Automatic driving simulation task execution method and device, electronic equipment and storage medium

Abstract

The invention discloses an automatic driving simulation task execution method and device, electronic equipment and a storage medium, and the method comprises the steps: selecting a to-be-simulated automatic driving simulation scene in response to the uncompleted establishment of a message queue, and splitting the to-be-simulated automatic driving simulation scene into a plurality of subtasks according to the characteristics of the simulation scene; dividing the sub-tasks into different batches and sending the batches to a message queue; and in response to completion of establishment of the message queue, continuously optimizing the simulation result on the premise of keeping the credibility of the simulation result. According to the invention, smooth execution of automatic driving simulation tasks in multiple types of scenes is realized on the premise of ensuring credibility.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to an autonomous driving simulation task execution method, device, electronic device, and storage medium. Background Technology

[0002] With the rapid development of autonomous driving technology, thorough simulation testing of autonomous driving scenarios is crucial. However, autonomous driving simulation tasks across multiple scenarios face numerous challenges during execution, such as the large computational load of complex scenario simulations and resource waste caused by unreasonable scheduling of different scenario types, which seriously affects simulation execution efficiency. At the same time, to ensure the safety and reliability of the autonomous driving system, the credibility of the simulation results must be guaranteed.

[0003] Therefore, designing a solution that balances credibility and accelerates the execution of autonomous driving simulation tasks in various scenarios has become an urgent problem to be solved. Summary of the Invention

[0004] This application provides an autonomous driving simulation task execution method, device, electronic device, and storage medium to achieve the smooth execution of autonomous driving simulation tasks in various scenarios while ensuring credibility.

[0005] The embodiments of this application adopt the following technical solutions:

[0006] In a first aspect, embodiments of this application provide a method for executing an autonomous driving simulation task, wherein the execution method includes:

[0007] In response to the incomplete establishment of the message queue, the autonomous driving simulation scenario to be simulated is selected and split into multiple sub-tasks according to the characteristics of the simulation scenario;

[0008] Based on the multiple subtasks, they are divided into different batches and sent to the message queue; and

[0009] In response to the completion of message queue establishment, the simulation results are continuously optimized while maintaining their credibility.

[0010] In some embodiments, the step of continuously optimizing the simulation results in response to the completion of message queue establishment, while maintaining the credibility of the simulation results, includes:

[0011] After the message queue is established, an autonomous driving simulation task is executed to obtain simulation results;

[0012] The simulation results are compared between the simulated vehicle and the actual vehicle in terms of: vehicle position, vehicle orientation, vehicle rotation angle, vehicle speed, and acceleration.

[0013] If the error of the comparison results is within the credibility range, the simulation results are continuously optimized.

[0014] In some embodiments, the step of continuously optimizing the simulation results in response to the completion of message queue establishment, while maintaining the credibility of the simulation results, includes:

[0015] A multi-precision model library is pre-established based on various types of scenarios and simulation task requirements;

[0016] Based on the multi-precision model library, different precisions are selected for adaptive switching when corresponding to different types of scenarios and simulation tasks, and the credibility of simulation results is guaranteed based on precision during the adaptive switching process.

[0017] In some embodiments, the step of continuously optimizing the simulation results in response to the completion of message queue establishment, while maintaining the credibility of the simulation results, includes:

[0018] The error of the simulation results generated by the simulation model is monitored and predicted in real time.

[0019] An error compensation model was trained using historical simulation data and actual test data.

[0020] During the simulation, the simulation model is dynamically corrected based on real-time data and the prediction results of the error compensation model.

[0021] In some embodiments, the method further includes:

[0022] If the error of the comparison result is determined to be outside the credibility range, the simulation task should be retried.

[0023] The priority of the simulation task retry is adjusted according to the dynamic priority algorithm;

[0024] In response to the stated priority, the simulation task is retried.

[0025] In some embodiments, after dividing the multiple subtasks into different batches and sending them to the message queue, the process includes:

[0026] The simulation process is accelerated using GPUs, depending on the computational intensity or task type of the simulation task.

