Test method and device of vehicle control system, vehicle and storage medium

Abstract

This application provides a testing method, apparatus, vehicle, and storage medium for a vehicle control system. The method includes: acquiring virtual scene parameters of a virtual test scenario, motion state parameters of a virtual vehicle, and a video injection mode of the vehicle control system; generating a virtual video stream of the virtual vehicle from multiple perspectives based on the virtual scene parameters and motion state parameters; injecting the virtual video stream into the vehicle control system based on the video injection mode, so that the vehicle control system generates control decision results; and determining the test results of the vehicle control system based on the control decision results. This application solves the technical problem of poor vehicle control performance.

Description

Technical Field

[0001] This application relates to the fields of vehicle control and autonomous driving, and more specifically, to a test method, apparatus, vehicle, and storage medium for a vehicle control system. Background Technology

[0002] In the field of vehicle technology, vehicle control systems (such as autonomous driving systems) rely on visual perception modules to perceive the environment in real time. The accuracy of control decisions made by the vehicle control system is crucial for autonomous driving. During the testing of vehicle control systems, multi-view video is generated through a simulation platform and injected into the onboard controller to reproduce extreme situations and verify system performance. Currently, in the industry, fixed-frequency injection cannot adapt to the dynamic load and network conditions of real-time operation of vehicle control systems, resulting in dropped video frames, latency jitter, or time asynchrony, which in turn leads to erroneous vehicle control commands and execution delays.

[0003] There is currently no good solution to the above problems. Summary of the Invention

[0004] This application provides a testing method, apparatus, vehicle, and storage medium for a vehicle control system, to at least solve the technical problem of poor vehicle control performance.

[0005] According to one aspect of the embodiments of this application, a testing method for a vehicle control system is provided, comprising: acquiring virtual scene parameters of a virtual test scenario, motion state parameters of a virtual vehicle, and a video injection mode of the vehicle control system, wherein the video injection mode is used to represent the injection mode for injecting video into the vehicle control system; generating a virtual video stream of the virtual vehicle from multiple perspectives based on the virtual scene parameters and motion state parameters; injecting the virtual video stream into the vehicle control system based on the video injection mode, so that the vehicle control system generates control decision results; and determining the test results of the vehicle control system based on the control decision results.

[0006] Furthermore, the method also includes: if a master-slave mode video injection instruction is received, determining the video injection mode as a master-slave video injection mode, wherein the master-slave video injection mode is used to inject a virtual video stream into the vehicle control system according to the control signal of the master device, and the master device is the vehicle control system; if no master-slave mode video injection instruction is received, obtaining the system status parameters of the vehicle control system, and determining the video injection mode of the vehicle control system based on the system status parameters and the video stream parameters of the virtual video stream.

[0007] Furthermore, based on system state parameters and video stream parameters of the virtual video stream, the video injection mode of the vehicle control system is determined, including: determining whether the virtual video stream meets the preset stability conditions of the vehicle control system based on the system state parameters and video stream parameters, wherein the preset stability conditions are used to represent the conditions for stable transmission of the virtual video stream to the vehicle control system; if the virtual video stream meets the preset stability conditions, the video injection mode is determined to be a continuous video injection mode, wherein the continuous video injection mode is used to inject the virtual video stream into the vehicle control system according to a preset video injection frequency; if the virtual video stream does not meet the preset stability conditions, the video injection mode is determined to be a triggered video injection mode, wherein the triggered video injection mode is used to inject the virtual video stream into the vehicle control system when a key video frame is detected.

[0008] Furthermore, the system status parameters of the vehicle control system are obtained, including: obtaining the network status parameters, system processing load, and video frame occupancy rate of the vehicle control system, wherein the system processing load is used to represent the occupancy ratio of computing resources in the vehicle control system, and the video frame occupancy rate is used to represent the occupancy ratio of video frames in the memory space of the vehicle control system; and the system status parameters are determined based on the network status parameters, system processing load, and video frame occupancy rate.

