A Valet Parking Simulation Method and Device

Abstract

The present application provides a valet parking simulation method and device, which reversely converts the simulated open parking lot map into target features that can be recognized by the parking controller in real time, and then broadcasts the target features in the area where the target simulated vehicle is located to the parking controller based on the positioning position, and reconstructs the broadcasted target features in real time in the area, thereby obtaining a universal open parking lot map of the target simulated vehicle in the area where it is located, and downloads and updates the universal open parking lot map to the parking controller for use in the parking controller's path planning. This can solve the problem of how to perform algorithm verification in the L4 valet parking system in the simulation system, and provide reliable guarantee for the algorithm verification of the L4 valet parking system.

Description

Technical Field

[0001] The present application relates to the field of simulation technology, and more specifically, to a valet parking simulation method and device. Background Art

[0002] Currently, the core difference between Automated Valet Parking (AVP) and AutoParking Assist (APA) lies in the addition of map information. Therefore, the core of valet parking simulation verification is the simulation of map information. However, most current simulation verification schemes remain at the L3+ level. Most of them require pre-downloading real closed parking lot maps in the parking controller, and then reproducing exactly the same closed parking lot map in the simulated parking lot map scenario for the simulation of valet parking map information.

[0003] To obtain map information, the parking controller needs to occupy resources for parsing. Therefore, only a few closed parking lot maps of limited size can be pre-downloaded. However, in real application scenarios, the number of urban open parking lots is huge. Pre-downloading all these huge parking lot information into the parking controller is obviously unrealistic due to hardware and software limiting factors such as memory and processors. Summary of the Invention

[0004] In view of this, to solve the above problems, the present application provides a valet parking simulation method and device, and the technical solutions are as follows:

[0005] On the one hand, the present application provides a valet parking simulation method, and the method includes:

[0006] Obtain a simulated open parking lot map;

[0007] Perform reverse transformation on the simulated open parking lot map to obtain target features that can be recognized by the parking controller; wherein, the simulated open parking lot map is a custom-built simulated open parking lot map, and the target features include road features and lane features;

[0008] Based on the positioning position of the target simulated vehicle, broadcast the target features within the area where the target simulated vehicle is located to the parking controller and perform real-time drawing and reconstruction of the broadcast target features within the area where the target simulated vehicle is located to obtain a general open parking lot map within the area where the target simulated vehicle is located, and download and update the general open parking lot map to the parking controller.

[0009] Optionally, the performing reverse transformation on the simulated open parking lot map includes:

[0010] Parse the simulated open parking lot map;

[0011] Extract the target features within the area where the target simulated vehicle is located from the parsing result of the simulated open parking lot map;

[0012] Based on the ADASIS protocol, determine the unavailable features in the target features within the area where the target simulated vehicle is located that are mapped to the simulated open parking lot map;

[0013] Transform the unavailable features.

[0014] Optionally, the transforming of the unavailable features includes:

[0015] Perform data missing transformation / data difference transformation on the unavailable features.

[0016] Optionally, the reverse transformation of the simulated open parking lot map further includes:

[0017] Screen the unique features of the open parking lot from the target features.

[0018] Optionally, the unique features include:

[0019] Parking space features, parking space line features, and parking space line corner features.

[0020] On the other hand, the present application provides a valet parking simulation device, the device includes:

[0021] A simulation map acquisition module, configured to acquire a simulated open parking lot map;

[0022] A general map transformation module, configured to perform reverse transformation on the simulated open parking lot map to obtain target features that can be recognized by a parking controller; wherein, the simulated open parking lot map is a custom-built simulated open parking lot map, and the target features include road features and lane features;

[0023] A broadcast and download module, configured to, based on the positioning position of the target simulated vehicle, broadcast the target features within the area where the target simulated vehicle is located to the parking controller and perform real-time rendering and reconstruction of the broadcast target features within the area where the target simulated vehicle is located to obtain a general open parking lot map within the area where the target simulated vehicle is located, and download and update the general open parking lot map to the parking controller.

[0024] Optionally, the general map transformation module is specifically configured to:

[0025] Parse the simulated open parking lot map; extract the target features within the area where the target simulated vehicle is located from the parsing result of the simulated open parking lot map; determine the unavailable features mapped to the simulated open parking lot map among the target features within the area where the target simulated vehicle is located based on the ADASIS protocol; transform the unavailable features.

[0026] Optionally, the general map transformation module for transforming the unavailable features is specifically configured to:

[0027] Perform data missing transformation / data difference transformation on the unavailable features.