[0027] In some embodiments, in response to the incomplete establishment of the message queue, an autonomous driving simulation scenario to be simulated is selected and divided into multiple sub-tasks based on the characteristics of the simulation scenario, including:

[0028] The simulation scenarios are divided into multiple sub-tasks based on different types of simulation scenarios, and simulation scenarios with the same characteristics are split into the same sub-task.

[0029] In some embodiments, in response to the incomplete establishment of the message queue, selecting an autonomous driving simulation scenario to be simulated and splitting it into multiple sub-tasks based on the characteristics of the simulation scenario includes:

[0030] Establish a monitoring mechanism to dynamically assess the resource requirements of each computing task and the load of computing nodes in the simulation scenario;

[0031] Based on the monitoring results, the resource allocation strategy and the requirements of different simulation scenarios are dynamically adjusted, and idle computing resources are allocated to computing tasks or reserved for computing tasks.

[0032] Secondly, embodiments of this application also provide an autonomous driving simulation task execution device, wherein the execution device includes:

[0033] The first processing module is used to respond to the incomplete establishment of the message queue by selecting the autonomous driving simulation scenario to be simulated and splitting it into multiple sub-tasks according to the characteristics of the simulation scenario.

[0034] The second processing module is used to divide the multiple subtasks into different batches and send them to the message queue; and

[0035] The persistence module is used to continuously optimize the simulation results while maintaining their credibility in response to the completion of message queue establishment.

[0036] Thirdly, embodiments of this application also provide an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the above-described method.

[0037] Fourthly, embodiments of this application also provide a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the above-described method.

[0038] The at least one technical solution adopted in this application embodiment can achieve the following beneficial effects: In response to the incomplete establishment of the message queue, an autonomous driving simulation scenario to be simulated is selected and multiple sub-tasks are obtained according to the characteristics of the simulation scenario. After the multiple sub-tasks are divided into different batches and sent to the message queue, in response to the completion of the message queue establishment, the simulation results are continuously optimized while maintaining the credibility of the simulation results. Through the above method, the smooth execution of autonomous driving simulation tasks of multiple types of scenarios is achieved while ensuring credibility. Attached Figure Description

[0039] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0040] Figure 1 This is a flowchart illustrating the method for executing autonomous driving simulation tasks in the embodiments of this application;

[0041] Figure 2 This is a schematic diagram illustrating the implementation principle of the autonomous driving simulation task execution method in the embodiments of this application;

[0042] Figure 3 This is a schematic diagram of the structure of the autonomous driving simulation task execution device in the embodiments of this application;

[0043] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0045] Based on the analysis of the current status and problems of autonomous driving simulation tasks in various scenarios, autonomous driving simulation technology is constantly developing and can now construct increasingly complex and realistic virtual driving environments, covering various road types, traffic conditions, driving behaviors, and weather conditions. Numerous research institutions and enterprises have invested heavily in the development of autonomous driving simulation platforms, making it possible to test large-scale simulation cases of different types of autonomous driving scenarios.

[0046] The inventor believes that the following shortcomings still exist:

[0047] (1) Low simulation efficiency. When facing large-scale, multi-type autonomous driving simulation scenarios, the traditional serial computing mode leads to excessively long task execution time. For example, for scenarios involving a large number of intelligent agent interactions and complex environmental changes, the amount of simulation computation increases exponentially, and ordinary computing resources are difficult to complete within a reasonable time. Due to the differences in resource requirements of different types of scenarios, the existing task scheduling strategy lacks flexibility and cannot make full use of computing resources, resulting in a situation where resource idleness and shortage coexist, further reducing simulation efficiency.

[0048] (2) Challenges in Ensuring Reliability. Simplified algorithms adopted to improve simulation speed may affect the accuracy and reliability of simulation results. In some cases, oversimplified environmental or vehicle dynamics models cannot realistically reflect various situations in actual driving, leading to significant deviations between the results and reality. Furthermore, the lack of effective verification and evaluation mechanisms makes it difficult to determine whether simulation results truly meet high reliability requirements. In complex autonomous driving scenarios, even small errors can accumulate and cause qualitative changes, thus affecting the decision-making and control of the entire autonomous driving system.