[0009] Furthermore, based on virtual scene parameters and motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated, including: generating an initial virtual video stream of the virtual vehicle from multiple perspectives based on virtual scene parameters and motion state parameters, wherein the initial virtual video stream contains multiple initial video frames; verifying the multiple initial video frames in the initial virtual video stream to obtain a verification result, wherein the verification result is used to indicate whether there are any abnormal video frames in the multiple initial video frames; updating the initial virtual video stream based on the verification result to obtain a virtual video stream.

[0010] Furthermore, the initial virtual video stream is updated based on the verification results to obtain a virtual video stream, including: if the verification result shows that there are abnormal video frames among multiple initial video frames, the abnormal video frames in the initial virtual video stream are removed to obtain a virtual video stream; if the verification result shows that there are no abnormal video frames among multiple initial video frames, the initial virtual video stream is determined to be a virtual video stream.

[0011] Furthermore, based on the video injection mode, a virtual video stream is injected into the vehicle control system so that the vehicle control system can generate control decision results. This includes: performing time calibration on the virtual video stream and motion state parameters to obtain calibration results; updating the timestamp of the virtual video stream based on the calibration results to obtain an updated virtual video stream; and injecting the updated virtual video stream into the vehicle control system based on the video injection mode so that the vehicle control system can generate control decision results.

[0012] According to another aspect of the embodiments of this application, a testing apparatus for a vehicle control system is also provided, comprising: a host computer for generating virtual scene parameters and motion state parameters of a virtual vehicle for a virtual test scenario; a scene generator for generating a virtual video stream of the virtual vehicle from multiple perspectives based on the virtual scene parameters and motion state parameters; a video injection module for injecting the virtual video stream into the vehicle control system according to a video injection mode; a vehicle control system for generating control decision results based on the virtual video stream; and a result generation module for generating test results of the vehicle control system based on the control decision results.

[0013] According to another aspect of the embodiments of this application, a vehicle is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0014] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0015] In this embodiment, virtual scene parameters of the virtual test scenario, motion state parameters of the virtual vehicle, and video injection mode of the vehicle control system are obtained. Based on the virtual scene parameters and motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated. The virtual video stream is injected into the vehicle control system based on the video injection mode, so that the vehicle control system can generate control decision results. Based on the control decision results, the test results of the vehicle control system are determined. By obtaining the virtual scene parameters of the virtual test scenario, the motion state parameters of the virtual vehicle, and the video injection mode supported by the vehicle control system, a virtual video stream that meets the requirements of multiple perspectives is generated, ensuring the authenticity, synchronization, and compatibility of the video stream, thereby ensuring the accuracy of the test results. Injecting the virtual video stream into the vehicle control system according to the video injection mode ensures that the vehicle control system generates control decision results, realizing dynamic matching of the optimal video injection mode under different conditions, thereby improving the stability, reliability, and resource efficiency of the testing process. The vehicle control system generates control decision results based on the injected video stream, and determines the test results of the vehicle control system based on these control decision results. This achieves high-fidelity verification of the decision-making behavior of the vehicle control system in a virtual simulation environment, improves the stability, accuracy and practicality of the test, and solves the technical problem of poor vehicle control performance. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0017] Figure 1 This is a flowchart of a test method for a vehicle control system according to an embodiment of this application;

[0018] Figure 2 This is a block diagram of the overall structure of an intelligent driving video injection and verification system based on adaptive multi-mode scheduling according to an embodiment of this application;

[0019] Figure 3 This is a flowchart of an intelligent driving video injection and verification system based on adaptive multi-mode scheduling according to an embodiment of this application;

[0020] Figure 4 This is a block diagram of the internal structure of a video injection card / field-programmable gate array according to an embodiment of this application;

[0021] Figure 5 This is a flowchart of an adaptive multi-mode video injection strategy according to an embodiment of this application;

[0022] Figure 6 This is a schematic diagram of a test apparatus for a vehicle control system according to an embodiment of this application;

[0023] Figure 7 This is a schematic diagram of a test apparatus for a vehicle control system according to an embodiment of this application. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] According to an embodiment of this application, a method embodiment for testing a vehicle control system is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] This embodiment provides a test method for a vehicle control system. Figure 1 This is a flowchart of a test method for a vehicle control system according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0028] Step S102: Obtain the virtual scene parameters of the virtual test scenario, the motion state parameters of the virtual vehicle, and the video injection mode of the vehicle control system.