[0028] Optionally, the general map transformation module is further configured to:

[0029] Filter out the unique features of the open parking lot from the target features.

[0030] Optionally, the unique features include:

[0031] Parking space features, parking space line features, and parking space line corner features.

[0032] Compared with the prior art, the beneficial effects achieved by this application are:

[0033] This application provides a valet parking simulation method and device. By reversely transforming the simulated open parking lot map into target features recognizable by the parking controller in real time, and then broadcasting the target features within the area where the target simulated vehicle is located to the parking controller based on the positioning position of the target simulated vehicle and reconstructing the broadcast target features in real time within this area, a general open parking lot map of the target simulated vehicle within its area is obtained. This general open parking lot map is downloaded and updated to the parking controller for the parking controller to use for path planning. This can solve the problem of how to perform algorithm verification for the L4-level valet parking system in the simulation system, providing a reliable guarantee for the algorithm verification of the L4-level valet parking system. Description of the Drawings

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0035] Figure 1 It is the flowchart of the valet parking simulation method provided by the embodiment of the present application;

[0036] Figure 2It is a partial method flowchart of the valet parking simulation method provided by the embodiments of this application;

[0037] Figure 3 It is a structural schematic diagram of the valet parking simulation device provided by the embodiments of this application. Detailed implementation manners

[0038] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the protection scope of this application.

[0039] To make the above objects, features, and advantages of this application more obvious and understandable, the following further details this application in conjunction with the drawings and specific implementation manners.

[0040] With the successive R & D of L4-level intelligent driving vehicles by major vehicle manufacturers, single real vehicle tests are no longer sufficient to verify various working conditions. For the test and verification of valet parking algorithms, it is no longer limited to monitoring the surrounding areas of the vehicle. In the current situation where the infrastructure is incomplete, it is unrealistic to rely on V2X and cloud technologies to enable the vehicle to obtain information from farther places. Therefore, it is a wiser choice to provide the vehicle with the ability to perceive distant information based on the map.

[0041] Currently, the core difference between Automated Valet Parking (AVP) and AutoParking Assist (APA) lies in the addition of map information. Therefore, the core of valet parking simulation verification is the simulation of map information. However, most current simulation verification schemes stay at the L3+ level. Most of them require pre-downloading the real closed parking lot map in the parking controller, and then reproducing the exact same closed parking lot map in the simulated parking lot map scenario for the simulation of valet parking map information. The main ways to unify the real closed parking lot map and the simulated closed parking lot map are as follows:

[0042] The first is forward map information conversion, that is, converting the real closed parking lot map downloaded in the parking controller into a format available for the simulation scenario and loading it into the simulation scenario to generate a virtual closed parking lot map. The second method is to manually build and reproduce the real closed parking lot map in the simulation scenario.

[0043] In the first method mentioned above, since the forward conversion can only obtain the real closed parking lot map in advance, the scenario is single, and it is not easy to obtain special scenarios. Moreover, the conversion requires the service support of a map provider, who will charge according to the number of conversions and the quantity of conversions. The customization is relatively high and it lacks universality, making the actual operation difficult. In the second method mentioned above, the workload is large, the reproducibility is low, and it is difficult to ensure that the manually built closed parking lot map is exactly the same as the real one, and it takes too long, making it difficult to ensure the development progress.

[0044] The inventor analyzed that the biggest problem with the above two map simulation schemes is that both require pre-selecting and downloading the real closed parking lot map in the parking controller, so that the parking controller knows all the information of the parking lot before simulation. However, in real application scenarios, there are often situations where the closed parking lot map has not been drawn and cannot be downloaded, or there is no closed parking lot available for valet parking in the area, and the valet parking function cannot be used.

[0045] In this way of pre-downloading the closed parking lot map to the parking controller, the parking controller needs to occupy resources for parsing to obtain the map information. Therefore, only a few closed parking lot maps of limited size can be pre-downloaded. However, in most real application scenarios, parking is done in urban open parking lots. The number of urban open parking lots is huge. If all this huge amount of parking lot information is pre-downloaded into the parking controller, due to hardware and software limiting factors such as memory and processor, this scheme is obviously unrealistic.

[0046] Therefore, if the parking controller can be made to know the open parking lot map information within a specified number of kilometers in real time and perform path planning in real time, these problems can be fundamentally solved, achieving the effect of verifying the L4-level valet parking algorithm, while the current two simulation schemes cannot complete the verification of the L4-level open parking lot algorithm.

