Automated driving interactive test scenario generation apparatus, method, device and medium

Abstract

The application relates to the technical field of automatic driving, in particular to an automatic driving interactive test scene generation device, method, equipment and medium, wherein the device comprises: an extender used for decomposing and supplementing a target natural language description to generate a plurality of scene descriptions; a weather simulator used for simulating a scene according to weather information to generate weather parameters; a self-vehicle locator used for searching a plurality of self-vehicle generated positions on a preset map according to a self-vehicle position; an adversarial vehicle locator used for searching a plurality of adversarial vehicle static positions on the preset map according to an adversarial vehicle relative position and the self-vehicle position; a behavior generator used for constructing a behavior topology network according to an interactive behavior library to generate a plurality of interactive dynamic behaviors; and a scene generation module used for constructing a plurality of comprehensive test scenes according to the related data obtained above. Thus, the problem that the matching degree of the interactivity of scene generation and user demand in the prior art is obviously defective is solved.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to an interactive test scenario generation device, method, equipment, and medium for autonomous driving based on a large language model. Background Technology

[0002] Currently, generating test and training scenarios for autonomous vehicles still faces several challenges. Scenarios collected from the real world often lack sufficient danger, with rarely occurring dangerous scenarios making up a very small percentage, thus failing to provide enough diversity for testing or training. To increase the proportion of dangerous scenarios in the overall test / training scenario library, existing methods mainly include selecting scenarios from real-world data and manually predefining adversarial vehicle models within scenarios. While these two methods can generate some scenarios, they have several drawbacks.

[0003] First, methods based on real-world data are limited by the availability and diversity of data, especially in acquiring information about security-critical scenarios. These scenarios are highly scarce and rarely occur in reality, making it difficult to achieve comprehensive coverage using real-world data. Furthermore, this approach requires extensive data collection and processing, which is costly from both economic and time perspectives.

[0004] Secondly, while methods based on adversarial vehicle models within manually predefined scenarios can create safety-critical scenarios and ensure the plausibility of vehicle behavior, these methods typically rely on manually set rules and scenarios, limiting the diversity and realism of the generated scenarios. Furthermore, these methods struggle to simulate complex and unpredictable driving behaviors, which is a key factor in the safety assessment of autonomous driving systems.

[0005] In recent years, methods based on Large Language Models (LLMs) have been proposed, utilizing natural language descriptions to generate scenes and leveraging the understanding capabilities of large language models to generate corresponding scene codes for playback in simulators. This approach can improve the flexibility and diversity of scene generation. However, existing methods still have shortcomings in terms of static scene diversity, the interactivity of background vehicle behavior, and integration with existing simulation platforms. For example, while 2D simulators like SUMO can quickly generate scenes, they lack the realism of 3D environments, while 3D simulators like CARLA require complex vehicle control logic. Furthermore, existing LLM methods often struggle to accurately control the behavior of background vehicles when processing natural language descriptions and are difficult to integrate with modern autonomous driving algorithms. The behavior of background vehicles often relies on predefined behavioral patterns rather than acting as intelligent, purpose-driven real vehicles. In addition, current scene generation tools based on large language models generally lack interactivity and dynamism; the behavior of background vehicles does not interact with their surrounding environment, resulting in scenes that are not realistic or challenging enough in simulation tests. Therefore, existing technologies have significant deficiencies in the interactivity and matching of scene generation with user needs. These shortcomings limit the training efficiency of autonomous driving systems and reduce their robustness to extreme driving environments. Therefore, the industry urgently needs a new solution capable of generating realistic, diverse, and interactive scenarios based on user needs. Summary of the Invention

[0006] This invention provides an interactive test scenario generation device, method, equipment, and medium for autonomous driving, to address the significant deficiencies in the interactivity and matching degree with user needs of existing technologies.