[0029] The aforementioned virtual test scenario refers to a digital 3D scene model constructed in a simulation environment to reproduce a real road environment. A virtual test scenario includes a set of parameterizable elements such as road geometry, static environmental features, dynamic traffic participants, and environmental conditions.

[0030] The aforementioned virtual scene parameters refer to a set of digital parameters used to quantitatively describe each component in a virtual test scene. Virtual scene parameters include, but are not limited to, road curvature radius, number and width of lanes, traffic light status and cycle, initial position and speed of dynamic obstacles, light intensity and color temperature, rainfall intensity, fog concentration, road surface friction coefficient, and sensor noise model parameters.

[0031] The aforementioned virtual vehicle can refer to a digital vehicle entity driven by a high-precision dynamic model in a simulation environment, which has the same motion characteristics as a real vehicle. Its appearance, size, sensor layout, and dynamic response are all modeled according to the parameters of a real vehicle.

[0032] The motion state parameters of the virtual vehicle mentioned above can refer to the multi-dimensional state vector of the real-time motion characteristics of the virtual vehicle in the virtual scene.

[0033] The aforementioned video injection modes refer to modes used to control how, when, and at what frequency video is injected into the vehicle control system. Video injection modes for the vehicle control system can include master-slave mode (injection triggered by the vehicle control system itself), continuous injection mode (continuous injection at a fixed frame rate), and triggered injection mode (injection only when a critical event frame is detected), etc.

[0034] In one optional embodiment, firstly, virtual scene parameters of the virtual test scenario are acquired, and scene element data such as road structure, weather, lighting, and dynamic targets are pre-set or generated in real time through a simulation platform. Secondly, motion state parameters of the virtual vehicle are acquired. These motion state parameters can be real-time vehicle pose, speed, acceleration, and other state quantities output by the vehicle dynamics model based on control commands or trajectory planning. Finally, the video injection mode of the vehicle control system is acquired. This can be determined directly by receiving an external mode video injection command, or, in the absence of an external mode video injection command, by reading the network status, processing load, and video frame occupancy rate built into the vehicle control system, and having the logic judgment module automatically match and determine the video injection mode according to preset rules.

[0035] In another alternative implementation, a default injection mode can be loaded via a preset configuration file, or the user can manually select the injection mode through a human-machine interface, or the vehicle control system can automatically load a matching preset injection mode based on the firmware version and hardware platform type upon startup. The aforementioned preset configuration file is a structured data file stored in the vehicle control system or testing equipment, which predefines a default video injection mode for use as a startup basis in the absence of external commands. The aforementioned preset injection mode can refer to a video injection strategy pre-configured and embedded in the vehicle control system during the vehicle control system design phase, based on a specific hardware or firmware combination, and can be directly invoked without dynamic calculation at runtime.

[0036] This application embodiment accurately obtains virtual scene parameters, vehicle motion state parameters, and video injection mode, providing scene description and motion benchmark for subsequent virtual video stream generation, providing vehicle control system state basis for control decision results, and ensuring that the subsequent vehicle control system test results have correct input. It achieves a high degree of consistency between multiple video streams and the real in-vehicle environment in terms of content authenticity, timing synchronization, and vehicle compatibility.

[0037] Step S104: Based on virtual scene parameters and motion state parameters, generate virtual video streams of the virtual vehicle from multiple perspectives.

[0038] The aforementioned virtual video stream of a virtual vehicle in multiple perspectives can refer to a continuous video sequence generated in real time by a 3D simulation engine based on virtual scene parameters and virtual vehicle motion state parameters, simulating the acquisition of multiple onboard cameras on the vehicle under different perspectives, focal lengths, and distortion characteristics.