[0047] Based on the above analysis of the simulation requirements of the L4-level valet parking system, this application proposes two key solutions:

[0048] One is to no longer use the forward conversion scheme of transforming the real parking lot map scenario into a simulation parking lot map scenario, but instead use the reverse conversion scheme of transforming the simulation parking lot map scenario into a real parking lot map scenario that the parking controller can recognize. This can solve the problem of limited algorithm verification due to the single scenario of the forward conversion scheme, and make the parking lot map scenario more flexible and rich by being arbitrarily built according to user needs.

[0049] Second, the custom-built simulation open parking lot map is converted into a general open parking lot map recognizable by the parking controller in real time within a specified kilometer range, broadcast on CAN / Ethernet, and the broadcast open parking lot map is reconstructed in real time within the specified kilometer range to reconstruct the parking lot map within the specified kilometer near the vehicle for the parking controller to use for path planning within the specified kilometer range. This can solve the problem of how to verify the algorithm in the L4 valet parking system in the simulation system and provide reliable guarantee for the algorithm verification of the L4 valet parking system. The implementation background of this application is described in detail below.

[0050] The valet parking simulation method provided by this application can be implemented by running a scenario reversal program. The scenario reversal program runs on the software and hardware platform of the hardware-in-the-loop system. When the scenario reversal program runs, it reversely converts the simulation open parking lot map into a general open parking lot map recognizable by the parking controller, broadcasts it on CAN / Ethernet, and reconstructs the parking lot map within the specified kilometer range for the parking controller loaded with the valet parking algorithm to use.

[0051] It should be noted that the simulation open parking lot map in this application can be built through the scenario editor of OpenDRIVE. OpenDRIVE is a high-precision map format that has been widely used in the unmanned driving or map positioning industry.

[0052] The hardware-in-the-loop system uses the parking controller loaded as a physical object to realize the real-time operation of the simulation vehicle model: verify all input signals related to the simulation parking controller, collect all output signals related to the verification of the parking controller and necessary input signals, run the vehicle model (including vehicle dynamics model, engine model, transmission model, etc.), simulate the motion posture of the vehicle and its response to control instructions, and run scenario software (such as VTD, VIRES Virtual Test Drive) to simulate various operating conditions of the vehicle. Algorithm verification personnel can also continuously enrich their simulation scenario library by building OpenDRIVE simulation scenarios.

[0053] The parking controller is loaded with the valet parking algorithm to be verified and tested, and it has interfaces such as CAN, Ethernet, IO, and LVDS. It interacts with the entire system information (including the parking lot map reconstructed by the scenario reversal program and the input signals input to the parking controller during the real-time operation of the vehicle model simulated by the hardware-in-the-loop system) to realize the functions of algorithms such as perception fusion, obstacle avoidance, vehicle positioning, path planning, task scheduling, trajectory tracking, lateral and longitudinal motion control, image display, and vehicle body safety and comfort control.

[0054] The hardware-in-the-loop system includes a real-time multiprocessor platform. The real-time simulator and the graphics workstation in the real-time multiprocessor platform operate in a master-slave combined simulation mode: the real-time simulator serves as the master, mainly used for running vehicle dynamics models, road traffic scene models, IO models, etc.; one graphics workstation serves as slave one, running test management (such as TCS, TestBase Control Software) and test software (such as TAE, Testcase Automation Executor); another graphics workstation serves as slave two, used for the calculation and output of the four-way camera sensors used in the 360-degree surround view simulation.

[0055] The hardware-in-the-loop system also includes components such as various IO boards. The IO boards are inserted in the real-time simulator and the external expansion chassis, mainly used for providing functions such as simulation and acquisition of general analog-digital IO signals, simulation and acquisition of bus communications such as CAN / LIN / automotive Ethernet, etc.

