An adaptive intelligent driving function simulation scene generalization method, system and application
By generating adaptive intelligent driving simulation scenarios through mapping tables and weighted Gaussian distributions, and combining multi-dimensional evaluation and closed-loop feedback, the adaptability and coverage of existing simulation scenario generalization methods are solved, thus achieving efficient simulation testing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI JIANGHUAI AUTOMOBILE GRP CORP LTD
- Filing Date
- 2026-03-09
- Publication Date
- 2026-06-09
Smart Images
Figure CN122174480A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of intelligent driving technology, and in particular to a simulation scenario generalization method, system and application for adaptive intelligent driving function. Background Technology
[0002] The research and development and large-scale mass production of intelligent driving technology cannot be separated from high-coverage, high-quality simulation test scenarios. However, traditional methods rely on the sampling and permutation of scenario parameter probability distributions, which have problems such as weak functional targeting and low matching degree between generated scenarios and target functions. This results in insufficient effectiveness of simulation testing and cannot efficiently meet the coverage needs of "long-tail scenarios".
[0003] Currently, there are various generalization techniques for intelligent driving simulation scenarios. For example, patent application CN117591405A discloses a scenario generalization method, device, electronic device, and medium for autonomous driving. This method includes: selecting multiple target scenario parameters from scenario parameters in a scenario file to be generalized; determining the distribution type of each target scenario parameter, including deterministic and probabilistic distributions; determining the parameter value of each target scenario parameter based on its distribution type; selecting multiple second sample values from multiple second parameter values according to weight values, and directly using the first parameter value as the first sample value; and determining multiple generalized scenario files based on the first sample value, the multiple second sample values, and the scenario file to be generalized. By employing the aforementioned scenario generalization method, device, electronic device, and medium for autonomous driving, the problem of low matching between generalized scenario files and simulation testing requirements is solved. However, this method uses existing scenario files as the scenario files to be generalized, and by extracting scenario parameters and determining the parameter probability distribution, it generalizes a series of new scenario files. Therefore, this method is based on existing scene files, and the number of targets and target logical behaviors in the generalized scene are strongly correlated with the basic scene files, which cannot cover a large number of complex logical scenes.
[0004] For example, patent application CN119538518A discloses a method and system for generalizing evaluation of autonomous driving simulation test scenarios. The method includes: acquiring and preprocessing real vehicle operation data to obtain basic scenario generalization data; fitting the basic scenario generalization data to a Gaussian mixture model to obtain the joint probability density function between scenario parameters; sampling generalized scenario parameters using the Hamiltonian Monte Carlo method based on the joint probability density function between scenario parameters to generate a generalized scenario parameter set; evaluating the generalized scenario based on the basic scenario generalization data and the generalized scenario parameter set, and obtaining the coverage of the generalized scenario parameter set by considering the complexity of the generalized scenario; the initialization parameters of the Gaussian mixture model are generated using the K-means++ algorithm. This method can improve the convergence speed and accuracy of generalized scenario construction and achieve effective evaluation of generalized scenario coverage; it solves the problem in existing technologies that cannot achieve comprehensive and effective evaluation of generalized scenarios. However, this method obtains the probability density function from preprocessed real vehicle data, then samples generalized scenario parameters to generate a scenario parameter set, and performs scenario generalization and evaluation based on the scenario parameter set. This method first relies on pre-processed real vehicle data, and the collection and processing of real vehicle data require a lot of equipment and manpower costs. In addition, real vehicle data is limited by the collection conditions and cannot fully cover a large number of "long tail" scenarios. The diversity of scenario types is insufficient and cannot effectively meet the simulation test scenario requirements of all intelligent driving functions.
[0005] Therefore, an automatic generalization method for simulation scenarios that can be adaptively combined with intelligent driving functions is needed to solve the shortcomings of existing technologies such as poor scenario adaptability, low testing efficiency, and insufficient coverage of extreme working conditions, so as to provide an efficient and accurate simulation environment for the research and development and verification of intelligent driving functions. Summary of the Invention
[0006] In view of the shortcomings of the prior art, the present invention provides a simulation scene generalization method, system and application for adaptive intelligent driving function, which can solve the problem that the existing simulation scene generalization method cannot adaptively generate scenes for intelligent driving function.