[0039] By fusing virtual scene parameters (such as roads, weather, and obstacles) with virtual vehicle motion state parameters (such as pose, speed, and acceleration), a 3D simulation engine can be driven to render visual data from multiple camera perspectives in real time, generating a virtual video stream consistent with the input of a real vehicle-mounted perception system. The aforementioned 3D simulation engine can refer to a software system used to construct and render a 3D virtual environment in real time. Based on input scene parameters (such as road structure, weather, lighting, and static / dynamic obstacles) and vehicle motion state parameters (such as position, attitude, and speed), it can simulate the imaging process of multiple cameras, generating a physically realistic temporal virtual video stream.

[0040] In another alternative embodiment, a virtual video stream can be generated based on an image synthesis method. Optionally, real road video or image databases can be used to synthesize multi-view videos through image transformation techniques (such as viewpoint transformation, optical flow interpolation, and semantic segmentation redrawing). The content of the screen can be dynamically adjusted in combination with vehicle motion parameters, which is suitable for testing needs with rich textures and limited scenes.

[0041] In another alternative embodiment, video stitching and projection mapping can be used to spatially project and temporally synchronize real-world videos from multiple cameras based on the motion parameters of the virtual vehicle, and then generate a virtual video stream through video stitching and background replacement.

[0042] This application generates multi-view virtual video streams to achieve high-fidelity simulation of real-world perceived input, ensuring spatiotemporal synchronization of visual data and reproducibility of scenes, improving test coverage and control decision verification accuracy, and effectively supporting the stability assessment of vehicle control systems.

[0043] Step S106: Inject the virtual video stream into the vehicle control system based on the video injection mode so that the vehicle control system can generate control decision results.

[0044] The aforementioned vehicle control system can refer to the electronic control unit system on the vehicle used to realize autonomous driving or assisted driving functions.

[0045] The aforementioned control decision result can refer to the driving decision command output by the vehicle control system based on the injected virtual video stream and its own perception and planning algorithms.

[0046] For example, control decision results can be acceleration / braking signals, steering angle commands, lane keeping status, emergency obstacle avoidance actions, etc., which are used to evaluate whether the behavior of the vehicle control system in the simulation environment conforms to the safe, compliant and expected driving strategy.

[0047] In one optional embodiment, firstly, the vehicle control system adaptively selects the injection mode based on the load and network status, accurately synchronizing the virtual video stream and injecting it into the vehicle control system. Subsequently, based on the injected video and motion parameters, the vehicle control system generates control commands such as braking and steering, completing a closed-loop response from perception to decision. Finally, the host computer collects the control decision results, compares them with the expected safe behavior, and evaluates the response latency and the effectiveness of hazard avoidance.

[0048] This application embodiment achieves accurate synchronous input of virtual video streams and vehicle control systems through an adaptive video injection mode, ensuring stable and timely sensing data, overcoming the limitations of real-world testing, and effectively improving the problem of poor vehicle control performance. By injecting virtual video streams into the vehicle control system based on the video injection mode, it ensures that the input matches the system state, improves the accuracy and stability of control decision results, and achieves an efficient and reliable simulation testing closed loop.

[0049] Step S108: Determine the test results of the vehicle control system based on the control decision results.

[0050] The aforementioned test results refer to the quantitative evaluation conclusions generated by comparing the control decision results output by the vehicle control system with preset safety specifications, expected behavior models, or ideal response trajectories. Test results can include decision delay, false trigger rate, obstacle avoidance success rate, trajectory deviation, number of violations, etc., used to objectively measure the functionality, stability, and safety performance of the vehicle control system in simulation scenarios.

[0051] In one optional embodiment, firstly, the control decision results generated by the vehicle control system based on the virtual video stream are acquired. Then, based on these control decision results, it is determined whether they meet preset safety and functional expectations, including response latency, consistency of decision logic, and effectiveness of risk avoidance. Finally, based on the determination results, the test results of the vehicle control system are determined to evaluate its control performance and stability in a virtual scenario.

