain path planning data, and perform a first simulated navigation based on the path planning data. That is, according to the path planning data, it automatically simulates the process of navigating from the navigation start position to the navigation end position in chronological order, and obtains simulated navigation data. Since the first simulated navigation is an ordered and complete process from the navigation start position to the navigation end position, the chronological order and spatial arrangement order of the generated simulated navigation data are determined. That is, the first time point corresponds to the first position, the second time point corresponds to the second position, and so on. The time interval between the first time point and the second time point is determined, the spatial arrangement order of the first position and the second position is determined, and so on. This ensures that no additional matching or synchronization work is required when generating complex scene data, which greatly improves the testing efficiency and scene generation efficiency.

[0083] The navigation engine 11 can import map data in an orderly manner based on the scene setting parameters of the simulated scene, construct a simulated map based on the map data, perform the first simulated navigation on the simulated map, obtain simulated navigation data, and input the simulated navigation data to subsequent modules and devices for scene generation. Map providers already have mature map data that can be used directly. Since this solution generates the simulated scene and the first reference data based on map data, as long as the positions are "legal" during the simulated navigation process, it can complete the simulated navigation automatically. If the planned path of the navigation engine 11 is inconsistent with that of the virtual algorithm during simulated navigation, the difference can be recorded at the location to eliminate the influence of data differences that may lead to misjudgment.

[0084] The navigation engine 11 also includes a coordinate transformation module, which is used to perform coordinate transformation between the vehicle speed and IMU coordinate systems and the simulated position coordinate system to ensure that the data matches correctly.

[0085] The differentiator 12 is used to differentiate the velocity information and curvature information in the simulated navigation data to obtain acceleration information and angular velocity information. In other words, the differentiator 12 can restore the velocity signal in the preset scene parameters to the acceleration signal and the curvature signal to the angular velocity signal to obtain the underlying sensor signal, which is used to simulate the output information of the vehicle sensor and thus support the full-link simulation test.

[0086] The interpolation alignment module 13 is used to perform interpolation processing on the path planning data and simulated navigation data to obtain interpolated data, match the interpolated data, the acceleration information and the angular velocity information, and perform alignment processing according to a preset period to obtain interpolated navigation data.

[0087] In one embodiment of this application, the interpolation alignment module 13 can perform linear interpolation in the straight line area of ​​the path planning data and interpolation according to curvature in the curve area, and interpolate to fill in the missing parts of the simulated navigation data caused by tunnels, viaducts, ramps, turns, acceleration and deceleration and other situations, so as to ensure data continuity and consistency.

[0088] The interpolation alignment module 13 can fine-tune the height and roll in the map data based on the relative height and road tilt angle in the simulated navigation data. Furthermore, it is necessary to ensure clock alignment within the interpolation period to avoid misalignment between position, acceleration, and angular velocity information in the interpolated data.

[0089] The first delayer 14 is used to delay the interpolated navigation data to obtain a first delay sequence for time synchronization with the second simulated navigation;

[0090] The first delay unit 14 is used to establish the first delay sequence. Taking the start time of the simulated navigation, i.e., the time when the first set of data is generated, as the starting point, the interpolated navigation data is time-delayed to obtain the first delay sequence. The first delay unit 14 is used to synchronize the first simulated navigation with the second simulated navigation. It is the first element (i.e., the "time" element) for realizing the synchronized spatiotemporal virtual mapping of this application, and a prerequisite for accessing real-time data such as road conditions, ephemeris, atmospheric parameters, RTK corrections, and other vehicle heatmaps to the simulator 15. By synchronizing the second simulated navigation with time, and using it in conjunction with the simulator, ephemeris data can be obtained from the satellite ground station and ionospheric parameters from the operator during testing, generating realistic positioning scene data. Simultaneously, after time synchronization, real-time road condition information and online maps can be directly obtained through the mobile network, ensuring the integrity of the data acquisition process and the consistency between the data link and the real situation, thus realizing the construction of a realistic scene. Finally, due to the time synchronization and the realism of the scene, various rendering data and HMI data integrated into the vehicle can be directly called during navigation testing, and system interaction tests such as functional testing and evaluation can be directly performed. The simulator 15 is used to synthesize the acceleration information and the angular velocity information to obtain position and attitude vector synthesis data; construct a simulated scene based on preset scene parameters, obtain real-time scene data corresponding to the second simulated navigation, generate radio frequency signals from the interpolated navigation data and the real-time scene data, and perform second simulated navigation in the simulated scene based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence, and simulate the signals of the on-board sensors of the first simulated vehicle to obtain first reference data for testing on-board navigation.

[0091] The simulator 15 is mainly used to vector synthesize characteristic parameters such as temperature drift and zero-bias curve of the inertial unit with the acceleration information and the angular velocity information to obtain position and attitude vector synthesis data. Simultaneously, it acquires real-time scene data corresponding to the second simulated navigation, such as ionospheric data, ephemeris of the current location, road conditions, and heat map data. It then generates radio frequency signals from the interpolated navigation data and the real-time scene data. Based on preset scene parameters, it constructs simulated scenes. For example, if the preset scene parameters are for highways or urban expressways, a simulated scene of an open sky can be constructed; if the preset scene parameters are for commercial districts or urban areas, a simulated scene of an urban canyon can be constructed.

[0092] Radio frequency signals are used to simulate received satellite navigation wireless signals during the second simulated navigation, supporting the acquisition of positioning data for the second simulated navigation. Position and attitude vector synthesis data is used to simulate the vehicle's operating state during the second simulated navigation. The first delay sequence is used to support time synchronization for the second simulated navigation: on the one hand, it supports the time synchronization of the simulator, obtains real-time ionospheric parameters, etc., to ensure the authenticity and integrity of the space satellite navigation signal in the second simulation, and can achieve full-link scenario coverage from the antenna to the positioning module to the navigation application of the real vehicle to support systematic testing; on the other hand, it completes the time synchronization of map data, acceleration information, angular velocity information, etc.

[0093] In the constructed simulated scenario, simulator 15 utilizes the previously generated radio frequency signals, position and attitude vector synthesis data, and the first delay sequence to perform a second simulated navigation. It simulates signals detected by the on-board sensors of the first simulated vehicle in a complex navigation scenario, such as: position information, time information, collision information (radar recognition results), visual information (feature point simulation), and lidar point cloud (recognition results), to obtain first reference data. This first reference data can be used as standard data and the original signal for testing in on-board navigation tests. Thus, when conducting on-board navigation tests, the first reference data can be used as the original signal for testing the on-board navigation test process, thereby obtaining on-board navigation test data. The on-board navigation test data is compared with the standard data (first reference data) to determine the deviation of the on-board navigation test data relative to the standard data (first reference data), thereby obtaining the on-board navigation test result.