[0007] To achieve the above and related objectives, the present invention adopts the following technical solution:
[0008] The first aspect of this invention provides a simulation scenario generalization method for adaptive intelligent driving functions, comprising the following steps:
[0009] Step S100: Based on the input intelligent driving function type, query the preset mapping table to obtain the initial dataset of scene dimension parameters; wherein, the mapping table stores the correspondence between intelligent driving function types and scene dimension parameters including at least map type, entity type and quantity, and entity behavior;
[0010] Step S200: Using the initial dataset and the first feedback information as input, a candidate scene parameter set for the current iteration is generated through a weighted Gaussian distribution; where,
[0011] In the first iteration, the first feedback information is the initial feedback information generated based on preset rules;
[0012] In non-first iterations, the first feedback information is the updated feedback information dynamically adjusted based on the simulation results of the previous iteration.
[0013] Step S300: Perform a multi-dimensional scenario effectiveness evaluation on the candidate scenario parameter set of the current iteration round, and select effective scenario parameters;
[0014] Step S400: Generate an executable scenario file based on the effective scenario parameters, and perform intelligent driving function simulation to obtain the simulation results of the current round;
[0015] Determine whether the preset iteration termination condition is met. If it is met, the process ends. If it is not met, generate update feedback information for the next iteration based on the simulation results of the current round, and return to step S200 to proceed with the next iteration.
[0016] Furthermore, in step S200, the initial feedback information includes preset scene parameter weight values and the initial scene parameter expectation domain determined based on the full parameter range of the initial dataset.
[0017] Furthermore, in step S200, the updated feedback information includes the adjusted scene parameter weight values and the expected domain of the next round of scene parameters determined based on the parameter distribution of the failed scenes in the previous round of simulation results.
[0018] Furthermore, in step S300, the multi-dimensional scenario effectiveness includes regulatory compliance, dynamic accuracy, and functional compliance.
[0019] Furthermore, in step S100, the initial dataset includes at least a map type dataset, an entity type and quantity dataset, and an entity behavior dataset. The entity behavior dataset is obtained by querying the mapping table based on the intelligent driving function type and the entity type and quantity dataset.
[0020] Furthermore, functional compliance is used to verify whether the scenarios defined by the candidate scenario parameter set in the current iteration round contain the core entities and behavioral elements required by the intelligent driving function type to trigger and test the function.
[0021] Furthermore, in step S400, the iteration termination conditions include all key test scenario types corresponding to the intelligent driving function types input in the mapping table have been covered, and scenario type diversity saturation has been achieved.
[0022] A second aspect of the present invention provides a simulation scenario generalization system for adaptive intelligent driving functions, comprising:
[0023] The data acquisition module is used to query a pre-set mapping table based on the input intelligent driving function type to obtain the initial dataset of scene dimension parameters. The mapping table stores the correspondence between intelligent driving function types and scene dimension parameters, including at least map type, entity type and quantity, and entity behavior.
[0024] The parameter generalization module takes the initial dataset and the first feedback information as input and generates a candidate scene parameter set for the current iteration using a weighted Gaussian distribution; among which,
[0025] In the first iteration, the first feedback information is the initial feedback information generated based on preset rules;
[0026] In non-first iterations, the first feedback information is the updated feedback information dynamically adjusted based on the simulation results of the previous iteration.
[0027] The generalization evaluation module is used to evaluate the effectiveness of the candidate scenario parameter set in the current iteration in multiple dimensions and select the effective scenario parameters.
[0028] The scenario simulation module is used to generate executable scenario files based on valid scenario parameters, perform intelligent driving function simulation, and obtain the simulation results of the current round.
[0029] Determine whether the preset iteration termination condition is met. If it is met, the process ends. If not, generate update feedback information for the next iteration based on the simulation results of the current round, and call the parameter generalization module to perform the next iteration.
[0030] A third aspect of the present invention provides a computer-readable storage medium storing computer-readable instructions, which, when executed by a computer's processor, cause the computer to perform the above-mentioned simulation scenario generalization method for adaptive intelligent driving functions.
[0031] A fourth aspect of the present invention provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the simulation scenario generalization method for the above-mentioned adaptive intelligent driving function.
[0032] The beneficial technical effects of this invention are as follows:
[0033] This invention maps intelligent driving function types to specific scene dimension parameters through a pre-defined mapping table, fundamentally ensuring a strong correlation between the generated simulation scenarios and the functions under test. This allows the invention to generalize matching simulation scenarios based on the intelligent driving function type, improving simulation effectiveness and reducing scene management costs. Furthermore, this invention dynamically adjusts the generated update feedback information based on the simulation results of the previous iteration, especially failed simulation results. Through closed-loop iteration, it continuously corrects parameter weights and the expected domain, enabling scene generation to focus on high-value or high-risk areas, thereby effectively improving the quality and efficiency of generalization.