[0056] The host is equipped with vehicle dynamics model software (such as Modelbase) for simulating the dynamic behavior of the vehicle, such as the processes of acceleration, braking, steering, etc. The host also includes the simulation software VTD (VIRES Virtual TestDrive) for simulating all scene elements in the virtual parking lot required for validating the valet parking algorithm. Through the above design, the hardware-in-the-loop system realizes vehicle simulation control (provided by the vehicle dynamics model software Modelbase, IO model, IO board, four-channel camera sensors used for simulated 360-degree surround view, etc.) and provides the scene simulation environment VTD (VIRES Virtual Test Drive). Vehicle simulation control is the information interaction between the simulated virtual vehicle and virtual vehicle control components such as virtual EPS (Electric Power Steering), ESP (Electronic Stability Program), and TCU (Transmission Control Unit), simulating the relevant functions, signals, and handshake logics of the vehicle. Among them, the virtual ESP (Electronic Stability Program) realizes the control of the virtual HCU (Hybrid Control unit) and TCU (Transmission Control Unit), and completes the longitudinal control of the vehicle by receiving the message information sent by the parking controller. The virtual EPS (Electric Power Steering) completes the handshake logic, steering control, and message sending of the steering system under the parking function, closes the loop of the parking controller, and completes the closed-loop control of the relevant functions of the simulated vehicle. The scene reversal program of the present application is arranged in the linux real-time system of the host, and when it runs, it can implement the valet parking simulation method provided by the present application. The valet parking simulation method will be described below.

[0057] An embodiment of the present application provides a valet parking simulation method, and the method flow chart of this method is as Figure 1 shown, including the following steps:

[0058] S10, obtain the map of the simulated open parking lot.

[0059] As above, in the embodiment of the present application, the map of the simulated open parking lot can be custom-built through the scene editor of OpenDRIVE. Using the OpenDRIVE scene editor, build an open parking lot simulation scene large enough according to the test requirements, import the map of the simulated open parking lot into the scene simulation software, such as VTD (VIRES Virtual Test Drive), and the scene reversal program interacts with the VTD software.

[0060] S20. Reverse-transform the simulated open parking lot map to obtain target features recognizable by the parking controller. The simulated open parking lot map is a custom-built simulated open parking lot map, and the target features include road features and lane features.

[0061] Since the parking controller needs to access and use the open parking lot map, vehicle position, speed, and other data to complete valet parking. However, the parking lot map database is inaccessible to applications other than the navigation system and is stored in the proprietary format of the in-vehicle multimedia navigation system. This application provides a solution to decompose and reorganize the simulated open parking lot map, thereby transforming it into a general open parking lot map recognizable by the parking controller. In the embodiments of the present invention, the target features include road features (such as slope, curvature) and lane features (such as the number of lanes, lane line type, line geometry, landmarks, road boundaries), etc.

[0062] In the specific implementation process, "reverse-transform the simulated open parking lot map" in step S20 can adopt the following steps, and the method flow chart is as Figure 2 shown:

[0063] S201. Analyze the simulated open parking lot map.

[0064] In the embodiments of this application, during the analysis, the simulated open parking lot map is offset, that is, the conversion of map data from the 84 coordinate system to the 02 coordinate system is completed. It should be noted that the offset is to deflect the map using a certain algorithm to prevent external attacks, and the related implementation is prior art and will not be elaborated here.

[0065] The simulated open parking lot map is built using the OpenDRIVE scenario editor. OpenDRIVE is a high-precision map format that has been widely used in the unmanned driving or map positioning industry. The parsing algorithm used to analyze the simulated open parking lot map is based on the ADASIS (Advanced Driver Assistance Systems interface Specification) protocol. The interface specification standard between the map data defined by the ADASIS protocol and the ADAS system is unified. It uses a formal language (Franca IDL) to define the relevant data structures.

[0066] S202. Extract the target features within the area where the target simulated vehicle is located from the analysis result of the simulated open parking lot map.

[0067] In the embodiments of the present application, target features in the simulated parking lot scene in front of the target simulated vehicle (i.e., the designated simulated vehicle) are initially extracted from the parsing result of the simulated open parking lot map, including road features (such as slope, curvature) and lane features (such as the number of lanes, lane line type, line geometry, landmarks, road boundaries), etc.

[0068] The ADASIS protocol focuses on autonomous driving functions such as HWP (High Way Pilot) and TJP (Traffic Jam Pilot). Based solely on this protocol, it cannot be well used for the valet parking function. By parsing the road features of the simulated open parking lot map based on this protocol, in order to ensure the comprehensiveness of the target features, some other algorithms are also needed to parse the lane features from the simulated open parking lot map. For the above-mentioned target features, optimization algorithms also need to be applied to complete two parts of work (i.e., Core One of the subsequent optimization algorithm and Core Two of the optimization algorithm) to make them applicable to the valet parking function system.