[0007] A first aspect of the present invention provides an interactive test scenario generation device for autonomous driving, comprising: an expander for decomposing and supplementing a target natural language description to generate multiple scenario descriptions, wherein the multiple scenario descriptions include weather information, vehicle location, relative location of adversary vehicles, and an interactive behavior library; a weather simulator for simulating scenarios based on the weather information to generate weather parameters that can be used for rendering a simulation environment; a vehicle locator for searching multiple vehicle generation locations on a preset autonomous driving simulator map based on the vehicle location; an adversary vehicle locator for searching multiple static locations of adversary vehicles on the preset autonomous driving simulator map based on the relative location of the adversary vehicles and the vehicle location; a behavior generator for constructing a behavior topology network based on the interactive behavior library and generating multiple interactive dynamic behaviors using the behavior topology network; and a scenario generation module for constructing multiple comprehensive test scenarios based on the weather parameters, the multiple vehicle generation locations, the multiple static locations of adversary vehicles, and the multiple interactive dynamic behaviors.

[0008] Optionally, the weather parameters include at least one of solar altitude angle, wind force, precipitation, fog concentration, and dust storm intensity.

[0009] Optionally, the vehicle locator includes:

[0010] The first construction unit is used to construct a location feasibility judgment function based on the vehicle's location; the first search unit is used to search for multiple vehicle generation locations that meet the preset location requirements based on the preset autonomous driving simulator map and using the location feasibility judgment function.

[0011] Optionally, the counter-vehicle locator includes:

[0012] The second construction unit is used to construct a surrounding vehicle position information function of the vehicle's position based on the relative positions of the opposing vehicles; the second search unit is used to input the vehicle's position into the surrounding vehicle position information function to search for the static positions of the multiple opposing vehicles in the preset autonomous driving simulator map.

[0013] Optionally, the behavior generator includes:

[0014] The third construction unit is used to select multiple discrete behaviors from the interactive behavior library and construct the behavior topology network based on the multiple discrete behaviors; the generation unit is used to generate interactive dynamic behaviors using the behavior topology network.

[0015] Optionally, the third construction unit includes a filtering subunit, a vehicle model selection subunit, a vehicle model configuration subunit, a behavior success / failure condition subunit, and a construction subunit, wherein,

[0016] The filtering subunit is used to filter multiple discrete behaviors that conform to the descriptions of the multiple scenarios in the interactive behavior library; the vehicle model selection subunit is used to select the corresponding autonomous driving model for the vehicle performing any discrete behavior; the vehicle model configuration subunit is used to interactively adjust the behavior parameters of the corresponding autonomous driving model according to each discrete behavior; the behavior success / failure condition subunit is used to define the motion state of each discrete behavior; the construction subunit is used to connect each discrete behavior sequentially or in parallel logical connection according to the motion state of each discrete behavior to form the behavior topology network.

[0017] A second aspect of the present invention provides a method for generating interactive test scenarios for autonomous driving, comprising the following steps: decomposing and supplementing the target natural language to generate multiple scenario descriptions, wherein the multiple scenario descriptions include weather information, vehicle location, relative positions of adversary vehicles, and an interactive behavior library; performing scenario simulation based on the weather information to generate weather parameters that can be used for rendering a simulation environment; searching for multiple vehicle generation locations on a preset autonomous driving simulator map based on the vehicle location; searching for multiple static positions of adversary vehicles on the preset autonomous driving simulator map based on the relative positions of the adversary vehicles and the vehicle location; constructing a behavior topology network based on the interactive behavior library, and generating multiple interactive dynamic behaviors using the behavior topology network; and constructing multiple comprehensive test scenarios based on the weather parameters, the multiple vehicle generation locations, the multiple static positions of the adversary vehicles, and the multiple interactive dynamic behaviors.

[0018] A third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for generating interactive test scenarios for autonomous driving as described in the above embodiments.

[0019] A fourth aspect of the present invention provides a computer program product, which, when executed by a processor, implements the above-described method for generating interactive test scenarios for autonomous driving.

[0020] A fifth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for generating interactive test scenarios for autonomous driving.