[0094] In addition, Simulator 15 also supports the import of inertial unit installation location and vehicle vibration parameters to complete the inertial data restoration, ensuring that it is as close as possible to the real vehicle state.

[0095] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a navigation start and end position input module 16 and a map module 17;

[0096] The navigation start and end position input module 16 is used to receive the input navigation start and end positions and send them to the navigation engine. It can be directly input through the computer call interface. For the test vehicle terminal with built-in navigation function and the navigation function has been implemented, the navigation start and end positions can be directly entered using the human machine interface (HMI) part of the test vehicle terminal. It can also notify the navigation engine 11 to call map data, complete route planning and simulate navigation. At this time, it is only necessary to connect the data output by the navigation engine 11 to the subsequent module for processing.

[0097] Map module 17 is used to provide the map data and send it to the navigation engine. This map module primarily provides the most basic data source for scene construction. The map data includes existing map data collected by map providers, as well as dedicated maps or crowdsourced maps collected by qualified surveying and mapping parties. Regardless of the type, the map information must include data such as road connectivity, road type, boundary conditions, traffic signs, latitude and longitude, relative curvature and height, and road direction to ensure the completeness of the subsequently constructed scene data.

[0098] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a timing module 18 and a memory 19;

[0099] The timing module 18 is used to synchronize the navigation engine 11, the differentiator 12, the interpolation alignment module 13, the first delay unit 14 and the simulator 15 to ensure that the time systems of each module are consistent and to obtain time synchronization information.

[0100] The memory 19 is used to store the navigation start and end positions, the first delay sequence, the path planning data, the acceleration information, the angular velocity information, and the time synchronization information.

[0101] In this embodiment of the application, the memory 19 is also used to store the original scene data of the generated simulation scene, which can be used as a simulation data source for the simulator to facilitate in-loop testing. The generated original scene data can also be used directly for algorithm testing. Since the existing simulator already has the ability to smooth and generate standard data, the solution and device do not need to smooth and generate message data separately.

[0102] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a channel simulator 20;

[0103] The channel simulator 20 is used to simulate and test the propagation performance of the Global Navigation Satellite System (GNSS) in complex environments. It can be left unselected for ordinary testing and can be replaced by real-time scene data (such as ionospheric parameters) downloaded from the simulator 15 and the simulated scene generated by the simulator 15.

[0104] Figure 1 In this system, the device under test (DUT) is both the source and user of information for scene generation. On one hand, it can directly use the DUT's HMI to input navigation start and end positions and simulate navigation. On the other hand, it can provide reference data (such as first reference data) for each simulated scene to the DUT to complete the test. Compared with traditional simulation methods, this system can perform accuracy testing and also evaluate the actual effect using the DUT's terminal HMI rendering, making the process more intuitive and visual.

[0105] Based on the foregoing embodiments, in another embodiment of this application, such as Figure 2 As shown, the reference data generation device for vehicle navigation testing also includes: a first position offsetter 21, at least one second delayer 22, at least one second position offsetter 23, a first modeling comparator 24, and a first stitching smoothing processing module 25.

[0106] Figure 2 In the middle, the map module is used to provide high-precision maps. Furthermore, the accuracy needs to be at the centimeter level. Here, you can still use ready-made maps from map providers, specially collected maps, or crowdsourced maps.

[0107] The first position offsetter 21 is used to offset the first delay sequence to obtain the first offset sequence of the first simulated vehicle in the lane change scenario. The first position offsetter 21 is mainly used to offset the current position by a certain distance to construct the lane change scenario. It ensures that the new position is in the correct lane by comparing it with the map boundary through the boundary judgment module.

[0108] The first position offsetter 21 is connected to multiple serial branches consisting of a second delayer 22 and a second position offsetter 23. Each second delayer 22 is used to time-delay the first offset sequence to obtain a second delayed sequence for time synchronization with the third simulated navigation. Each second position offsetter 23 is used to offset the second delayed sequence to obtain a second offset sequence for at least one second simulated vehicle. The second delayer 22, the second position offsetter 23, and the first position offsetter 21 are used in series to offset the data obtained from the first simulated navigation, constructing complex scene data to support advanced navigation applications such as autonomous driving, V2X, and ADAS. Specifically, while simulating lane changes, a vehicle's information can be simulated in the same time and space, and other vehicle information in the current lane and target lane can be simulated through time delay and position offset. Simultaneously, the offset construction of multiple signals by several second position offsetters 23 can realize the simulation of different vehicle positions and generate scenes with different traffic density. The specific implementation process is as follows: The navigation engine 11 uses the ordered basic position information generated by a virtual navigation to shape the spatial position through a series network consisting of a first position offsetter 21 and several second position offsetters 23, thereby constructing the scene in the spatial dimension. This is another element (i.e., the "spatial" element) for realizing the synchronous spatiotemporal virtual mapping of this application. Further, after the first simulated navigation completes the spatial construction and generates a second offset sequence through the first position offsetter 21 and several second position offsetters 23, the modeling comparator 24 determines the conflict between the map boundary and the boundary of different vehicle positions after the first simulated navigation offset, generating judgment data as the basic signals of radar, vision, and positioning required to support advanced navigation such as autonomous driving. Further, the vision simulator 28, radar simulator 29, and simulator 15 further generate second reference data, namely visual information, radar & lidar, and radio frequency signal information in the real scene. Finally, this information is fully connected to the vehicle controller through the actual interface of the real vehicle for systematic and complete link autonomous driving navigation business testing.

[0109] The first splicing and smoothing processing module 25 is used to splice and smooth the first delayed sequence to obtain a first smoothed delayed sequence;

[0110] The first stitching smoothing module 25 smooths the offset scene data to avoid abrupt changes in position and distortion, and at the same time compensates for the acceleration and angular velocity introduced by the position offset to complete the attitude data synthesis (Gi=G0i+G1i, A=A0i+A1i, G0i is the value obtained by differentiating each axis of the gyroscope, G1i is the component of each axis introduced by the offset, A0i is the acceleration value obtained by differentiating the vehicle speed, A0i=f'(vi), A1i is the component of each axis of the accelerometer introduced by the offset).