[0049] To address the aforementioned shortcomings, the autonomous driving simulation task execution method in this application proposes an innovative design scheme through in-depth analysis of the limitations of existing technologies. By optimizing the simulation process, improving the algorithm model, and optimizing the hardware cluster acceleration parallel architecture, the method significantly accelerates the execution of autonomous driving simulation tasks in various scenarios while ensuring the reliability of simulation results, thus providing strong support for the development of autonomous driving technology.

[0050] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0051] This application provides a method for executing autonomous driving simulation tasks, such as... Figure 1 As shown, a flowchart of an autonomous driving simulation task execution method in an embodiment of this application is provided. The method includes at least the following steps S110 to S130:

[0052] In step S110, in response to the incomplete establishment of the message queue, the autonomous driving simulation scenario to be simulated is selected and split into multiple sub-tasks according to the characteristics of the simulation scenario.

[0053] A message queue (MQ) is a middleware used for asynchronous communication in distributed systems; essentially, it's a container for storing and transmitting messages. Before creating a message queue, the autonomous driving simulation scenario to be simulated is selected, and then it's broken down into multiple sub-tasks based on the characteristics of the simulation scenario. For example, after splitting the scenario into multiple sub-tasks, information such as the autonomous driving image, the cluster where the simulation is located, the simulator, and the priority can be selected.

[0054] Step S120: Divide the multiple subtasks into different batches and send them to the message queue.

[0055] The multiple subtasks are divided into multiple batches and sent to the message queue.

[0056] Step S130: In response to the completion of establishing the message queue, the simulation results are continuously optimized while maintaining the credibility of the simulation results.

[0057] Once the message queue is established, it can buffer messages to handle high concurrency. This allows for continuous optimization of the simulation results while maintaining their credibility in autonomous driving simulation scenarios.

[0058] By using the above methods, we can ensure the efficiency of autonomous driving simulation tasks in various scenarios while maintaining the high reliability of the simulation results.

[0059] Based on the above methods, a method for accelerating the execution of autonomous driving simulation tasks in various scenarios while maintaining credibility is proposed. Through task allocation and scheduling strategies based on a hybrid computing model, algorithm optimization and model improvement under a credibility guarantee mechanism, and collaborative design of hardware acceleration and software optimization, the execution efficiency of simulation tasks is effectively improved while ensuring the credibility of simulation results. Furthermore, specific experimental results verify the effectiveness and feasibility of the proposed method, providing strong technical support for the development of autonomous driving technology.

[0060] In one embodiment of this application, the step of continuously optimizing the simulation results in response to the completion of establishing the message queue, while maintaining the credibility of the simulation results, includes: after establishing the message queue, executing an autonomous driving simulation task to obtain simulation results; determining the comparison results between the simulated vehicle and the actual vehicle in the simulation results: vehicle position comparison results, vehicle orientation comparison results, vehicle rotation angle comparison results, vehicle speed and acceleration comparison results; and continuously optimizing the simulation results if the error of the comparison results is within the credibility range.

[0061] like Figure 2 As shown, the simulation engine consumes message queue (MQ) messages to download resources (including but not limited to map resources and simulation task cases), loads configurations (mainly vehicle parameters), and executes the simulation.

[0062] After establishing the message queue, the autonomous driving simulation task is executed to obtain simulation results. It is then necessary to determine whether the simulation results match the actual vehicle's performance based on credibility criteria. Specifically, this includes comparing the simulation vehicle's position, orientation, rotation angle, speed, and acceleration against the actual vehicle. If the error in the comparison results is within the credibility range, the credibility criteria are considered met, and continuous optimization of the simulation results is required.

[0063] During model switching, a rigorous calibration and verification mechanism is established. By comparing and analyzing the output results of models with different precision levels, model parameters are adjusted to ensure that the output results of the low-precision model are consistent with those of the high-precision model within the error range, thereby guaranteeing the reliability of the simulation results.

[0064] Specifically, by comparing the vehicle's position, yaw direction, rotation angle, speed, and acceleration, the simulated vehicle and the actual vehicle in the entire scene are determined.