[0021] The interactive test scenario generation device, method, equipment, and medium for autonomous driving proposed in this invention can generate realistic and diverse 3D scenarios based on user natural language descriptions, including road maps, weather conditions, and background vehicle behavior. Through modular design and the collaborative work of multiple LLM agents, the efficiency and controllability of scenario generation are ensured, achieving significant progress in scenario realism and consistency between background vehicle behavior and semantic descriptions. It can integrate various non-learning or learning-based vehicle control algorithms to guarantee the rationality of vehicle behavior, ensuring that vehicle behavior is constrained not only by dynamics but also by the control model. The generated scenarios not only simulate complex traffic environments but also ensure that the behavior of background vehicles is physically feasible and conforms to real-world driving rules. Therefore, users can generate highly realistic and reliable test scenarios, thereby improving the performance of autonomous driving systems. It improves the training efficiency and safety of the system; it enables real-time interaction between the vehicle and the driver, and between vehicles, to obtain dynamic scenarios. Furthermore, it can respond with corresponding adversarial actions depending on the driver's behavior, resulting in highly realistic scenarios that can dynamically adjust vehicle behavior based on user needs. This allows the autonomous driving system to face more diverse and challenging scenarios during training, further enhancing its robustness and safety. Users can quickly generate test scenarios that meet actual needs, reducing reliance on real data, lowering testing costs, and improving the training efficiency and safety of the autonomous driving system. Through its unique modular design, integrated vehicle control algorithms, and real-time interaction capabilities, it solves several key problems in existing technologies, providing a more effective and flexible solution for the testing and training of autonomous vehicles.

[0022] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0023] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0024] Figure 1 This is a block diagram of an interactive test scenario generation device for autonomous driving provided in an embodiment of the present invention;

[0025] Figure 2 This is a detailed flowchart of an interactive test scenario generation device for autonomous driving provided in an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram illustrating the working principle of the behavior generator provided in an embodiment of the present invention;

[0027] Figure 4This is a schematic diagram of the behavioral topology network provided in an embodiment of the present invention;

[0028] Figure 5 This is a flowchart illustrating a method for generating an interactive test scenario for autonomous driving, as provided in an embodiment of the present invention.

[0029] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0030] Explanation of reference numerals in the attached figures:

[0031] 10-Autonomous driving interactive test scenario generation device, 101-Writer, 102-Weather simulator, 103-Vehicle locator, 104-Adversarial vehicle locator, 105-Behavior generator, 106-Scene generation module, 60-Electronic device, 601-Memory, 602-Processor and 603-Communication interface. Detailed Implementation

[0032] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0033] The following description, with reference to the accompanying drawings, describes an apparatus, method, device, and medium for generating interactive test scenarios for autonomous driving according to embodiments of the present invention.

[0034] Figure 1 This is a block diagram of an interactive test scenario generation device for autonomous driving provided in an embodiment of the present invention.

[0035] like Figure 1 As shown, the interactive autonomous driving test scenario generation device 10 includes: a writer 101, a weather simulator 102, a vehicle locator 103, an adversarial vehicle locator 104, a behavior generator 105, and a scenario generation module 106. The device proposed in this embodiment of the invention inputs the user's natural language description into the writer. The writer refines and segments the information contained in the description into descriptions tailored to different scenario content. These descriptions are then passed to all other large language model agents. The outputs of each agent are aggregated and stored in a scenario parameter file, and the scenario player performs scenario playback and evaluation.

[0036] The system comprises the following components: Expander 101 decomposes and supplements the target natural language description to generate multiple scene descriptions, including weather information, vehicle location, relative positions of adversary vehicles, and an interactive behavior library. Weather simulator 102 simulates scenes based on weather information to generate weather parameters suitable for rendering the simulation environment. Vehicle locator 103 searches for multiple vehicle generation locations on a preset autonomous driving simulator map based on the vehicle's location. Adversary vehicle locator 104 searches for multiple static adversary vehicle locations on the preset autonomous driving simulator map based on the relative positions of adversary vehicles and the vehicle's location. Behavior generator 105 constructs a behavior topology network based on the interactive behavior library and generates multiple interactive dynamic behaviors using this network. Scene generation module 106 constructs multiple comprehensive test scenarios based on weather parameters, multiple vehicle generation locations, multiple static positions of adversary vehicles, and multiple interactive dynamic behaviors.