[0034] This invention employs a weighted Gaussian distribution as the generalization algorithm. By centrally sampling high-weight parameters and expanding the exploration of low-weight parameters in the parameter space, a balance can be achieved between scenario diversity and test effectiveness. Simultaneously, by combining multi-dimensional evaluations of regulatory compliance, dynamic accuracy, and functional compliance, a large number of invalid or non-compliant scenarios are filtered out during the scenario generation stage, thereby ensuring the usability of the output simulation scenarios.
[0035] In summary, this invention, through a collaborative mechanism of mapping table-oriented guidance, weighted Gaussian distribution balance generation, multi-dimensional evaluation filtering, and closed-loop feedback of simulation results, can systematically solve the challenges of the relevance, diversity, effectiveness, and coverage of simulation scenario generalization in intelligent driving simulation testing.
[0036] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description
[0037] The accompanying drawings, incorporated in and forming part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without inventive effort. In the drawings:
[0038] Figure 1 A flowchart of the simulation scenario generalization method for the adaptive intelligent driving function of this application;
[0039] Figure 2 This is a framework diagram of the simulation scenario generalization system for the adaptive intelligent driving function of this application.
[0040] Figure 3 A flowchart of another exemplary system of this application;
[0041] Figure 4 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. Detailed Implementation
[0042] Unless otherwise defined, all technical and / or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should be understood that certain features of the invention (described in the context of separate embodiments for clarity) may also be provided in a single embodiment. Conversely, multiple features of the invention (described in the context of a single embodiment for brevity) may also be provided separately or in any suitable combination or, where appropriate, in any other described embodiment of the invention. Certain features described in the context of various embodiments will not be considered essential features of those embodiments unless the embodiment is inoperable without those elements. The invention is further illustrated below by specific examples; however, it should be noted that the specific process conditions and results described in the embodiments of the invention are merely illustrative and should not be construed as limiting the scope of protection of the invention. All equivalent changes or modifications made in accordance with the spirit and essence of the invention should be covered within the scope of protection of the invention.
[0043] Please see Figure 1 The flowchart of the simulation scenario generalization method for the adaptive intelligent driving function of this application is described in detail below:
[0044] Step S100: Based on the input intelligent driving function type, query the preset mapping table to obtain the initial dataset of scene dimension parameters; wherein, the mapping table stores the correspondence between intelligent driving function types and scene dimension parameters including at least map type, entity type and quantity, and entity behavior.
[0045] Specifically, the initial dataset of this application includes at least a map type dataset, an entity type and quantity dataset, and an entity behavior dataset. The entity behavior dataset is obtained by querying the mapping table based on the intelligent driving function type and the entity type and quantity dataset. The map type dataset includes, but is not limited to, a set of map types such as straight roads, curves, ramps, and intersections; the entity type and quantity dataset includes, but is not limited to, entity types such as vehicles and pedestrians, and a set of the quantity of each type of entity in the scene, such as 1 main vehicle, 3 other vehicles, and 2 pedestrians; the entity behavior dataset is a set of all expected entity behaviors queried from the mapping table, such as acceleration, deceleration, lane changing, and following other vehicles.
[0046] More specifically, the mapping table includes at least three types of mapping relationships:
[0047] 1) Intelligent driving function type - map type: For example, adaptive cruise function corresponds to straight road and curve map; automatic parking function corresponds to parking lot and roadside parking space map; intersection turning function corresponds to intersection and ramp map.
[0048] 2) Intelligent driving function type - entity type and quantity: For example, adaptive cruise corresponds to 1 main vehicle and several side vehicles; intersection turning corresponds to 1 to 5 pedestrians, etc.
[0049] 3) Intelligent driving function type - entity type and quantity - entity behavior: for example, adaptive cruise control corresponds to acceleration, deceleration, maintaining distance, lane change, etc.
[0050] Step S200: Using the initial dataset and the first feedback information as input, a candidate scenario parameter set for the current iteration is generated through a weighted Gaussian distribution; wherein, in the first iteration, the first feedback information is the initial feedback information generated based on preset rules; in subsequent iterations, the first feedback information is the updated feedback information dynamically adjusted based on the simulation results of the previous iteration.