[0065] In one embodiment of this application, the step of continuously optimizing the simulation results in response to the completion of establishing the message queue while maintaining the credibility of the simulation results includes: pre-establishing a multi-precision model library according to various types of scenarios and simulation task requirements; selecting different precisions for adaptive switching when corresponding to different types of scenarios and simulation tasks according to the multi-precision model library; and ensuring the credibility of the simulation results based on precision during the adaptive switching process.

[0066] Based on the characteristics of various scenarios and the different needs of simulation stages, a multi-precision model library is established. This library not only allows for adaptive switching between different precision levels to accommodate different scenarios and simulation tasks, but also ensures the reliability of simulation results based on precision during the adaptive switching process.

[0067] Specifically, in the initial stage, high-precision physical and environmental models are used for rapid modeling and preliminary verification. For example, when simulating vehicle dynamics, a high-precision mechanical model is used to accurately calculate the forces acting on the vehicle. Then, during the simulation, based on real-time monitoring data and error analysis results, the model is adaptively switched to a model of appropriate precision. Alternatively, for some local areas or non-critical parameters that have a minor impact on the overall results, a low-precision model can be used for rapid calculation to reduce the computational load.

[0068] In one embodiment of this application, the step of continuously optimizing the simulation results in response to the completion of establishing the message queue, while maintaining the credibility of the simulation results, includes: real-time monitoring and prediction of the error of the simulation results generated by the simulation model; training an error compensation model using historical simulation data and actual test data; and dynamically correcting the simulation model during the simulation execution process according to real-time data and the prediction results of the error compensation model.

[0069] In machine learning-based error compensation and correction, machine learning algorithms are used to monitor and predict errors generated during simulation in real time. A large amount of historical simulation data and actual test data is collected to train the error compensation model. For example, neural networks are used to model and compensate for errors in environmental perception data, improving the accuracy of the agent's environmental perception.

[0070] During the simulation, the simulation model is dynamically corrected based on real-time data and the prediction results of the error compensation model. For example, when a deviation is found between the simulated vehicle speed and the actual situation, the parameters in the speed calculation model are adjusted in a timely manner to reduce the accumulation of errors and ensure the reliability of the simulation results.

[0071] In one embodiment of this application, the method further includes: selecting to retry the simulation task when the error of the comparison result is determined to be outside the confidence range; adjusting the priority of the simulation task retry according to a dynamic priority algorithm; and retrying the simulation task in response to the priority.

[0072] Retry design refers to the principle that the priority of retry scenario simulation task cases should not be too high to avoid scenario simulation task cases with unreliable results repeatedly encountering the simulation engine (such as machine performance bottlenecks). At the same time, the priority of retry design should not be too low to avoid task timeouts due to the inability to obtain computing resources.

[0073] Preferably, the scenario simulation task case for retrying in this application embodiment is designed with a dynamic priority based on simulated annealing mechanism, which can ensure credibility and improve execution efficiency each time.

[0074] In one embodiment of this application, after the multiple subtasks are divided into different batches and sent to the message queue, the following steps are included: in response to the task computation intensity or task processing type in the simulation task, the simulation process is accelerated using a GPU.

[0075] like Figure 2 As shown, for the GPU-accelerated computation process, for computationally intensive tasks in autonomous driving simulation, such as large-scale environment rendering in Worldsim and complex physics calculations in Logsim, the parallel computing capabilities of the graphics processing unit (GPU) are utilized for acceleration. The data processing and rendering tasks of the environment model are distributed to the GPU, and multiple stream processors on the GPU perform simultaneous off-processing, thereby improving the speed and quality of environment rendering.

[0076] In one embodiment of this application, in response to the incomplete establishment of the message queue, an autonomous driving simulation scenario to be simulated is selected and divided into multiple sub-tasks according to the characteristics of the simulation scenario, including: layering different types of simulation scenarios to obtain multiple sub-tasks, and splitting simulation scenarios with the same characteristics into the same sub-task.

[0077] Autonomous driving simulation tasks across various scenarios are stratified according to their nature and complexity. For example, simulation scenario type, simulation vehicle type, simulation algorithm module, simulation map, simulation driving behavior, and simulation obstacle features are used as dimensions for stratified task decomposition. Furthermore, scenarios are grouped by features, dividing the task into multiple sub-tasks. These sub-tasks allow for on-demand server configuration for scenarios with varying resource requirements, while also enabling unified processing of scenarios with common characteristics, avoiding unnecessary overhead caused by repeated resource switching during simulation execution.