[0037] In some embodiments, such as Figure 2 As shown, the interpreter 101 is responsible for interpreting the user's natural language description, breaking it down into multiple aspects, such as weather conditions, the vehicle's location, the relative position of opposing vehicles and their interactive behaviors, etc. If the user does not specify certain information, the interpreter will supplement it based on its understanding of the scene.

[0038] In some embodiments, the weather simulator 102 (Weather Report) is responsible for simulating weather conditions during the simulation process. This is crucial for training or testing involving perception. The weather simulator takes the weather description described in the enlarger as input and generates weather parameters that can be used for rendering the simulation environment, including information such as solar altitude angle, wind speed, precipitation, fog concentration, and dust storm intensity. Weather simulation will have a significant impact on vehicle perception in the simulation environment, enhancing the realism and diversity of the generated scenes.

[0039] In some embodiments, the vehicle locator 103 (Ego Locator) includes:

[0040] The first construction unit is used to construct a location feasibility judgment function based on the vehicle's location;

[0041] The first search unit is used to search for multiple vehicle-generated locations that meet the preset location requirements based on the preset autonomous driving simulator map and using a location feasibility judgment function.

[0042] Specifically, the vehicle locator searches for suitable vehicle generation points in the CARLA autonomous driving simulator map, such as "at an intersection," "in a roundabout," or "on a long straight road." The locator takes the vehicle location description from the encoder as input and outputs a feasibility judgment function for the generated vehicle location. Possible vehicle location information is input into this function; if the location matches the input semantic description, it returns true; otherwise, it returns false.

[0043] In some embodiments, the Adv Locator 104 includes:

[0044] The second construction unit is used to construct the vehicle's surrounding vehicle position information function based on the relative position of the opposing vehicle;

[0045] The second search unit is used to input the vehicle's position into the surrounding vehicle position information function in order to search for multiple static positions of adversarial vehicles in the preset autonomous driving simulator map.

[0046] Specifically, the adversarial vehicle locator is responsible for generating the static positions of adversarial vehicles, ensuring that these vehicles and the driver vehicle are placed compliantly on the road. This adversarial vehicle locator takes the relative positions of the adversarial vehicles from the encoder's output as input and outputs a function that searches for surrounding vehicle positions based on the driver vehicle's position. If the position exists, is located in the driving lane, and has no other obstacles, the function returns the adversarial vehicle position information; otherwise, it returns false. This function is combined with the feasibility judgment function for generating the driver vehicle's position generated by the driver vehicle locator, which searches the map for all possible driver and surrounding vehicle positions and stores them in the corresponding scene parameter file.

[0047] In some embodiments, the action generator 105 includes:

[0048] The third building unit is used to select multiple discrete behaviors from the interactive behavior library and construct a behavior topology network based on the multiple discrete behaviors.

[0049] The generation unit is used to generate interactive dynamic behaviors using behavioral topology networks.

[0050] Specifically, such as Figure 3 As shown, the behavior generator is a key component for dynamic interactivity in traffic scenarios, responsible for generating the dynamic behaviors of adversarial vehicles. This module selects appropriate discrete behaviors, configures these behaviors to meet practical needs, and connects them using parallel or sequential logic to construct an interactive behavior topology network. The behavior generator takes the adversarial vehicle behaviors described in the expander's output as input and outputs a behavior topology network in coded form.

[0051] In some embodiments, the third construction unit includes a filtering subunit, a vehicle model selection subunit, a vehicle model configuration subunit, a behavior success / failure condition subunit, and a construction subunit, wherein...

[0052] The filtering subunit is used to filter multiple discrete behaviors in the interaction behavior library that match multiple scenario descriptions;

[0053] The vehicle model selection subunit is used to select the corresponding autonomous driving model for a vehicle performing any discrete behavior;

[0054] The vehicle model configuration subunit is used to interactively adjust the behavior parameters of the corresponding autonomous driving model based on each discrete behavior.