[0111] The first modeling comparator 24 is used to model vehicles based on the first offset sequence, the first smooth delay sequence and each of the second offset sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scene and the multiple types of simulated vehicles to obtain conflict determination results;

[0112] The first modeling comparator 24 mainly includes a modeler and a comparator. The modeler can abstract each position coordinate into rectangles of different lengths and widths with the position coordinate as the geometric center. By adjusting the size of the rectangles, different types of vehicles can be simulated, such as a sedan of 3.5*6m and a truck of 4.2*18m. The comparator compares with the map boundary to avoid vehicles appearing in illegal positions. On the other hand, it constructs rectangles based on the modeler and calculates the distance between the current position and the edge of the second position offset 23. If there is a conflict, the offset needs to be adjusted. The parallel time t is calculated based on the second delay 22 [t=(L1+L2) / (v1-v2), and x / (v1-v2)>0, where x is the distance between the two vehicles, positive when the target vehicle is in front and negative when the current vehicle is in front, L1 is the length of the current vehicle, L2 is the length of the target vehicle, v1 is the speed of the current vehicle, and v2 is the speed of the target vehicle]. It is important to note that the coordinate system here should be consistent with the engine's output coordinate system to avoid problems caused by inconsistencies. This article uses the WGS84 coordinate system. All coordinate systems not listed separately are in the WGS84 coordinate system. Other coordinate systems can be handled in the same way.

[0113] The simulator 15 is also used to perform a third simulated navigation based on the simulated scenario and the conflict determination result, and to simulate the signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles to obtain second reference data for testing the on-board navigation.

[0114] In the constructed simulated scenario, simulator 15 performs a third simulated navigation based on the simulated scenario and the conflict determination results. It simulates signals detected by the onboard sensors of the first and second simulated vehicles in a complex navigation scenario, such as position information, time information, collision information (radar recognition results), visual information (feature point simulation), and lidar point clouds (recognition results), to obtain second reference data. This second reference data can be used as standard data and raw signals for onboard navigation testing. As mentioned earlier, the simulator generates radio frequency signals in the real scenario by acquiring ionospheric data and ephemeris data through time synchronization. Here, the offset network, in conjunction with various simulators, generates collision information (radar recognition results), visual information (feature point simulation), and lidar point clouds (recognition results), all of which are results recognized by the onboard sensors. This can equivalently replace the onboard sensors accessing the autonomous driving domain controller and cockpit domain controller. It should be noted that the data access generated in this invention is through existing physical interfaces of the actual vehicle rather than calling various data interfaces, ensuring the integrity and systematic nature of various data transmission links. This provides scenario data support for autonomous driving control, avoidance, and target recognition, enabling real-vehicle operation scenario testing. In this way, during in-vehicle navigation testing, the second reference data can be used as the original signal for the test process, thereby obtaining in-vehicle navigation test data. This test data is then compared with the standard data (the second reference data) to determine the deviation of the in-vehicle navigation test data relative to the standard data (the second reference data), thus obtaining the in-vehicle navigation test result. Simultaneously, this invention achieves spatiotemporal virtual synchronization and ensures that the controller data transmission link is consistent with the actual vehicle. During in-vehicle navigation testing, navigation-related human-machine interface (HMI) rendering information and the in-vehicle voice interaction system can be directly activated, providing users with an immersive scene recreation. Therefore, the scenarios generated by this invention include various navigation test scenarios as well as some scenarios perceived by the user. Simultaneously, evaluation and experience-related experiments can be conducted in the laboratory, which is very user-friendly for non-professional testers.

[0115] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a first traffic sign modeler 26, a third delayer 27, a visual simulator 28, and a radar simulator 29;

[0116] The third delay unit 27 is used to perform a time delay based on the first delay sequence to obtain a third delay sequence;

[0117] First traffic sign modeler 26, used to simulate traffic signs based on the third delay sequence;

[0118] The first traffic sign modeler 26 works in conjunction with the third delayer 27. The third delayer 27, through series connection with the first delayer 14, forms a time sliding mechanism, forward-retrieves a position a certain distance from the current location (s = v * t3, where v is the simulated speed of the current scene, and t3 is the sliding time of the third delayer 27) and marks it. When the marked point is read during scene application, the simulator is awakened to output visual or lidar recognition results to support advanced navigation applications. In addition, the first traffic sign modeler 26 establishes a process from far to near, from edge to center and back to edge, based on the current location and the recognition object: for signs at high positions, the recognition process is mainly constructed by modeling a triangle formed by sliding distance based on the size and height of the sign. Traffic markings can be constructed based on the geometric area formed by high-precision map point cloud or standard lane lines, camera installation height and tilt angle, and the relative distance from the vehicle centerline to the left and right lanes according to calibration parameters to build a visual field.

[0119] The visual simulator 28 is used to simulate the movement of the first simulated vehicle and each of the second simulated vehicles according to the indication signals of the traffic signs; the visual simulator 28 is used to simulate the process of traffic signs and traffic markings moving from a distance, by converting the output signal of the traffic sign modeler into feature values ​​extracted by visual recognition.

[0120] The radar simulator 29 is used to simulate the distance and time between the first simulated vehicle and each of the second simulated vehicles. The radar simulator 29 is used to calculate the distance between the current position and the edge of the position modeling rectangle constructed by the modeler and the second position offset device 23, as well as to simulate the distance and time from the radar sensor in parallel time, to support autonomous driving-related tests such as automatic following and AEB.

[0121] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a first cache 30;

[0122] The first cache 30 is used to temporarily store the first delay sequence and the interpolation navigation data. The first cache 30 is mainly used to temporarily store data such as the first delay sequence and the interpolation navigation data to avoid data loss.

[0123] The first splicing and smoothing processing module 25 is further configured to perform splicing and smoothing processing on the interpolated navigation data to obtain smooth navigation data;

[0124] The memory 19 is also used to store the first smoothing delay sequence and smoothing navigation data.