[0052] In this embodiment, virtual scene parameters of the virtual test scenario, motion state parameters of the virtual vehicle, and video injection mode of the vehicle control system are obtained. Based on the virtual scene parameters and motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated. The virtual video stream is injected into the vehicle control system based on the video injection mode, so that the vehicle control system can generate control decision results. Based on the control decision results, the test results of the vehicle control system are determined. By obtaining the virtual scene parameters of the virtual test scenario, the motion state parameters of the virtual vehicle, and the video injection mode supported by the vehicle control system, a virtual video stream that meets the requirements of multiple perspectives is generated, ensuring the authenticity, synchronization, and compatibility of the video stream, thereby ensuring the accuracy of the test results. Injecting the virtual video stream into the vehicle control system according to the video injection mode ensures that the vehicle control system generates control decision results, realizing dynamic matching of the optimal video injection mode under different conditions, thereby improving the stability, reliability, and resource efficiency of the testing process. The vehicle control system generates control decision results based on the injected video stream, and determines the test results of the vehicle control system based on these control decision results. This achieves high-fidelity verification of the decision-making behavior of the vehicle control system in a virtual simulation environment, improves the stability, accuracy and practicality of the test, and solves the technical problem of poor vehicle control performance.

[0053] Optionally, upon receiving a master-slave mode video injection instruction, the video injection mode is determined to be a master-slave video injection mode, wherein the master-slave video injection mode is used to inject a virtual video stream into the vehicle control system according to the control signal of the master device, and the master device is the vehicle control system; if no master-slave mode video injection instruction is received, the system status parameters of the vehicle control system are obtained, and the video injection mode of the vehicle control system is determined based on the system status parameters and the video stream parameters of the virtual video stream.

[0054] The aforementioned master-slave mode video injection command can be a communication command actively issued by the vehicle control system to explicitly request to take over control of video injection.

[0055] The aforementioned master-slave video injection mode can refer to a collaborative working mode in which the vehicle control system acts as the master device, actively sending control signals to instruct external devices to inject virtual video streams according to timing and triggering requirements. This mode is used to inject virtual video streams into the vehicle control system based on the master device's control signals.

[0056] The aforementioned system status parameters can refer to key indicators that are collected in real time by the vehicle control system during operation and can be used to evaluate its resource carrying capacity and communication stability.

[0057] The aforementioned video stream parameters can refer to quantitative indicators that characterize the video stream's features and transmission requirements, such as frame rate, resolution, bit rate, timestamp accuracy, data packet size, and transmission interval.

[0058] In one optional embodiment, when the vehicle control system receives a master-slave mode video injection command, it immediately confirms the current video injection mode as a master-slave mode. Based on this, the vehicle control system, acting as the master device, controls the injection timing and rhythm of the virtual video stream in real time according to internal scheduling signals. If the master-slave mode video injection command is not received, the system status parameters of the vehicle control system are automatically read, including network status parameters, system processing load, and video frame occupancy rate. These parameters are then combined with the video stream parameters of the virtual video stream, which may include frame rate, resolution, and data throughput, to determine the video injection mode of the vehicle control system.

[0059] For example, the vehicle control system determines whether the virtual video stream meets the system's preset stability conditions. If the system processing load is less than 70%, the video frame occupancy rate is less than 60%, and the network jitter is less than 5ms, then the stability conditions are met, and the video injection mode is set to continuous video injection mode, continuously injecting the virtual video stream at a fixed frequency. Otherwise, the stability conditions are not met, and the system switches to trigger-based video injection mode, triggering injection only when key video frames are detected (such as pedestrians entering or traffic lights changing color), in order to reduce the system load and ensure the effectiveness of injection.

[0060] This application embodiment switches the injection mode by using a master-slave mode video injection command. In the absence of a master-slave mode video injection command, the video injection mode of the vehicle control system is determined by using the system status parameters and video stream parameters of the vehicle control system. This ensures stable injection of the virtual video stream, improves the accuracy of control decisions, and achieves efficient and reliable testing.