[0055] The behavior success / failure condition sub-unit is used to define the motion state of each discrete behavior;

[0056] Sub-units are constructed to connect each discrete behavior sequentially or in parallel logical connection according to the motion state of each discrete behavior, so as to form a behavior topology network.

[0057] Specifically, the core of the behavior generator 105 lies in constructing a behavior topology web composed of discrete behaviors. The first step in the workflow of this behavior generator is the construction of discrete behaviors. Discrete behaviors, as the smallest unit of behavior that can be executed by a vehicle, include vehicle model selection subunits (Agent Selection), vehicle model configuration subunits (Agent Configuration), and behavior success / fail condition subunits (Behavior Success / Fail Condition). The vehicle model selection subunit selects a suitable autonomous driving model for the vehicle performing a specific discrete behavior. In this embodiment, multiple models are compatible, such as the CARLA built-in autonomous driving model, the ACC following model, and the PlanT agent based on imitation learning. The vehicle model configuration subunit, based on the selected vehicle model, interactively adjusts the behavior parameters (such as speed, acceleration, heading angle, etc.) input by the LLM model. For example, when configuring the "follow vehicle" behavior, parameters such as the following target, target speed, and following distance can be adjusted to achieve precise control of the vehicle's behavior. The optional success / failure condition subunits are used to define the operational state of discrete behaviors. For example, the success condition for the "stop vehicle" behavior can be set to the vehicle speed dropping to zero. Through the above steps, this invention can construct diverse and highly controllable discrete behaviors, laying the foundation for the generation of subsequent complex scenarios.

[0058] The behavior generator proposed in this invention constructs an interactive behavior topology network by flexibly combining discrete behaviors, thereby endowing the scene with dynamic interactivity. In practical applications, the behavior generator first selects scene-related behavior units from a discrete behavior library based on the scene description. Subsequently, according to the specific requirements of the scene, the selected behavior units are configured with parameters. For example, when the scene description includes "vehicle A quickly cuts into vehicle B's lane," the "cut into lane" behavior unit is selected, the target vehicle is set to vehicle B, and the cutting speed is set to a high value. To simulate more complex vehicle behaviors, this invention also supports connecting multiple discrete behaviors using sequential or parallel logic to form behavior sequences or combinations. For example, behaviors can be executed in the order of "accelerate-cut into lane," or "follow vehicle" and "keep lane" behaviors can be executed simultaneously. Figure 4 In the network, modules represent discrete behaviors, while arrows represent relationships between behaviors. The network structure can be updated in real time based on the environment and vehicle status. Figure 4 Taking the behavioral topology network as an example, adversary vehicle 1 travels straight to the intersection and turns right without interacting with any other vehicles. If no dangerous situation is detected, this vehicle will complete the predetermined behavior independently. Adversary vehicles 2 and 3 interact with the vehicle itself. Adversary vehicle 2 reads the position and speed information of the vehicle in real time, increases its speed to overtake the vehicle, and changes lanes before fully overtaking it. Adversary vehicle 3 also reads the position and speed information of the adversary vehicles in real time and plans its following strategy based on its own adaptive cruise control algorithm. Adversary vehicle 4 only interacts with adversary vehicle 3, reading its speed and position information in real time to follow it. This dynamic adjustment mechanism gives background vehicles the ability to respond to changes in the behavior of the vehicle itself, realizing dynamic interaction between vehicles.

[0059] The behavioral topology network design enables mutual reference and interaction between background vehicles and between background vehicles and the driver vehicle. This interactivity makes the scene more dynamic and realistic, allowing the behavior of surrounding vehicles to be dynamically adjusted based on the behavior of the driver vehicle, thereby improving the safety and challenge of the scene. In this way, users can easily generate test scenarios that meet real-world needs, reducing reliance on real data, lowering testing costs, and improving the training efficiency and safety of autonomous driving systems. This unique design enables the embodiments of the present invention to significantly improve the training efficiency and safety of autonomous driving systems, reduce testing costs, and provide a new solution for the development of autonomous driving technology.