[0125] This application achieves automatic simulation data generation during the first navigation phase by using a single simulated navigation system in conjunction with multiple second position offsetters and second and third delayers. This reduces the two key steps of traditional simulation test scenario setup and time synchronization, enabling the rapid generation of large amounts of data for various navigation scenarios. In other words, this application achieves a synchronized spatiotemporal virtual mapping area that is synchronous with the real-time environment and reproduces the actual position to the greatest extent possible through a virtual single navigation phase and multiple delayers and position offsetters. This synchronized spatiotemporal virtual mapping enables the rapid generation of large batches of navigation and positioning data. Simultaneously, the virtual single navigation phase perfectly solves the data synchronization problem. Synchronization with the simulator allows for the generation and direct retrieval of various spatial signals for the simulated position, avoiding spatiotemporal asynchrony issues during simulator testing and the inability to cover spatial transmission during simulation testing. This allows for the maximum reproduction of various scenarios in actual vehicle operation, achieving full-link scenario reconstruction covering the physical layer. By using a delay-offset network to process virtual navigation data in an orderly manner, the radar simulator, simulator, and vision simulator are ensured to work in an orderly manner. Simple navigation data and scenarios are constructed into comprehensive, multi-system vehicle-mounted positioning scenario data covering GNSS positioning, visual recognition, and radar systems. This data can be used for testing autonomous driving positioning algorithms and training high-precision positioning big data. It can also be used as a sensor data source for hardware-in-the-loop testing, navigation and positioning function testing, and evaluation experiments. It can completely replace whole-vehicle testing to a certain extent.

[0126] Based on the foregoing embodiments, in another embodiment of this application, such as Figure 3 As shown, the number of the first delay unit 14 is at least two. The reference data generation device for vehicle navigation testing also includes: at least two third position offset units 31, at least two fourth delay units 32, a second modeling comparator 33, and at least two fifth delay units 34. The first delay unit 14, the third position offset unit 31, and the fourth delay unit 32 constitute at least two branches connected in parallel between the interpolation alignment module 13 and the second modeling comparator 33.

[0127] Each of the third position offsetters 31 is used to offset each of the first delay sequences to obtain a third offset sequence of multiple third simulated vehicles;

[0128] Each of the fourth delayers 32 is used to time delay each of the third offset sequences to obtain a fourth delay sequence for time synchronization with the fourth simulated navigation;

[0129] The second modeling comparator 33 is used to model vehicles based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scene and the multiple types of simulated vehicles to obtain conflict determination results. This application uses multiple third position offsetters and fourth delayers connected in parallel to generate multi-target, multi-scene data through a simple simulated navigation. At this time, each branch data is a first reference data, which is used as the original signal for testing. Multi-vehicle operation simulation tests can be performed to achieve scene dynamic simulation. At this time, unless an online map mode is used or other real-time information needs to be obtained, the relative relationships of each scene in this environment are determined, so a time synchronization system can be omitted to reduce system complexity. Furthermore, the vehicle heat map of a specific road at a certain moment (i.e., scene information of other vehicles at the same moment) can be accessed for autonomous driving navigation-related tests and data training. Specifically, the scene data of the vehicle on a certain road after the first simulated navigation, which is constructed by position information, time information, collision information (radar recognition results), visual information (feature point simulation), and lidar point cloud (recognition results), is stored in the memory 19. During testing, the initial position is input to start the test. Simultaneously, a large number of other vehicle heatmaps for that road are accessed via a branch network consisting of three position offsetters 31 and a fourth delayer 32. The simulated test vehicle is placed in this scenario for training various obstacle avoidance, oncoming traffic, and other autonomous driving algorithms, as well as synchronous training on traffic flow and road conditions. This is equivalent to the simulated test vehicle and actual road vehicles operating simultaneously in a synchronous virtual space-time, ensuring the realism of autonomous driving training. Passing a large-scale big data simulation test on a specific road before conducting real-world testing with an autonomous vehicle on the corresponding road can reduce road test risks. Furthermore, once the autonomous driving solution matures, data training can be performed on various road heatmaps to support vehicle loopback detection and improve accuracy before releasing it to users, ensuring safety. In addition, various parallel data accesses can enable extensive V2X scenario testing. For example, when only one data access is available besides the tested vehicle, AEB and V2V tests can be performed; when pedestrian information is accessed, V2P tests can be performed. This application achieves a synchronous spatio-temporal virtual mapping area that is synchronized with the real-world location and reproduces the actual location to the greatest extent possible through a virtual navigation and multiple delayers and position offsetters. By using a virtual mapping system that synchronizes time and space, the system enables the rapid generation of navigation and positioning data in large quantities. Simultaneously, by accessing real-time heatmap data, testing can be conducted directly in the synchronized virtual time and space. On the one hand, there is no need to manually build complex interactive scenarios, and on the other hand, the scenarios and data are real. This is not only useful for training and testing autonomous driving solutions that use loop closure detection, but also has practical significance for improving their accuracy, prediction, and response.(Loop closure detection: When the previous scene and data are identified for the second time, a fast response will be made. Since the scene is real when the vehicle is trained by this invention, the trained vehicle can respond quickly in actual application, thus improving its performance.)

[0130] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: a second traffic sign modeler 35, at least two fifth delayers 34, and a time recovery module 36;

[0131] The time recovery module 36 is used to synchronize the simulated time with the real time. The time recovery module is mainly used to synchronize the time of the generated scene application with the real time according to the time delay sequence. At the same time, since the device does not have a simulator, it can restore the data to a standard format starting from the current time according to the protocol.

[0132] The fifth delayer 34 is used to delay the initial time to obtain a delay time;

[0133] The second traffic sign modeler 35 is used to simulate traffic signs based on the third delay sequence and the delay time.

[0134] In one embodiment of this application, the reference data generation device for vehicle navigation testing further includes: at least two second caches 37 and at least two second stitching smoothing processing modules 38;

[0135] Each of the second caches 37 is used to temporarily store each of the first delay sequences and the interpolated navigation data;

[0136] The second splicing and smoothing processing module 38 is used to splice and smooth each of the first delay sequences to obtain a second smooth delay sequence;

[0137] The memory 19 is also used to store the determination result, the second smooth delay sequence, the indication signal of the second traffic sign modeler, and the delay time.

[0138] The main difference between this embodiment and the previous embodiment is that multiple third position offsetters and fourth delayers are connected in parallel here. This generates a scenario where multiple vehicles are operating simultaneously, creating a multi-path dynamic scenario. In the previous embodiment, the serial connection method only simulated the target's movement; other data were generated as corresponding simulated signals based on time delay sequences, resulting in a semi-static state. This embodiment uses multi-path simulation, which can achieve precise positioning between two targets, crucial for simulating and measuring AEB and V2X. Furthermore, if conditions permit, the target vehicle scenario can be constructed using only the previous embodiment. Since the simulation time is consistent with reality, only the space is virtual, a heatmap of vehicle positions on the actual road can be integrated. This virtualizes the target vehicle to be tested and other vehicles in a certain space at the same time in a "simultaneous" virtual space, which is significant for V2X and autonomous driving big data training. Moreover, considering the large number of simulators required for multi-path signals, resulting in high investment and cost, and taking into account the large interactive characteristics of multi-path testing and the flexibility of the solution, Device 3 does not consider full-link scenario reconstruction in-loop testing based on simulators. It is mainly used for scenario and data generation and autonomous driving data training.