[0051] Specifically, the initial feedback information in this application includes preset scenario parameter weight values and an initial scenario parameter expectation domain determined based on the full parameter range of the initial dataset. The preset scenario parameter weight values reflect the criticality of each scenario dimension parameter to the intelligent driving function test. These weight values guide the weighted Gaussian distribution to prioritize sampling high-weight parameters during the initial generalization, ensuring that the initially generated candidate scenarios focus on core testing requirements and avoiding low effective simulation scenario generation rates due to excessive randomness. The preset scenario parameter weight values in this application can be set based on prior knowledge such as industry standards, functional safety requirements, and historical test data statistics, and are dynamically adjusted in subsequent iterations based on simulation results. For example, high-weight parameters, such as 0.8 to 1.0, are scenario parameters strongly correlated with intelligent driving functions, mandated by regulations, or prone to failure. For instance, in adaptive cruise control, the weight value for the vehicle in front's deceleration behavior is 0.9, and the weight value for the following distance is 0.8. Medium-weighted parameters, such as 0.4 to 0.7, are auxiliary parameters for testing intelligent driving functions but are not core parameters. For example, in lane keeping function, the weight value for curve curvature is 0.6 and the weight value for lane line clarity is 0.5. Low-weighted parameters, such as 0.1 to 0.3, have little impact on testing intelligent driving functions and are mainly used to enrich the diversity of scenarios. For example, in parking lot scenarios, the weight value for the color of surrounding vehicles is 0.1.
[0052] More specifically, the expected domain of the initial scene parameters in this application is the allowable sampling range of each scene dimension parameter. During the first iteration, since there are no historical simulation results to constrain it, the expected domain can directly take the full parameter range of the initial dataset to ensure the breadth of the initial exploration. For example, the expected domain for map types: if the initial dataset includes {straight roads, curves, ramps, intersections}, then during the first iteration, the expected domain for map types will be these four categories, without additional restrictions.
[0053] Specifically, the updated feedback information includes adjusted scene parameter weights and the expected domain of scene parameters for the next round, determined based on the parameter distribution of failed scenarios in the previous simulation results. The adjusted scene parameter weights refer to the dynamic redistribution of the importance or priority of each scene dimension parameter in the previous iteration. This adjustment directly affects the sampling probability when generating new scenarios using a weighted Gaussian distribution. For example, if a parameter dimension is associated with a large number of irrelevant or inefficient scenarios, its weight can be lowered; if a parameter dimension generated high-value scenarios in the previous round, its weight can be increased. More specifically, the expected domain of scene parameters for the next round refers to the key sampling range and boundaries for scene parameter values allowed in the next iteration. For example, statistical analysis of parameters from all marked failed scenarios in the previous simulation can identify which parameters have abnormally high frequencies in these failed scenarios. If a parameter is found to frequently cause failures, the expected domain for the next round will be strategically narrowed towards the vicinity of that parameter to reduce the expected domain. Furthermore, the expected domain in this application can be a range of parameter values or a combination of parameters.
[0054] Specifically, this application uses a weighted Gaussian distribution to automatically focus on high-value areas, such as failed scenarios and uncovered scenarios, while retaining a certain degree of randomness to explore new scenarios, thus achieving adaptive generalization and precise focusing.
[0055] Step S300: Perform a multi-dimensional scenario effectiveness evaluation on the candidate scenario parameter set of the current iteration round, and select effective scenario parameters.
[0056] Specifically, the multi-dimensional scenario validity of this application includes regulatory compliance, dynamic accuracy, and functional compliance. Regulatory compliance verifies whether the simulation scenario defined by the candidate scenario parameter set conforms to real-world road traffic regulations, standards, and driving rules, thereby ensuring the simulation scenario possesses realistic legality and compliance. Dynamic accuracy verifies whether the candidate scenario parameter set conforms to the physical motion laws of entities such as vehicles and pedestrians in the real world, thereby ensuring the simulation scenario possesses physical realism and providing reliable verification for the intelligent driving algorithm. Functional compliance verifies whether the scenario defined by the candidate scenario parameter set in the current iteration contains the core entities and behavioral elements required for triggering and testing the intelligent driving function type; the core entity elements are the traffic participants necessary for triggering and testing the function; the core behavioral elements are the key behaviors and states that can stimulate the typical operating conditions and boundary conditions of the intelligent driving function.