[0060] In addition, such as Figure 2As shown, the embodiments of the present invention may further include: a random traffic generator, which randomly generates traffic flow around the vehicle based on the vehicle's position and the position of the adversary vehicle to simulate a real traffic scenario, and then inputs it into the scenario generation module 106, together with weather parameters, multiple vehicle generation positions, multiple adversary vehicle static positions and multiple interactive dynamic behaviors to construct multiple comprehensive test scenarios.

[0061] In summary, the interactive test scenario generation device for autonomous driving proposed in the embodiments of the present invention has the following beneficial effects:

[0062] (1) By integrating multiple learning-based or non-learning-based vehicle control algorithms, the behavior of the background vehicle is ensured to be constrained not only by dynamics but also by the control model, thereby improving the realism and reliability of scene generation. This design allows users to generate highly realistic and reliable test scenarios, reducing reliance on real data, lowering testing costs, and improving the training efficiency and safety of autonomous driving systems;

[0063] (2) It can realize real-time interaction between the surrounding vehicle and the self vehicle, and between surrounding vehicles, to obtain dynamic scenes; and it can make corresponding countermeasures when the self vehicle behaves differently. This makes the generated scenes not only highly realistic, but also dynamically adjust the vehicle behavior in the scene according to the user's needs. It also enables the autonomous driving system to face more diverse and challenging scenes during the training process, thereby further improving the robustness and safety of the system.

[0064] (3) Through its unique modular design, integrated vehicle control algorithm and real-time interaction capabilities, it provides a more effective and flexible solution for the testing and training of autonomous vehicles. This innovative solution not only improves the efficiency and realism of scene generation, but also provides a new solution for the development of autonomous driving technology.

[0065] Next, with reference to the accompanying drawings, a method for generating interactive test scenarios for autonomous driving according to an embodiment of the present invention is described.

[0066] Figure 5 This is a flowchart illustrating a method for generating an interactive test scenario for autonomous driving, as provided in an embodiment of the present invention.

[0067] like Figure 5 As shown, the method for generating interactive test scenarios for autonomous driving includes the following steps:

[0068] In step S501, the target natural language is decomposed and supplemented to generate multiple scene descriptions, including weather information, vehicle location, relative position of adversarial vehicles, and interactive behavior library.

[0069] In step S502, a scene simulation is performed based on weather information to generate weather parameters that can be used for rendering the simulation environment.

[0070] In some embodiments, weather parameters include at least one of solar altitude angle, wind force, precipitation, fog concentration, and dust storm intensity.

[0071] In step S503, multiple vehicle generation locations are searched on a preset autonomous driving simulator map based on the vehicle's location.

[0072] In some embodiments, searching for multiple vehicle-generated locations on a preset autonomous driving simulator map based on the vehicle's location includes:

[0073] Construct a location feasibility judgment function based on the vehicle's location;

[0074] Based on a preset autonomous driving simulator map, a location feasibility judgment function is used to search for multiple autonomous vehicle generation locations that meet the preset location requirements.

[0075] In step S504, multiple static positions of the opposing vehicles are searched on a preset autonomous driving simulator map based on the relative positions of the opposing vehicles and the position of the own vehicle.

[0076] In some embodiments, searching for multiple static locations of adversary vehicles on a preset autonomous driving simulator map based on the relative positions of the adversary vehicles and the vehicle's position includes:

[0077] A surrounding vehicle position information function is constructed based on the relative positions of the opposing vehicles;

[0078] The vehicle's position is input into the surrounding vehicle position information function to search for multiple static positions of adversarial vehicles in a preset autonomous driving simulator map.

[0079] In step S505, a behavior topology network is constructed based on the interactive behavior library, and multiple interactive dynamic behaviors are generated using the behavior topology network.