[0139] In another embodiment of this application, a method for generating reference data for vehicle navigation testing is also provided, such as... Figure 4 As shown, it includes:

[0140] Step S101: Import map data and construct a simulated map based on the map data to obtain the input navigation start and end positions;

[0141] In this embodiment, the map data can be existing map data collected by a map provider, or it can be a dedicated map or a crowdsourced map collected by an organization with surveying and mapping qualifications. The map data includes: road connectivity, road type, boundary conditions, traffic signs, latitude and longitude, relative curvature and altitude, road direction, etc. The start and end locations include: the starting and ending locations of the navigation (i.e., the destination).

[0142] It is important to note that if the deflected map data is imported here, the terminal needs to avoid secondary deflection during use to prevent mismatch between the positioning data and the map.

[0143] Step S102: Perform simulated navigation based on the navigation start and end positions in the simulated map to obtain the path planning data and simulated navigation data of the first simulated vehicle;

[0144] Virtual navigation can be performed by inputting the start and end locations through a computer's map engine interface. Virtual navigation includes path planning and simulated navigation. If the vehicle navigation test does not require assessing the navigation capabilities of the test scenario, it can also be based on the shortest path principle [let D]. i,j,kLet D be the shortest path from the input starting point i to the navigation ending point j, and let the main roads and forks along the way be denoted as (1...k). i,j,k =min(D i,k,k-1 +D k,j,k-1 D i,j,k-1 Alternatively, you may need to manually match the routing rules. As long as the route is not planned to a non-road or other illegal area, it will not affect the result. Complete the path planning between the start and end points on the map.

[0145] Alternatively, a navigation application can be used to input the start and end points and complete the navigation route calculation process. Finally, during the virtual navigation process on the application side, map data is systematically called to construct the scene and generate signal sources.

[0146] Step S103: Differentiate the velocity information and curvature information in the simulated navigation data to obtain acceleration information and angular velocity information;

[0147] Step S104: Interpolate the path planning data and simulated navigation data to obtain interpolated data; match the interpolated data, the acceleration information and the angular velocity information, and align them according to a preset period to obtain interpolated navigation data.

[0148] Since the map data is fixed, but the density of the actual data on the time axis varies due to differences in speed, attitude, and scene during the simulation, and also varies with simulation speed, position spacing, and landing points on the map, interpolation processing of the map data is required to cover the needs of different scenarios. Specifically:

[0149] Perform linear interpolation in the linear region [x] i =x i-1 +(1 / k)*f(v i ),x i-1 f(v) represents the position at the previous time step. i Let v be the simulated vehicle speed at time i. i The vehicle speed converted to the simulation requirements;

[0150]

[0151] m is the scaling factor; if rapid completion is not required, m = 1; otherwise, a larger m results in faster completion, but it should not exceed the data latency. h is the scaling factor for speed-limited or speed-regulated scenarios; a value between 0.8 and 1.2 is recommended (except for special scenarios such as speeding zones). k is the data frequency (k = 1 or k = 10; generally, GPS uses 1 or 10, and IMU uses 10 or 100). Data curves are interpolated according to curvature [x]. i =x i-1 +(1 / k)*f(vi )*F(e i ),e i To ensure data continuity and consistency for the current position curvature, fine-tuning of height and roll is required based on relative height and road inclination angle. Furthermore, since applications have varying data requirements, clock alignment must be ensured within the interpolation cycle to prevent misalignment between position and attitude information. For example, if position information requires 1Hz and attitude data requires 10Hz, then for every position data output, 10 attitude data points must be aligned, and so on. If the actual scene map used already covers these requirements, this step can be omitted.

[0152] Step S105: Delay the interpolated navigation data to obtain a first delay sequence for time synchronization with the second simulated navigation;

[0153] Establishing a delay sequence: This involves adding time stamps to the aforementioned interpolated navigation data. The simulated navigation start time (i.e., the time when the first set of data is generated) is taken as the time starting point T0. The data requires an interval F (F = 1 or 10, specifically based on the GPS data output frequency as a basic reference value). The interpolated data is sequentially delayed based on T0 to form a delay sequence (Ti = T0 + i * F). In application scenarios, the simulated time is used as the time starting point, and this sequence is used to align time with the real world in an orderly manner. This allows for simulation testing using dynamic map layer data and the use of parameters such as the ionosphere during simulation, further ensuring data authenticity and enabling the use of heatmaps from other terminals for big data testing and training.

[0154] Step S106: The acceleration information and the angular velocity information are synthesized to obtain position and attitude vector synthesis data; a simulated scene is constructed based on preset scene parameters, and real-time scene data corresponding to the second simulated navigation is obtained. The interpolated navigation data and the real-time scene data are used to generate radio frequency signals. In the simulated scene, the second simulated navigation is performed based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence. The signals of the on-board sensors of the first simulated vehicle are simulated to obtain first reference data for testing on-board navigation.

[0155] In this embodiment, A. Scene parameter settings: The preset scene parameters are pre-set by the user and include multiple operating condition parameters. This allows for rapid virtual navigation and shortens scene generation time, but a speed model needs to be input to determine the position and density of the generated scene relative to the map, as well as the main acceleration components. Speed ​​information for each stage of virtual navigation can be automatically generated based on road speed limits, traffic regulations, and driving habits. Alternatively, speeds can be manually set based on testing needs, such as setting highways to 100 km / h and 90° curves to 20 km / h to cover extreme operating conditions and scenarios. During rapid traversal, speeds need to be set according to v. i =f -1 (v 1i ), [v1i is the manually set velocity at time i, f -1 (v 1i ) is f(v i The inverse function of ) is used for conversion. Furthermore, it is necessary to set the location and density of conflicts and positional offsets on the map to simulate the timing and frequency of radar signal simulation, as well as lane change and visual simulation timing, etc.