[0069] In the embodiments of the present application, Core One of the optimization algorithm is to solve the problem of inconsistency between the data in the simulated open parking lot map and the data defined by the ADASIS protocol after establishing a complete mapping relationship between the simulated development parking lot map and the ADASIS protocol broadcast data. That is, when mapping to the OpenDRIVE simulation map according to the map data that needs to be broadcast by the ADASIS protocol and finding that an attribute is missing or the data expression form is unavailable, processing and conversion are required. The simulated open parking lot map mainly includes three characteristics: road reference line, lanes, and road features. The features (i.e., elements) concerned by the valet parking function extracted by the optimization algorithm include the road center line, heading description, road height undulation, road cross slope, lane connection relationship and road connection relationship, road type definition method, lane widening and diversion and lane narrowing and merging description, object expression method of the tunnel, description of the road curve, and signal light expression, etc.

[0070] S203. Based on the ADASIS protocol, determine the unavailable features in the target features within the area where the target simulated vehicle is located and mapped to the simulated open parking lot map.

[0071] In the embodiments of the present application, the features that are unavailable due to data missing or different data expression forms when mapped to the simulated open parking lot map based on the ADASIS protocol (i.e., unavailable features) mainly include, but are not limited to: data format differences in guardrail attributes, road boundary expression methods, median strips, curvature description methods, traffic lights, zebra crossings, road reference lines, etc.

[0072] For example, the map data broadcast by the ADASIS protocol expresses cross and longitudinal slope data through the absolute height calculation of lane lines. However, the simulated open parking lot map can only express the road surface height and the relative height between the lane lines and the road surface. Therefore, the absolute height data of the lane lines needs to be converted.

[0073] S204. Convert the unavailable features.

[0074] In the embodiments of the present application, all the conversion methods for all differences are not listed. For the conversion technologies adopted for all differences, the conversion methods will be different according to different types. Only the data missing and data differences proposed in this article are used to illustrate the conversion methods. Specifically:

[0075] In the embodiments of the present application, the data missing conversion is completed by adding road markings to the simulated open parking lot map. The lane confluence and divergence of the simulated open parking lot map are expressed in the form of a cubic polynomial. However, when there are changes in the number of lanes, road types, or road intersections, the Road needs to be interrupted and then mapped to the map data broadcast by ADSIS to make it have a unique lane ID (i.e., road marking), and it is 0, -1, -2, -3... without wrong order, disorder, or repetition, so as to supplement the data missing in the conversion process.

[0076] Furthermore, in the embodiments of the present application, the data difference conversion is completed by road geometry decomposition. For the conversion of the definition methods of attributes such as LaneCurvature and Lane Slope in the road reference line, in OpenDRIVE, a mathematical formula (decompose the road geometry of the road reference line in OpenDRIVE into straight lines, spirals, arcs, cubic polynomials, and cubic polynomial parametric equations) is used to give the definition of a section of road. However, the positions, curvatures, etc. of the points on the road center line defined by the map data broadcast according to the ADASIS protocol are not in the same coordinate system. To complete the mapping relationship, the optimization algorithm first defines the simulated initial position in the simulated parking lot map scene as the starting point, establishes a unified coordinate system, and then decomposes the road geometry of the simulated open parking lot map into straight lines, arcs, and cubic polynomials to obtain the attributes and coordinates of each point, and maps to obtain the positions, curvatures, etc. of the points on the road center line of the broadcast map, so as to solve the problem of inconsistent data between the two.

[0077] On this basis, to further enrich the target features, the step of "performing reverse conversion on the simulated open parking lot map" in step S20 further includes the following steps:

[0078] Screen the unique features of the open parking lot from the target features.

[0079] In the embodiments of the present application, the second core of the optimization algorithm is to screen and obtain the unique features of the open parking lot through a large number of real open parking lot map materials. The unique features include but are not limited to:

[0080] 1) Parking space features. Specifically, parking space location information, parking space type (horizontal, vertical, diagonal), horizontal parking space length, vertical parking space width, and diagonal parking space width.

[0081] 2) Parking space line features. Specifically, parking space line specifications (the width of the parking space line itself, the distance from the line edge to the inside, the distance from the line edge to the center point of the parking space line, and the wear rate of the parking space line), the width of the parking space line (the distance between the center points of two lines), and the angle of the parking space line (the angle between the long side of the parking space line and the side edge of the vehicle).

[0082] 3) Parking space line corner point features. Specifically, the type of parking space line corner points (such as T-shaped corner points, L-shaped corner points, and mixed T-shaped and L-shaped corner points).

[0083] Summarize the above unique features as a supplement to the ADASIS protocol map data to obtain a general open parking lot map, including but not limited to information such as possible routes, vehicle positions, alternative routes, and route trajectories (road geometry, maximum speed, intersections), etc. The detailed routes can be accurately transmitted to information such as each lane, road surface markings, and parking space position types. Thus, the open parking lot map data required for the parking controller to complete valet parking can be obtained.