[0078] In one embodiment of this application, in response to the incomplete establishment of the message queue, selecting an autonomous driving simulation scenario to be simulated and dividing it into multiple sub-tasks according to the characteristics of the simulation scenario includes: establishing a monitoring mechanism to dynamically evaluate the resource requirements of each computing task and the load of computing nodes in the simulation scenario; dynamically adjusting the resource allocation strategy and the requirements of different simulation scenarios according to the monitoring results, and allocating idle computing resources to computing tasks or reserving computing resources for computing tasks.

[0079] Establish a real-time monitoring mechanism to dynamically assess the resource requirements of each computing task and the load of computing nodes. Based on the monitoring results, dynamically adjust the resource allocation strategy to rationally allocate idle computing resources to urgently needed tasks. For example, when a computing node completes its current task, it should be immediately allocated to other running and resource-intensive tasks to improve the overall utilization of computing resources.

[0080] 1. Execution Time Comparison. Various representative autonomous driving simulation tasks across multiple scenarios were selected for testing. Experimental results show that the average execution time of the simulation tasks is significantly reduced compared to traditional methods after adopting the reliability-considered acceleration design scheme proposed in the application embodiments. For example, in 500 scenarios involving complex urban road scenes and multiple intelligent agent interactions, the traditional method took 2 hours to execute, while the new scheme reduced the execution time to within 40 minutes, demonstrating a significant acceleration effect.

[0081] 2. Credibility Assessment. Comparative analysis of simulation results and actual driving data using Logsim-to-Worldsim conversion revealed that the new approach effectively improves simulation efficiency while maintaining the credibility of the simulation results. In terms of accuracy, the simulation results using the new approach are 7.5% more accurate than those using traditional methods. Regarding deviation from actual driving data, the new approach keeps the average deviation within an acceptable range, meeting the high credibility requirements of autonomous driving simulation.

[0082] Specific simulation experiment setup: (1) Experimental environment: An autonomous driving simulation experimental platform with more than 60 computing nodes was constructed. Each computing node is equipped with a high-performance CPU, GPU, and large-capacity memory and storage devices. The experimental platform supports various types of autonomous driving simulation scenarios and can test and evaluate different algorithms and designs. (2) Test indicators: The performance of the designed scheme is evaluated by multiple indicators such as the execution time of the simulation task, the accuracy of the simulation results, and the deviation from actual driving data. Among them, the execution time reflects the degree of improvement in simulation efficiency, while the accuracy and deviation reflect the credibility of the simulation results.

[0083] This application also provides an autonomous driving simulation task execution device 300, such as... Figure 3As shown, a structural schematic diagram of the autonomous driving simulation task execution device 300 in this application embodiment is provided. The autonomous driving simulation task execution device 300 includes at least: a first processing module 310, a second processing module 320, and a persistence module 330, wherein:

[0084] In one embodiment of this application, the first processing module 310 is specifically used to: in response to the incomplete establishment of the message queue, select the autonomous driving simulation scenario to be simulated and split it into multiple sub-tasks according to the characteristics of the simulation scenario.

[0085] A message queue (MQ) is a middleware used for asynchronous communication in distributed systems; essentially, it's a container for storing and transmitting messages. Before creating a message queue, the autonomous driving simulation scenario to be simulated is selected, and then it's broken down into multiple sub-tasks based on the characteristics of the simulation scenario. For example, after splitting the scenario into multiple sub-tasks, information such as the autonomous driving image, the cluster where the simulation is located, the simulator, and the priority can be selected.

[0086] In one embodiment of this application, the second processing module 320 is specifically used to: divide the multiple subtasks into different batches and send them to the message queue.

[0087] The multiple subtasks are divided into multiple batches and sent to the message queue.

[0088] In one embodiment of this application, the persistence module 330 is specifically used to: continuously optimize the simulation results while maintaining the credibility of the simulation results in response to the completion of establishing the message queue.