[0156] B. Position Offset: Based on scene density parameters, the current position is offset by a certain distance in the WGS84 coordinate system according to a certain frequency (density) using a position offsetter. The new position can be used to simulate lane change scenarios or as a source of information from other vehicles, in order to simulate the distance, azimuth, etc. of radar signals. First, according to the formulas a1=(N+H)cosbcosL, b1=(N+H)cosBsinL, c1=[N(1-e 2 )+H]sinB, where (L,B,H) are the geodetic latitude, longitude, and geodetic height of the simulated object, respectively; (a1,b1,c1) is the position calculated in the spatial rectangular coordinate system; and N is the radius of the ellipsoid's prime meridian. e is the first eccentricity of the ellipsoid. a and b represent the Earth's major and minor axes, respectively. In a Cartesian coordinate system, the target vehicle is modeled using a rectangle centered at (a1, b1, c1) and defined by vehicle length L1 and width W1. Interactive vehicles are modeled using a rectangle centered at (a2, b2, c2) and defined by vehicle length L2 and width W2, and so on, to construct information for other vehicles. At this point, information is generated based on the distance between two points within each rectangle. It can determine the relative position of the two vehicles and output a radar simulation signal.

[0157] C. Secondary time offset, which is achieved by concatenating with the delay sequence to realize time sliding, thereby marking the time when traffic signs appear to simulate the process from far to near, thus realizing the simulation of the visual recognition process (t=t1±t2,t1 is the delay time of delay device 1,t2 is the delay time of delay device 2).

[0158] When simulating the signals of the on-board sensors of the first simulated vehicle based on the simulated scenario and the first delay sequence, conflict determination is required. This is mainly based on the map boundary to determine the offset and delayed data, avoiding the location from appearing in non-road areas such as green belts, lakes, and residential buildings, which will be referred to as "illegal" locations. In addition, when simulating the offset location as a simulation of other vehicles, it is also necessary to determine the distance and orientation between the two locations in order to determine the vehicle distance between the two locations in order to simulate the corresponding radar sensing data. The determination scheme is similar to the offset and scenario construction process. At this time, the target vehicle changes to the boundary location information based on map type and semantics.

[0159] In this embodiment, the offset scene data can be smoothed to avoid abrupt changes and distortions in position, while compensating for the acceleration and angular velocity of the position offset. Generally, the simulator comes with a "Smooth Movement" component. For multi-target vehicle simulation without a simulator, the actual running trajectory can be smoothly fitted using cyclotron curves, polynomial curves, Bézier curves, or spline curves based on kinematic and dynamic constraints. This is a mature technical solution and will not be described further here.

[0160] In practical applications, an exemplary method for generating reference data for in-vehicle navigation testing can be as follows: Figure 5 As shown, map data can be imported first; virtual navigation can be performed by inputting the start and end positions of navigation: path planning can be completed for simulated navigation or fast data traversal; interpolation and period alignment can be performed: interpolation processing and period alignment within different data sources can be completed based on map data and type; a delay sequence can be established: a delay sequence can be established based on the start time to complete the overall clock delay of the data source; offset and scene construction can be performed: data position offset can be used to construct other vehicle information; a second time delay can be used to simulate visual signals; based on the conflict determination result, it can be determined whether any simulated vehicle has a conflict with the position of other simulated vehicles, and based on the map data, it can be determined whether any simulated vehicle exceeds the map boundary; if based on the conflict determination result, it can be determined that the positions of each simulated vehicle do not conflict with the positions of other simulated vehicles, and based on the map data, it can be determined that each simulated vehicle has not exceeded the map boundary, then each offset delay sequence is spliced ​​and smoothed to generate a simulation data source (i.e., the first reference data).

[0161] If, based on the conflict determination result, it is determined that any simulated vehicle is in conflict with the positions of other simulated vehicles, or, based on the map data, it is determined that any simulated vehicle exceeds the map boundary, then it is determined whether a collision scenario needs to be constructed. If a collision scenario needs to be constructed, then a collision scenario is constructed for the simulated vehicles with conflicting positions, or for the simulated vehicles that exceed the map boundary, and then the corresponding offset delay sequence is spliced ​​and smoothed to generate a simulation data source (i.e., the first reference data).

[0162] To facilitate understanding, the following will be combined with... Figures 6-7 The principle of the above scheme will be explained, and the following will be used: Figure 1 The in-vehicle navigation test reference data generation device shown, when generating ordinary navigation test scenarios and data based on the in-vehicle navigation test reference data generation method, is used in this case because, due to map accuracy limitations, the scenarios are all at the road level. Unless it is necessary to test and evaluate the functional performance when the positioning is inaccurate, no offset or scene construction is required. The in-vehicle navigation test reference data generation process is as follows: Figure 6 As shown.

[0163] After the map import is complete, the start and end positions are entered. Virtual (simulated) navigation begins at time S1 and reaches the virtual navigation endpoint at time S4. The corresponding map data during the virtual navigation process are S5 to S8. The velocity in this process is differentiated to obtain the basic acceleration, and the curvature is differentiated to obtain the basic angular rate. Simultaneously, based on local gravitational acceleration, slope, curvature, pitch, and roll angles, three-dimensional acceleration and angular velocity are synthesized. Figure 6 The location S6 in the middle does not have corresponding data on the map, so interpolation is required. After smoothing, the result can be obtained. Figure 7 The scene and data shown are as follows: S9 to S12 constitute the delay sequence, and S13 to S16 constitute the position and attitude data.

[0164] During testing, stored scene data (location information, delay sequence, start and end positions), IMU inertial navigation specifications, vehicle data, and network data (connecting to online maps, downloading ionospheric data, and downloading target location ephemeris books) are simultaneously imported into or connected to the simulator. The simulator is mainly used to convert position signals into radio frequency signals. It can also match different channel models based on sub-scenes fed back from the map. For example, an open sky scene can be selected for highways and urban expressways, while an urban canyon scene can be selected for commercial districts and urban areas. Since the delay unit can achieve virtual synchronization with real time, the ionospheric parameters of the location can be downloaded for direct simulation and correction simulation. In addition, the simulator also supports the import of inertial unit temperature drift, zero-bias curves, installation location, and vehicle vibration parameters to complete the inertial data restoration.

[0165] In another embodiment of this application, the method for generating reference data for vehicle navigation testing further includes:

[0166] The first delay sequence is offset to obtain the first offset sequence of the first simulated vehicle in the lane change scenario;

[0167] The first offset sequence is time-delayed to obtain a second delayed sequence for time synchronization with the third simulated navigation. The second delayed sequence is then offset to obtain a second offset sequence for at least one second simulated vehicle.