[0084] S30. Based on the positioning position of the target simulation vehicle, broadcast the target features within the area where the target simulation vehicle is located to the parking controller and perform real-time rendering and reconstruction of the broadcast target features within the area where the target simulation vehicle is located to obtain a general open parking lot map within the area where the target simulation vehicle is located, and download and update the general open parking lot map to the parking controller.

[0085] In the embodiment of the present application, the target features within the area where the target simulation vehicle is located and the positioning position of the target simulation vehicle are broadcast and transmitted in the ADASIS protocol format on the Ethernet / CAN through the vehicle bus. In addition, real-time rendering and reconstruction are performed on the target features within the area where the target simulation vehicle is located to obtain the corresponding general open parking lot map, so as to provide in advance for the parking controller an open parking lot map and parking space features (such as road / lane geometry, slope, lane markings, parking space information, etc.) based on the position in front of the vehicle for the parking controller to use for path planning.

[0086] In the embodiments of the present application, a sensor fusion positioning method is adopted to obtain the positioning position of the target simulated vehicle, that is, to perform fusion positioning on the output information of multiple simulation sensors. Specifically: The perception positioning system is arranged in the Linux real-time system of the host. Its simulation sensors are all built using models. The sensor simulation hardware relies on the real-time system of the host, and the software relies on the sensor model of VTD for parameterization. The specific simulation method can refer to the prior art, mainly including a simulated 360-degree panoramic camera, a front-view camera, a simulated millimeter-wave radar, a simulated ultrasonic radar, and a simulated lidar. Then, based on the simulated parking lot map data, the recognition results of feature points by the vehicle simulation vision module (360-degree panoramic camera and front-view system), the vehicle simulation GNSS (Global Navigation Satellite System) module, and the vehicle simulation gyroscope module sensors are fused, combined with vehicle state information such as wheel speed, to provide the output information of multiple simulation sensors required by the parking controller. After performing multi-sensor fusion positioning, the positioning position of the vehicle itself (i.e., the target simulated vehicle) in the open parking lot map is obtained, providing a basis for broadcasting within a specified kilometer range of the open parking lot map, and at the same time assisting the parking controller to complete functions such as detecting parking spaces, obstacle avoidance, and path replanning.

[0087] Thus, after obtaining the positioning position of the target simulated vehicle, the area within a specified kilometer range centered on this positioning position in the general open parking lot map can be used as the area where the target simulated vehicle is located. It should be noted that the GNSS signal used to obtain the positioning position of the vehicle itself (i.e., the target simulated vehicle) in the general open parking lot map comes from a simulated TBOX (the simulated TBOX is arranged in the Linux real-time system of the host and is built using a model. The specific simulation method can refer to the prior art). The GNSS information required for positioning directly adopts the method of simulating through the in-vehicle Ethernet (SOMEIP) bus. The GNSS simulation value is output through the scenario simulation system software (VTD, VIRES Virtual Test Drive), and the GNSS Ethernet signal is sent to the simulation central gateway through the in-vehicle Ethernet simulation board (the simulation central gateway is arranged in the Linux real-time system of the host and is built using a model. The specific simulation method can refer to the prior art). In the present application, the scenario reversal program receives the forwarding of the simulation central gateway. The parking controller receives the GNSS signal sent by the simulated TBOX and the simulated vehicle gyroscope signal sent by the scenario reversal program through the in-vehicle Ethernet and the bus gateway, and obtains the open parking lot map within a certain kilometer range in the coordinate system.

[0088] It should also be noted that the target features within the area where the target simulated vehicle is located mainly include, but are not limited to:

[0089] 1) Position features: path number, offset, speed, relative direction to the road, current lane, confidence level, and time relative to the previous GPS (Global Positioning System) information.

[0090] 2) Segment features: path number, highway class, type (main road, roundabout, parking lot), road composition (highway, single or double lane), speed limit, number of lanes, direction, (tunnel, bridge, fork, emergency lane, calculated path, service area, and complex intersection signs).

[0091] 3) STUB (corner points between similar Segments) features: path number, sub-path number, turning angle (angle with the next road segment), probability of being an intersection, road type and composition, number of lanes in both forward and reverse directions, turning point (turning onto another road), and whether it is a complex intersection.

[0092] 4) Profile features: path number, profile type, profile sequence points, curvature. Metadata: country code, region code, driving position, speed unit, and other information.