[0080] In some embodiments, a behavior topology network is constructed based on an interactive behavior library, and multiple interactive dynamic behaviors are generated using the behavior topology network, including:

[0081] Select multiple discrete behaviors from the interactive behavior library, and construct a behavior topology network based on the multiple discrete behaviors;

[0082] Generate interactive dynamic behaviors using behavioral topology networks.

[0083] In some embodiments, selecting multiple discrete behaviors from an interactive behavior library and constructing a behavior topology network based on the multiple discrete behaviors includes:

[0084] Filter multiple discrete behaviors from the interactive behavior library that match multiple scenario descriptions;

[0085] Select the corresponding autonomous driving model for a vehicle performing any discrete action;

[0086] The behavioral parameters of the corresponding autonomous driving model are interactively adjusted based on each discrete behavior.

[0087] Define the motion state of each discrete behavior;

[0088] Based on the motion state of each discrete behavior, each discrete behavior is sequentially or logically connected in parallel to form a behavior topology network.

[0089] In step S506, multiple comprehensive test scenarios are constructed based on weather parameters, multiple self-generated vehicle locations, multiple static locations of adversarial vehicles, and multiple interactive dynamic behaviors.

[0090] It should be noted that the foregoing explanation of the embodiment of the autonomous driving interactive test scenario generation device also applies to the autonomous driving interactive test scenario generation method of this embodiment, and will not be repeated here.

[0091] The interactive test scenario generation method for autonomous driving proposed in this embodiment of the invention has the following beneficial effects:

[0092] (1) By integrating multiple learning-based or non-learning-based vehicle control algorithms, the behavior of the background vehicle is ensured to be constrained not only by dynamics but also by the control model, thereby improving the realism and reliability of scene generation. This design allows users to generate highly realistic and reliable test scenarios, reducing reliance on real data, lowering testing costs, and improving the training efficiency and safety of autonomous driving systems;

[0093] (2) It can realize real-time interaction between the surrounding vehicle and the self vehicle, and between surrounding vehicles, to obtain dynamic scenes; and it can make corresponding countermeasures when the self vehicle behaves differently. This makes the generated scenes not only highly realistic, but also dynamically adjust the vehicle behavior in the scene according to the user's needs. It also enables the autonomous driving system to face more diverse and challenging scenes during the training process, thereby further improving the robustness and safety of the system.

[0094] (3) Through its unique modular design, integrated vehicle control algorithm and real-time interaction capabilities, it provides a more effective and flexible solution for the testing and training of autonomous vehicles. This innovative solution not only improves the efficiency and realism of scene generation, but also provides a new solution for the development of autonomous driving technology.

[0095] Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0096] The electronic device may include: a memory 601, a processor 602, and a computer program stored on the memory 601 and capable of running on the processor 602.

[0097] When the processor 602 executes the program, it implements the method for generating interactive test scenarios for autonomous driving provided in the above embodiments.

[0098] Furthermore, electronic devices also include:

[0099] Communication interface 603 is used for communication between memory 601 and processor 602.

[0100] The memory 601 is used to store computer programs that can run on the processor 602.

[0101] The memory 601 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0102] If the memory 601, processor 602, and communication interface 603 are implemented independently, then the communication interface 603, memory 601, and processor 602 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0103] Optionally, in a specific implementation, if the memory 601, processor 602, and communication interface 603 are integrated on a single chip, then the memory 601, processor 602, and communication interface 603 can communicate with each other through an internal interface.

[0104] Processor 602 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0105] This invention also provides a computer program product, which, when executed by a processor, implements the above-described method for generating interactive test scenarios for autonomous driving.

[0106] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for generating interactive test scenarios for autonomous driving.

[0107] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0108] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0109] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0110] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0111] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0112] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it includes one or a combination of the steps of the method embodiments.