[0057] Step S400: Generate an executable scenario file based on valid scenario parameters, and perform intelligent driving function simulation to obtain the simulation results of the current round; determine whether the preset iteration termination condition is met. If it is met, the process ends; if it is not met, generate update feedback information for the next iteration round based on the simulation results of the current round, and return to step S200 to proceed with the next iteration.
[0058] Specifically, the scenario file of this application can be based on the commonly used openScenario standard or a custom standard; and the scenario file can simulate the input intelligent driving function through a professional simulation platform, and feed the simulation results back to step S200 for feedback information update. On the other hand, the simulation results can be used to assist in verifying the simulation performance of the current effective scenario parameters, and indirectly support the evaluation of effectiveness judgment.
[0059] Specifically, the iteration termination conditions in this application include the coverage of all key test scenario types corresponding to the intelligent driving function types input in the mapping table, and scenario type diversity saturation. In the mapping table of this application, each intelligent driving function type is associated with a set of key test scenario types. For example, automatic emergency braking is associated with a stationary vehicle in front, pedestrians crossing, and overtaking on a curve. When the simulation results contain all these key scenario types, the function coverage can be considered complete, thus ensuring that each intelligent driving function is fully tested in the simulation scenario generalization and combined with the long-tail scenario coverage target. Scenario type diversity saturation in this application means that during the iteration process, newly generated candidate scenarios no longer increase the diversity in type distribution, thereby avoiding invalid iterations, saving computational resources, and reducing scenario management costs.
[0060] Please see Figure 2 Here is a framework diagram of the simulation scenario generalization system 200 for the adaptive intelligent driving function of this application, including:
[0061] The data acquisition module 210 is used to query a preset mapping table based on the input intelligent driving function type to obtain the initial dataset of scene dimension parameters; wherein, the mapping table stores the correspondence between intelligent driving function types and scene dimension parameters including at least map type, entity type and quantity, and entity behavior;
[0062] The parameter generalization module 220 is used to take the initial dataset and the first feedback information as input, and generate a candidate scene parameter set for the current iteration round through a weighted Gaussian distribution; wherein,
[0063] In the first iteration, the first feedback information is the initial feedback information generated based on preset rules;
[0064] In non-first iterations, the first feedback information is the updated feedback information dynamically adjusted based on the simulation results of the previous iteration.
[0065] The generalization evaluation module 230 is used to evaluate the effectiveness of the candidate scenario parameter set in the current iteration in multiple dimensions and select effective scenario parameters.
[0066] The scenario simulation module 240 is used to generate an executable scenario file based on valid scenario parameters, and to perform intelligent driving function simulation to obtain the simulation results of the current round.
[0067] Determine whether the preset iteration termination condition is met. If it is met, the process ends. If not, generate update feedback information for the next iteration based on the simulation results of the current round, and call the parameter generalization module to perform the next iteration.
[0068] Specifically, in combination Figure 3 The data acquisition module 210 of this application includes a map sampling module, an entity type and quantity sampling module, and an entity behavior sampling module, which are used to acquire map type datasets, entity type and quantity datasets, and entity behavior datasets, respectively.
[0069] Specifically, in combination Figure 3 The system described in this application employs a closed-loop mechanism of input-acquisition-generalization-evaluation-generation-feedback. The simulation results of the current round are fed back to the generalization evaluation module, which generates updated feedback information based on this and inputs this updated feedback information to the parameter generalization module, driving the parameter generalization module to perform the next round of generalization generation, thereby achieving adaptive scene expansion.
[0070] It should be noted that the simulation scenario generalization system for adaptive intelligent driving functions provided in the above embodiments and the simulation scenario generalization method for adaptive intelligent driving functions provided in the above embodiments belong to the same concept. The specific methods of execution of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the simulation scenario generalization system for adaptive intelligent driving functions provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.
[0071] Embodiments of this application also provide a computer device, including: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, enable the computer device to implement the simulation scenario generalization method for adaptive intelligent driving functions provided in the above embodiments.
[0072] Figure 4 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. It should be noted that... Figure 4The computer system 400 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0073] like Figure 4 As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage section 408 into a random access memory (RAM) 403, such as performing the methods described in the above embodiments. Various programs and data required for system operation are also stored in the RAM 403. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404. The following components are connected to the I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (local area network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A driver 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as disks, optical discs, magneto-optical discs, semiconductor memories, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.
[0074] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer tool programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs various functions defined in the system of this application.
[0075] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, flash memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. Computer programs contained on computer-readable media can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.
[0076] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0077] The units described in the embodiments of this application can be implemented by tools or by hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the unit itself.