[0093] 5) Parking features: parking position information, parking type (horizontal, vertical, angled), horizontal parking length, vertical parking width, angled parking width.

[0094] 6) Parking line features: parking space line specifications (width of the parking line itself, distance from the line edge to the inside, distance from the line edge to the center point of the parking line, wear rate of the parking line), width of the parking line (distance between the center points of two lines), angle of the parking line (angle between the long side of the parking line and the side edge of the vehicle).

[0095] 7) Parking line corner point features: such as T-shaped corner points, L-shaped corner points, and mixed T-shaped and L-shaped corner points.

[0096] Furthermore, in the embodiments of the present application, the purpose of reconstructing the open parking lot map scenario is to build an OpenDRIVE simulation parking lot map scenario based on a scenario simulation software (such as VTD). After the vehicle (i.e., the target simulation vehicle) is positioned, the target features within the area where the target simulation vehicle is located are broadcasted to real-time draw and reconstruct the open parking lot map within a specified number of kilometers (i.e., the general open parking lot map), so as to make the data in the general open parking lot map consistent with that in the simulated open parking lot scenario. Therefore, the core of the drawing and reconstruction algorithm is to reorganize the map scenario features within the specified number of kilometers broadcasted into the following types of data: intersections, lane lines, geometric parameters at the near end of the line, geometric parameters at the far end of the line, traffic signs, traffic lights, ProfileType, parking information, and other information.

[0097] In practical applications, furthermore, through graphical display in the host computer software / mobile phone software (the software implementation method can refer to the existing technology), users can monitor through the graphical interface of the reconstructed map and perform operations such as clicking to set the target parking space and specifying the pick-up point; or, completely by the parking controller receiving the broadcast map data, the path planning information after map reconstruction, and the parking space information in the parking lot through Ethernet, combined with real-time scanning of the parking spaces and parked vehicles by sensors such as 360 panoramic cameras and ultrasonic radars, to complete the optimal path planning and the selection of the nearest parking space.

[0098] Thus, the parking controller does not need to pre-download the parking lot map and can complete the real-time reconstruction of the open parking lot map within a specified kilometer range in an open parking lot where the parking lot map is previously unknown. At the same time, the parking controller receives and fuses various sensor information output by the positioning, and through fusion and decision-making, finally determines the nearest optimal parking space to complete the path planning, issues control instructions to control the movement of the simulation vehicle, and realizes autonomous parking in and out without a driver in the open parking lot, realizing the low-speed L4-level valet parking process.

[0099] In summary, on the one hand, this application obtains the ADASIS protocol data by parsing the map data broadcast through the in-vehicle Ethernet (SOMEIP) protocol, and at the same time adds the unique features of the open parking lot screened from the target features, and gives them to the parking controller through the methods of broadcast and reconstruction. At the same time, it also receives the positioning position obtained by the sensor fusion positioning implemented based on the hardware-in-the-loop system, reads the recorded simulation GNSS and simulation vehicle gyroscope data of the simulation, performs fusion positioning, and obtains the open parking lot map data within the specified kilometer range of the broadcast, including the connection information between the map roads, the lane attributes of a specific section of a specific road, lane lines, parking space information, attribute points, and the latitude, longitude, slope, curvature, etc. data of each point, and the parking space information data, reconstructs the parking lot map within the specified kilometer range, and draws and visualizes the reconstructed parking lot map in the host computer software / mobile phone software for the path planning of the valet parking controller.

[0100] Therefore, the valet parking simulation method based on the open parking lot provided by this application uses the simulated open parking lot map as the map data source. After vehicle positioning, within the specified kilometer range, it completes the data broadcast of the target features on the CAN / Ethernet and reconstructs the general open parking lot map within the specified kilometer range, solving the problem of single simulation scenarios based on the real parking lot map scenario. At the same time, it also solves the limitation that most L3+ valet parking simulation systems need to pre-download the closed parking lot map in the parking controller and cannot complete the valet parking simulation in an open parking lot where the parking lot map is unknown, providing a solution for the simulation verification of the valet parking algorithm of the L4-level valet parking system in the open parking lot.