[0168] The first delayed sequence is spliced ​​and smoothed to obtain the first smoothed delayed sequence;

[0169] Vehicle modeling is performed based on the first offset sequence, the first smooth delay sequence, and each of the second offset sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results.

[0170] Based on the simulated scenario and the conflict determination result, a third simulated navigation is performed. The signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles are simulated to obtain second reference data for testing of the on-board navigation.

[0171] To facilitate understanding, the following will be combined with... Figures 8-10 The principle of the above scheme will be explained, and the following will be used: Figure 2 The reference data generation device for vehicle navigation testing shown herein, when generating high-precision navigation and positioning test scenarios and data based on the reference data generation method for vehicle navigation testing, the reference data generation process for vehicle navigation testing is as follows: Figure 8 As shown. The process of generating position and attitude is basically the same as that of ordinary navigation test scenarios and data generation, and will not be described in detail here.

[0172] like Figure 8 As shown, the imported map is a lane-level high-precision map, which allows for the construction of complex scenes. For example... Figure 8 As shown, position S22 is passed through delay unit I and offset unit I to obtain S18, and position S24 is passed through delay unit I and offset unit I to obtain S26; S22 is passed through offset unit II and delay units I and II to obtain S17, S23, S19, and S25 respectively, and S27 is passed through offset unit II and delay units I and II to obtain S28 respectively; delay unit III constructs a time sliding sequence through delay sequence III, and at position S18, it starts to simulate camera recognition data based on traffic sign modeler and marks it in delay sequence I based on sliding time to ensure that visual recognition data is simulated according to the set time during scene testing; position conflict judgment compares the new position after deflection and delay. If S26 and S28 conflict, the position offset value will be modified. If the positions of S26 and S28 are still close, a simulated radar signal will be generated. The modeler mainly sets a certain range of target position as "vehicle" (i.e., a range of 3.5*6m centered on the target position) and determines the parallel time based on the time difference between delay unit II and delay unit I and the relative length value of "vehicle". After smoothing, the generated signal is shown in the figure. Figure 9 The scene and data shown are saved. The simulator usage is the same as when generating normal navigation scenes and data, and will not be elaborated here. Furthermore, different types of "vehicles" can be constructed by modifying the target position size of Offset II, such as... Figure 10 As shown.

[0173] In another embodiment of this application, the method for generating reference data for vehicle navigation testing further includes:

[0174] The first delay sequence is offset to obtain a third offset sequence of multiple third simulated vehicles;

[0175] Each of the third offset sequences is time-delayed to obtain a fourth delayed sequence for time synchronization with the fourth simulated navigation;

[0176] Vehicle modeling is performed based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results.

[0177] A fourth simulated navigation is performed based on the simulated scenario and the conflict determination result. The signals of the on-board sensors of multiple third simulated vehicles are simulated to obtain third reference data for testing of on-board navigation.

[0178] To facilitate understanding, the following will be combined with... Figure 11 The principle of the above scheme will be explained, and the following will be used: Figure 3 The reference data generation device for vehicle navigation testing shown herein, when generating high-precision navigation and positioning test scenarios and data based on the reference data generation method for vehicle navigation testing, the reference data generation process for vehicle navigation testing is as follows: Figure 11 As shown. The position and attitude generation process is basically the same as the high-precision navigation and positioning test scenario and data generation in the previous embodiment, and will not be repeated here. The main difference is that there are only multiple offset units I and delay units I connected in parallel. At this time, in addition to the target vehicle S32, S33, and S34, S29, S30, S31, S35, S36, S37, and S28 correspond to data sources with temporal and spatial characteristics generated by different offset units I and delay units I, as shown. Figure 11 As shown. Figure 8 The high-precision navigation and positioning test scenarios and data shown only have temporal and spatial characteristics for target vehicles S21, S22, S24, and S24; other vehicles only have temporal characteristics (these are specific data generated at fixed times based on scenario setting parameters and the modeler). Furthermore, S29, S30, S31, S35, S36, S37, and S28 can also be based on road vehicle heatmaps (i.e., the feature points and distribution overview of other vehicles on the map at this time). This allows for multi-path testing and integration with actual road conditions, playing a crucial role in V2X and autonomous driving simulation testing.

[0179] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0180] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A reference data generation device for in-vehicle navigation test, characterized by, include: Navigation engine, differentiator, interpolation alignment module, first delay unit, and simulator; The navigation engine is used to import map data, construct a simulated map based on the map data, obtain the input navigation start and end positions, perform a first simulated navigation based on the navigation start and end positions in the simulated map, and obtain the path planning data and simulated navigation data of the first simulated vehicle. The differentiator is used to differentiate the velocity information and curvature information in the simulated navigation data to obtain acceleration information and angular velocity information. The interpolation alignment module is used to interpolate the path planning data and simulated navigation data to obtain interpolated data, match the interpolated data, the acceleration information and the angular velocity information, and perform alignment processing according to a preset period to obtain interpolated navigation data. The first delayer is used to delay the interpolated navigation data to obtain a first delay sequence for time synchronization with the second simulated navigation; The simulator is used to synthesize the acceleration information and the angular velocity information to obtain position and attitude vector synthesis data; A simulated scenario is constructed based on preset scenario parameters. Real-time scenario data corresponding to the second simulated navigation is obtained. The interpolated navigation data and the real-time scenario data are used to generate radio frequency signals. In the simulated scenario, the second simulated navigation is performed based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence. The signals of the on-board sensors of the first simulated vehicle are simulated to obtain first reference data for testing on-board navigation.

2. The reference data generation device for vehicle navigation testing according to claim 1, characterized in that, Also includes: Navigation start and end location input module and map module; The navigation start and end position input module is used to receive the input navigation start and end positions and send them to the navigation engine; The map module is used to provide the map data and send the map data to the navigation engine.

3. The reference data generation apparatus for in-vehicle navigation testing according to claim 2, characterized by, Also includes: Time synchronization module and memory; The timing module is used to synchronize the navigation engine, the differentiator, the interpolation alignment module, the first delay unit, and the simulator to obtain time synchronization information. The memory is used to store the navigation start and end positions, the first delay sequence, the path planning data, the acceleration information, the angular velocity information, and the time synchronization information.