[0089] Once the message queue is established, it can buffer messages to handle high concurrency. This allows for continuous optimization of the simulation results while maintaining their credibility in autonomous driving simulation scenarios.

[0090] In one embodiment of this application, the persistence module 330 is further configured to:

[0091] After the message queue is established, an autonomous driving simulation task is executed to obtain simulation results;

[0092] The simulation results are compared between the simulated vehicle and the actual vehicle in terms of: vehicle position, vehicle orientation, vehicle rotation angle, vehicle speed, and acceleration.

[0093] If the error of the comparison results is within the credibility range, the simulation results are continuously optimized.

[0094] In one embodiment of this application, the persistence module 330 is further configured to:

[0095] A multi-precision model library is pre-established based on various types of scenarios and simulation task requirements;

[0096] Based on the multi-precision model library, different precisions are selected for adaptive switching when corresponding to different types of scenarios and simulation tasks, and the credibility of simulation results is guaranteed based on precision during the adaptive switching process.

[0097] In one embodiment of this application, the persistence module 330 is further configured to:

[0098] The error of the simulation results generated by the simulation model is monitored and predicted in real time.

[0099] An error compensation model was trained using historical simulation data and actual test data.

[0100] During the simulation, the simulation model is dynamically corrected based on real-time data and the prediction results of the error compensation model.

[0101] In one embodiment of this application, a retry module is further included, for:

[0102] If the error of the comparison result is determined to be outside the credibility range, the simulation task should be retried.

[0103] The priority of the simulation task retry is adjusted according to the dynamic priority algorithm;

[0104] In response to the stated priority, the simulation task is retried.

[0105] In one embodiment of this application, the second processing module 320 is further configured to:

[0106] The simulation process is accelerated using GPUs, depending on the computational intensity or task type of the simulation task.

[0107] In one embodiment of this application, the first processing module 310 is further configured to:

[0108] The simulation scenarios are divided into multiple sub-tasks based on different types of simulation scenarios, and simulation scenarios with the same characteristics are split into the same sub-task.

[0109] In one embodiment of this application, the first processing module 310 is further configured to:

[0110] Establish a monitoring mechanism to dynamically assess the resource requirements of each computing task and the load of computing nodes in the simulation scenario;

[0111] Based on the monitoring results, the resource allocation strategy and the requirements of different simulation scenarios are dynamically adjusted, and idle computing resources are allocated to computing tasks or reserved for computing tasks.

[0112] It is understood that the above-mentioned autonomous driving simulation task execution device can realize each step of the autonomous driving simulation task execution method provided in the foregoing embodiments. The relevant explanations of the autonomous driving simulation task execution method are applicable to the autonomous driving simulation task execution device, and will not be repeated here.

[0113] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Please refer to it. Figure 4 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for other business operations.

[0114] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0115] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0116] The processor reads the corresponding computer program from non-volatile memory into main memory and then executes it, forming an autonomous driving simulation task execution device at the logical level. The processor executes the program stored in memory and specifically performs the following operations:

[0117] In response to the incomplete establishment of the message queue, the autonomous driving simulation scenario to be simulated is selected and split into multiple sub-tasks according to the characteristics of the simulation scenario;

[0118] Based on the multiple subtasks, they are divided into different batches and sent to the message queue; and

[0119] In response to the completion of message queue establishment, the simulation results are continuously optimized while maintaining their credibility.

[0120] The above is as stated in this application. Figure 1 The method for executing autonomous driving simulation task execution devices disclosed in the illustrated embodiments can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0121] The electronic device can also perform Figure 1 The method for executing autonomous driving simulation tasks by the autonomous driving simulation task execution device is described, and the autonomous driving simulation task execution device is implemented in... Figure 1 The functions of the embodiments shown are not described in detail here.

[0122] This application also proposes a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by an electronic device including multiple applications, enable the electronic device to perform... Figure 1The method executed by the autonomous driving simulation task execution device in the illustrated embodiment is specifically used to execute:

[0123] In response to the incomplete establishment of the message queue, the autonomous driving simulation scenario to be simulated is selected and split into multiple sub-tasks according to the characteristics of the simulation scenario;

[0124] Based on the multiple subtasks, they are divided into different batches and sent to the message queue; and

[0125] In response to the completion of message queue establishment, the simulation results are continuously optimized while maintaining their credibility.