[0078] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the simulation scenario generalization method for the adaptive intelligent driving function as described above. This computer-readable storage medium may be included in the computer device described in the above embodiments, or it may exist independently and not assembled into the computer device.
[0079] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the simulation scenario generalization method for the adaptive intelligent driving function provided in the various embodiments above.
[0080] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.
Claims
1. A simulation scenario generalization method for adaptive intelligent driving functions, characterized in that, Includes the following steps: Step S100: Based on the input intelligent driving function type, query the preset mapping table to obtain the initial dataset of scene dimension parameters; wherein, the mapping table stores the correspondence between intelligent driving function types and scene dimension parameters including at least map type, entity type and quantity, and entity behavior; Step S200: Using the initial dataset and the first feedback information as input, a candidate scene parameter set for the current iteration is generated through a weighted Gaussian distribution; wherein, In the first iteration, the first feedback information is the initial feedback information generated based on preset rules; In non-first iterations, the first feedback information is an updated feedback information dynamically adjusted based on the simulation results of the previous iteration. Step S300: Perform a multi-dimensional scene validity evaluation on the candidate scene parameter set of the current iteration round, and select valid scene parameters; Step S400: Generate an executable scenario file based on the effective scenario parameters, and perform intelligent driving function simulation to obtain the simulation results of the current round; Determine whether the preset iteration termination condition is met. If it is met, the process ends. If it is not met, generate the update feedback information for the next iteration based on the simulation results of the current round, and return to step S200 to proceed with the next iteration.
2. The simulation scene generalization method according to claim 1, characterized in that, In step S200, the initial feedback information includes preset scene parameter weight values and an initial scene parameter expectation domain determined based on the full parameter range of the initial dataset.
3. The simulation scene generalization method according to claim 1, characterized in that, In step S200, the updated feedback information includes the adjusted scene parameter weight values and the expected domain of the next round of scene parameters determined based on the parameter distribution of the failed scenes in the previous round of simulation results.
4. The simulation scene generalization method according to claim 1, characterized in that, In step S300, the multi-dimensional scenario validity includes regulatory compliance, dynamic accuracy, and functional compliance.
5. The simulation scene generalization method according to claim 1, characterized in that, In step S100, the initial dataset includes at least a map type dataset, an entity type and quantity dataset, and an entity behavior dataset. The entity behavior dataset is obtained by querying the mapping table based on the intelligent driving function type and the entity type and quantity dataset.
6. The simulation scene generalization method according to claim 4, characterized in that, The functional compliance is used to verify whether the scenario defined by the candidate scenario parameter set of the current iteration contains the core entities and behavioral elements required by the intelligent driving function type for triggering and testing the function.
7. The simulation scene generalization method according to claim 1, characterized in that, In step S400, the iteration termination condition includes covering all key test scenario types corresponding to the intelligent driving function type input in the mapping table, and scenario type diversity saturation.
8. A simulation scenario generalization system for adaptive intelligent driving functions, characterized in that, include: The data acquisition module is used to query a pre-set mapping table based on the input intelligent driving function type to obtain an initial dataset of scene dimension parameters; wherein, the mapping table stores the correspondence between intelligent driving function types and scene dimension parameters including at least map type, entity type and quantity, and entity behavior; The parameter generalization module is used to take the initial dataset and the first feedback information as input, and generate a candidate scene parameter set for the current iteration round through a weighted Gaussian distribution; wherein, In the first iteration, the first feedback information is the initial feedback information generated based on preset rules; In non-first iterations, the first feedback information is an updated feedback information dynamically adjusted based on the simulation results of the previous iteration. The generalization evaluation module is used to perform multi-dimensional scenario effectiveness evaluation on the candidate scenario parameter set of the current iteration round and filter out effective scenario parameters; The scenario simulation module is used to generate an executable scenario file based on the effective scenario parameters, and to perform intelligent driving function simulation to obtain the simulation results of the current round. Determine whether the preset iteration termination condition is met. If it is met, the process ends. If it is not met, generate the update feedback information for the next iteration based on the simulation results of the current round, and call the parameter generalization module to perform the next iteration.
9. A computer-readable storage medium, characterized in that, It stores computer-readable instructions, which, when executed by the computer's processor, cause the computer to perform the simulation scenario generalization method for the adaptive intelligent driving function as described in any one of claims 1 to 7.
10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the simulation scenario generalization method for the adaptive intelligent driving function as described in any one of claims 1 to 7.