4. The reference data generation device for vehicle navigation testing according to claim 3, characterized in that, Also includes: The system comprises a first position offsetter, at least one second delayer, at least one second position offsetter, a first modeling comparator, and a first stitching smoothing module. The first position offset is used to offset the first delay sequence to obtain the first offset sequence of the first simulated vehicle in the lane change scenario. The first position offset is connected to multiple serial branches composed of a second delay and a second position offset. Each of the second delayers is used to time delay the first offset sequence to obtain a second delayed sequence for time synchronization with the third simulated navigation; each of the second position offsetters is used to offset the second delayed sequence to obtain a second offset sequence for at least one second simulated vehicle. The first splicing and smoothing processing module is used to splice and smooth the first delayed sequence to obtain a first smoothed delayed sequence. The first modeling comparator is used to model vehicles based on the first offset sequence, the first smooth delay sequence and each of the second offset sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results; The simulator is also used to perform a third simulated navigation based on the simulated scenario and the conflict determination result, and to simulate the signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles to obtain second reference data for testing the on-board navigation.

5. The reference data generation device for vehicle navigation testing according to claim 4, characterized in that, Also includes: First traffic sign modeler, third delayer, visual simulator, and radar simulator; The third delayer is used to perform a time delay based on the first delay sequence to obtain a third delay sequence; A first traffic sign modeler is used to simulate traffic signs based on the third delay sequence; The visual simulator is used to simulate the movement of the first simulated vehicle and each of the second simulated vehicles according to the indication signals of the traffic signs; The radar simulator is used to simulate the distance and time between the first simulated vehicle and each of the second simulated vehicles.

6. The reference data generation apparatus for in-vehicle navigation testing according to claim 4, wherein Also includes: First cache; The first cache is used to temporarily store the first delay sequence and the interpolated navigation data; The first splicing and smoothing processing module is further configured to perform splicing and smoothing processing on the interpolated navigation data to obtain smooth navigation data; The memory is also used to store the first smoothing delay sequence and smoothing navigation data.

7. The reference data generation apparatus for in-vehicle navigation testing according to claim 3, wherein The number of the first delay unit is at least two, and the reference data generation device for vehicle navigation testing further includes: at least two third position offset units, at least two fourth delay units, a second modeling comparator, and at least two fifth delay units, wherein the first delay unit, the third position offset unit, and the fourth delay unit constitute at least two branches connected in parallel between the interpolation alignment module and the second modeling comparator; Each of the third position offsetters is used to offset each of the first delay sequences to obtain a third offset sequence of multiple third simulated vehicles; Each of the fourth delayers is used to time delay each of the third offset sequences to obtain a fourth delayed sequence for time synchronization with the fourth simulated navigation; The second modeling comparator is used to model vehicles based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles, and to perform conflict determination based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results; The simulator is also used to perform a fourth simulated navigation based on the simulated scenario and the conflict determination result, and to simulate the signals of the on-board sensors of multiple third simulated vehicles to obtain third reference data for testing on-board navigation.

8. The reference data generation apparatus for in-vehicle navigation testing according to claim 7, characterized by, Also includes: Second traffic sign modeler, at least two fifth delayers and time recovery module; The time recovery module is used to synchronize simulated time with real time; The fifth delayer is used to delay the initial time to obtain the delay time; A second traffic sign modeler is used to simulate traffic signs based on a third delay sequence and the delay time.

9. The reference data generation apparatus for in-vehicle navigation testing according to claim 8, characterized by, Also includes: At least two secondary caches and at least two secondary splicing smoothing modules; Each of the second caches is used to temporarily store each of the first delay sequences and the interpolated navigation data; The second splicing and smoothing processing module is used to splice and smooth each of the first delay sequences to obtain a second smooth delay sequence; The memory is also used to store the determination result, the second smooth delay sequence, the indication signal of the second traffic sign modeler, and the delay time.

10. A method for generating reference data for vehicle navigation testing, characterized in that, include: Import map data and construct a simulated map based on the map data to obtain the input navigation start and end positions; Based on the navigation start and end positions, a first simulated navigation is performed on the simulated map to obtain the path planning data and simulated navigation data of the first simulated vehicle; The velocity and curvature information in the simulated navigation data are differentiated to obtain acceleration and angular velocity information. The path planning data and simulated navigation data are interpolated to obtain interpolated data. The interpolated data, the acceleration information and the angular velocity information are matched and aligned according to a preset period to obtain interpolated navigation data. The interpolated navigation data is time-delayed to obtain a first delay sequence, which is then synchronized with the second simulated navigation. The acceleration information and the angular velocity information are combined to obtain position and attitude vector composite data; A simulated scenario is constructed based on preset scenario parameters. Real-time scenario data corresponding to the second simulated navigation is obtained. The interpolated navigation data and the real-time scenario data are used to generate radio frequency signals. In the simulated scenario, the second simulated navigation is performed based on the radio frequency signals, the position and attitude vector synthesis data and the first delay sequence. The signals of the on-board sensors of the first simulated vehicle are simulated to obtain first reference data for testing on-board navigation.

11. The method for generating reference data for vehicle navigation testing according to claim 10, characterized in that, Also includes: The first delay sequence is offset to obtain the first offset sequence of the first simulated vehicle in the lane change scenario; The first offset sequence is time-delayed to obtain a second delayed sequence for time synchronization with the third simulated navigation. The second delayed sequence is then offset to obtain a second offset sequence for at least one second simulated vehicle. The first delayed sequence is spliced ​​and smoothed to obtain the first smoothed delayed sequence; Vehicle modeling is performed based on the first offset sequence, the first smooth delay sequence, and each of the second offset sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results. Based on the simulated scenario and the conflict determination result, a third simulated navigation is performed. The signals of the on-board sensors of the first simulated vehicle and each of the second simulated vehicles are simulated to obtain second reference data for testing of the on-board navigation.

12. The method of claim 10, wherein, Also includes: The first delay sequence is offset to obtain a third offset sequence of multiple third simulated vehicles; Each of the third offset sequences is time-delayed to obtain a fourth delayed sequence for time synchronization with the fourth simulated navigation; Vehicle modeling is performed based on multiple third offset sequences, second smooth delay sequences, and each of the fourth delay sequences to obtain multiple types of simulated vehicles. Conflict determination is then performed based on the simulated scenario and the multiple types of simulated vehicles to obtain conflict determination results. A fourth simulated navigation is performed based on the simulated scenario and the conflict determination result. The signals of the on-board sensors of multiple third simulated vehicles are simulated to obtain third reference data for testing of on-board navigation.

Images