[0126] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0127] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0128] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0130] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0131] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0132] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0133] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0134] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. An automated driving simulation task execution method, wherein, The execution method comprises: in response to the uncompletion of establishing the message queue, selecting an autonomous driving simulation scene to be simulated and splitting into multiple sub-tasks according to the simulation scene characteristics; based on the multiple sub-tasks, dividing into different batches and sending to the message queue; and in response to the completion of establishing the message queue, continuously optimizing the simulation result under the premise of maintaining the credibility of the simulation result.

2. The method of claim 1, wherein, The response to the completion of establishing the message queue, continuously optimizing the simulation result under the premise of maintaining the credibility of the simulation result, comprises: after the completion of establishing the message queue, performing an autonomous driving simulation task to obtain a simulation result; judging the comparison results between the simulation vehicle and the actual vehicle in the simulation result, including the position comparison result of the vehicle, the orientation comparison result of the vehicle, the rotation angle comparison result of the vehicle, the speed comparison result of the vehicle, and the acceleration comparison result of the vehicle; in the case that the error of the comparison result is within the credibility range, continuously optimizing the simulation result.

3. The method of claim 2, wherein, The response to the completion of establishing the message queue, continuously optimizing the simulation result under the premise of maintaining the credibility of the simulation result, comprises: pre-establishing a multi-precision model library according to multiple types of scenes and simulation task requirements; according to the multi-precision model library, selecting different precision adaptive switching corresponding to different types of scenes and simulation tasks, and guaranteeing the credibility of the simulation result based on the precision in the process of adaptive switching.

4. The method of claim 2, wherein, The response to the completion of establishing the message queue, continuously optimizing the simulation result under the premise of maintaining the credibility of the simulation result, comprises: real-time monitoring and predicting the error of the simulation result generated by the simulation model; training an error compensation model by using historical simulation data and actual test data; in the process of simulation execution, dynamically correcting the simulation model according to real-time data and the prediction result of the error compensation model.

5. The method of claim 2, wherein, The method further comprises: in the case that the error of the comparison result is not within the credibility range, selecting simulation task retry; adjusting the priority of the simulation task retry according to a dynamic priority algorithm; in response to the priority, performing simulation task retry.

6. The method of claim 1, wherein, After the multiple sub-tasks are divided into different batches and sent to the message queue, it comprises: in response to the task calculation intensity or task processing type in the simulation task, accelerating the simulation process by using GPU.

7. The method of claim 1, wherein, In response to the uncompletion of establishing the message queue, selecting an autonomous driving simulation scene to be simulated and splitting into multiple sub-tasks according to the simulation scene characteristics, comprises: dividing into multiple sub-tasks according to different types of simulation scenes, and splitting simulation scenes with the same characteristics into the same sub-task.

8. The method of claim 1, wherein, In response to the uncompletion of establishing the message queue, selecting an autonomous driving simulation scene to be simulated and splitting into multiple sub-tasks according to the simulation scene characteristics, comprises: establishing a monitoring mechanism to dynamically evaluate the resource demand of each calculation task and the load condition of the calculation node in the simulation scene; according to the monitoring result, dynamically adjusting the resource allocation strategy and the requirements of different simulation scenes, and allocating idle calculation resources to the calculation task or reserving calculation resources for the calculation task.

9. An automatic driving simulation task execution apparatus, wherein, The execution device comprises: The first processing module is configured to select an automatic driving simulation scene to be simulated and split the simulation scene into a plurality of subtasks according to a simulation scene feature in response to an uncompletion of establishing a message queue. The second processing module is configured to send the plurality of subtasks to the message queue in different batches. The persistence module is configured to continuously optimize the simulation result under the premise of maintaining the credibility of the simulation result in response to a completion of establishing the message queue. 10.An electronic device, comprising: a processor; and a memory arranged to store computer-executable instructions that, when executed, cause the processor to perform the method of any of claims 1-8. 11.A computer-readable storage medium storing one or more programs, the one or more programs, when executed by an electronic device including multiple applications, causing the electronic device to perform the method of any of claims 1-8.

Images