[0113] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0114] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. An interactive test scenario generation device for autonomous driving, characterized in that, include: An expander is used to decompose and supplement the target natural language description to generate multiple scene descriptions, wherein the multiple scene descriptions include weather information, vehicle location, relative position of adversarial vehicles, and interactive behavior library; A weather simulator is used to simulate scenarios based on the weather information in order to generate weather parameters that can be used for rendering the simulation environment; The vehicle locator is used to search for multiple vehicle-generated locations on a preset autonomous driving simulator map based on the vehicle's location. An adversary vehicle locator is used to search for multiple static locations of adversary vehicles on a preset autonomous driving simulator map based on the relative positions of the adversary vehicles and the position of the vehicle. A behavior generator is used to construct a behavior topology network based on the interactive behavior library, and to generate multiple interactive dynamic behaviors using the behavior topology network, specifically including: The third construction unit is used to select multiple discrete behaviors from the interactive behavior library and construct the behavior topology network based on the multiple discrete behaviors. The generation unit is used to generate interactive dynamic behaviors using the behavior topology network; The third construction unit includes a filtering subunit, a vehicle model selection subunit, a vehicle model configuration subunit, a behavior success / failure condition subunit, and a construction subunit. The filtering subunit is used to filter multiple discrete behaviors that conform to the descriptions of the multiple scenarios in the interactive behavior library; The vehicle model selection subunit is used to select the corresponding autonomous driving model for a vehicle performing any discrete behavior; The vehicle model configuration subunit is used to interactively adjust the behavior parameters of its corresponding autonomous driving model according to each discrete behavior. The success / failure condition subunit is used to define the motion state of each discrete behavior. The construction subunit is used to sequentially or in parallel logical connection each discrete behavior according to the motion state of each discrete behavior, so as to form the behavior topology network; The scene generation module is used to construct multiple comprehensive test scenarios based on the weather parameters, the multiple self-generated vehicle locations, the multiple static locations of the multiple adversarial vehicles, and the multiple interactive dynamic behaviors.

2. The interactive test scenario generation device for autonomous driving according to claim 1, characterized in that, The weather parameters include at least one of the following: solar altitude angle, wind force, precipitation, fog concentration, and dust storm intensity.

3. The interactive test scenario generation device for autonomous driving according to claim 1, characterized in that, The vehicle locator includes: The first construction unit is used to construct a location feasibility judgment function based on the vehicle's location; The first search unit is used to search for multiple vehicle-generated locations that meet the preset location requirements based on the preset autonomous driving simulator map and using the location feasibility judgment function.

4. The interactive test scenario generation device for autonomous driving according to claim 1, characterized in that, The counter-vehicle locator includes: The second construction unit is used to construct a surrounding vehicle position information function of the vehicle position based on the relative position of the opposing vehicle; The second search unit is used to input the vehicle's position into the surrounding vehicle position information function in order to search for the static positions of the multiple adversarial vehicles in the preset autonomous driving simulator map.

5. A method for generating interactive test scenarios for autonomous driving, characterized in that, The autonomous driving interactive test scenario generation device according to any one of claims 1-4 includes the following steps: The target natural language is decomposed and supplemented to generate multiple scene descriptions, wherein the multiple scene descriptions include weather information, vehicle location, relative position of adversarial vehicles, and interactive behavior library; The scene is simulated based on the weather information to generate weather parameters that can be used for rendering the simulation environment; Based on the vehicle's location, multiple vehicle locations are generated by searching on a preset autonomous driving simulator map; Based on the relative positions of the opposing vehicles and the position of the autonomous vehicle, multiple static positions of the opposing vehicles are searched on the preset autonomous driving simulator map; A behavior topology network is constructed based on the interactive behavior library, and multiple interactive dynamic behaviors are generated using the behavior topology network. Multiple comprehensive test scenarios are constructed based on the weather parameters, the multiple self-generated vehicle locations, the multiple static locations of the multiple adversarial vehicles, and the multiple interactive dynamic behaviors.

6. An electronic device, characterized in that, include: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the autonomous driving interactive test scenario generation method as described in claim 5.

7. A computer program product, characterized in that, When the computer program / instruction is executed by the processor, it implements the autonomous driving interactive test scenario generation method as described in claim 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the autonomous driving interactive test scenario generation method as described in claim 5.

Images