[0101] Based on the valet parking simulation method provided in the above embodiments, the embodiments of the present application correspondingly provide a device for executing the above valet parking simulation method. The structural schematic diagram of the device is as shown in Figure 3 shown, including:

[0102] A simulation map acquisition module 10, configured to acquire a simulation open parking lot map;

[0103] A general map conversion module 20, configured to perform reverse conversion on the simulation open parking lot map to obtain target features that can be recognized by the parking controller; wherein, the simulation open parking lot map is a custom-built simulation open parking lot map, and the target features include road features and lane features;

[0104] A broadcast and download module 30, configured to, based on the positioning position of the target simulated vehicle, broadcast the target features within the area where the target simulated vehicle is located to the parking controller and perform real-time drawing and reconstruction of the broadcast target features within the area where the target simulated vehicle is located, obtain a general open parking lot map within the area where the target simulated vehicle is located, and download and update the general open parking lot map to the parking controller.

[0105] Optionally, the general map conversion module 20 is specifically configured to:

[0106] Parse the simulation open parking lot map; extract the target features within the area where the target simulated vehicle is located from the parsing result of the simulation open parking lot map; determine the unavailable features mapped to the simulation open parking lot map among the target features within the area where the target simulated vehicle is located based on the ADASIS protocol; and convert the unavailable features.

[0107] Optionally, the general map conversion module 20 for converting the unavailable features is specifically configured to:

[0108] Perform data missing conversion / data difference conversion on the unavailable features.

[0109] Optionally, the general map conversion module 20 is further configured to:

[0110] Screen the unique features of the open parking lot from the target features.

[0111] Optionally, the unique features include:

[0112] Parking space features, parking space line features, and parking space line corner features.

[0113] It should be noted that for the detailed functions of each functional module in the embodiments of the present application, reference can be made to the corresponding disclosed parts in the above method embodiments, which will not be elaborated herein.

[0114] The above has introduced in detail a valet parking simulation method and device provided by the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

[0115] It should be noted that the various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0116] It should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes the inherent elements thereof, or further includes the inherent elements for these process, methods, articles or devices. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article or device including the said element.

[0117] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A valet parking simulation method, characterized in that The method includes: Obtaining a simulation open parking lot map; Parsing the simulation open parking lot map, and extracting target features within the area where the target simulated vehicle is located from the parsing result of the simulation open parking lot map; Based on the ADASIS protocol, determining features that are mapped to the simulation open parking lot map and are unavailable due to data missing or differences in data representation forms among the target features within the area where the target simulated vehicle is located; For the unavailable features, completing data missing transformation by adding road markings to the simulation open parking lot map, or completing data difference transformation by road geometry decomposition, to obtain target features that can be recognized by the parking controller; wherein, the simulation open parking lot map is a custom-built simulation open parking lot map, and the target features that can be recognized by the parking controller include road features and lane features; Based on the positioning location of the target simulated vehicle, broadcasting the target features within the area where the target simulated vehicle is located to the parking controller and performing real-time drawing and reconstruction of the broadcast target features within the area where the target simulated vehicle is located, to obtain a general open parking lot map within the area where the target simulated vehicle is located, and downloading and updating the general open parking lot map to the parking controller.

2. The method according to claim 1, characterized in that, It further includes: Screening out the unique features of the open parking lot from the target features within the area where the target simulated vehicle is located.

3. The method according to claim 2, wherein The unique features include: Parking space features, parking space line features, and parking space line corner point features.

4. A valet parking simulation device, characterized in that, The device includes: A simulation map acquisition module, configured to obtain a simulation open parking lot map; A general map conversion module, configured to parse the simulation open parking lot map, and extract target features within the area where the target simulated vehicle is located from the parsing result of the simulation open parking lot map; based on the ADASIS protocol, determining features that are mapped to the simulation open parking lot map and are unavailable due to data missing or differences in data representation forms among the target features within the area where the target simulated vehicle is located; for the unavailable features, completing data missing transformation by adding road markings to the simulation open parking lot map, or completing data difference transformation by road geometry decomposition, to obtain target features that can be recognized by the parking controller; wherein, the simulation open parking lot map is a custom-built simulation open parking lot map, and the target features that can be recognized by the parking controller include road features and lane features; A broadcast and download module, configured to broadcast the target features within the area where the target simulated vehicle is located to the parking controller based on the positioning location of the target simulated vehicle and perform real-time drawing and reconstruction of the broadcast target features within the area where the target simulated vehicle is located, to obtain a general open parking lot map within the area where the target simulated vehicle is located, and download and update the general open parking lot map to the parking controller.

5. The device according to claim 4, further characterized in that, The general map conversion module is further configured to: Screen out the unique features of the open parking lot from the target features within the area where the target simulated vehicle is located.

6. The device according to claim 5, characterized in that, The unique features include: Parking space features, parking space line features, and parking space line corner point features.

Images