[0061] Optionally, based on system state parameters and video stream parameters of the virtual video stream, the video injection mode of the vehicle control system is determined, including: determining whether the virtual video stream meets the preset stability conditions of the vehicle control system based on the system state parameters and video stream parameters, wherein the preset stability conditions are used to represent the conditions for stable transmission of the virtual video stream to the vehicle control system; if the virtual video stream meets the preset stability conditions, the video injection mode is determined to be a continuous video injection mode, wherein the continuous video injection mode is used to inject the virtual video stream into the vehicle control system according to a preset video injection frequency; if the virtual video stream does not meet the preset stability conditions, the video injection mode is determined to be a triggered video injection mode, wherein the triggered video injection mode is used to inject the virtual video stream into the vehicle control system when a key video frame is detected.

[0062] The aforementioned preset stability conditions can refer to the conditions for the stable transmission of virtual video streams to the vehicle control system.

[0063] The aforementioned continuous video injection mode can refer to the injection of virtual video into the vehicle control system according to a preset video injection frequency.

[0064] The aforementioned preset video injection frequency can refer to the fixed number of virtual video frames injected into the vehicle control system per unit time.

[0065] The aforementioned triggered video injection mode refers to a mode that injects a virtual video stream into the vehicle control system when a key video frame is detected.

[0066] In one optional embodiment, after initiating the virtual test process, the vehicle control system first reads its current system status parameters, including but not limited to packet loss rate of the network interface, processor workload ratio, and video frame buffer occupancy rate in memory. It then obtains virtual video stream parameters from the scene generator, which may include video frame rate, resolution, and data volume per frame. The vehicle control system compares these parameters with preset stability conditions. If all parameters meet the preset stability conditions, the virtual video stream is determined to have stable transmission capability. The system automatically sets the video injection mode to continuous video injection mode and continuously injects multi-view virtual video streams into the vehicle control system at a preset frequency, ensuring that the control algorithm completes continuous decisions such as lane keeping and forward collision warning under continuous visual input. If any parameter exceeds the limit, the stability conditions are not met, and the vehicle control system immediately switches to trigger-based video injection mode. Only when the scene generator identifies a key video frame will the corresponding frame and its timestamp be injected into the vehicle control system; injection is paused at other times.

[0067] For example, preset stability conditions can be defined as a network packet loss rate of less than 1%, a processor load of less than 65%, and a video frame buffer occupancy rate of less than 70%. The scene generator can identify key video frames as pedestrians crossing the street, red lights turning on, or vehicles braking suddenly in front.

[0068] In this embodiment, the video injection mode is further determined by whether the vehicle control system state parameters and the video stream parameters of the virtual video stream meet preset stability conditions. If the virtual video stream meets the preset stability conditions, the video injection mode is continuous video injection; otherwise, it is triggered video injection. The vehicle control system implements intelligent injection strategy scheduling under different operating loads and resource constraints, improving the stability and resource utilization of the testing process. When system resources are sufficient, the continuous video injection mode is automatically enabled to ensure high-frequency, low-latency synchronization between the video stream and the control closed loop, covering the algorithm response capability under continuous driving scenarios. When system resources are scarce or there is a potential risk of congestion, the system automatically switches to triggered video injection mode, injecting video only at critical event frames, effectively avoiding frame drops, latency jitter, or system stuttering caused by data overload, ensuring the reliability of control decisions and the interpretability of test results.

[0069] Optionally, the system status parameters of the vehicle control system are obtained, including: obtaining the network status parameters, system processing load, and video frame occupancy rate of the vehicle control system, wherein the system processing load is used to represent the occupancy ratio of computing resources in the vehicle control system, and the video frame occupancy rate is used to represent the occupancy ratio of video frames in the memory space of the vehicle control system; and the system status parameters are determined based on the network status parameters, system processing load, and video frame occupancy rate.

[0070] The aforementioned network status parameters may refer to network performance indicators when the vehicle control system communicates with external devices.

[0071] The aforementioned system processing load can refer to the current occupancy rate of computing resources such as processors and graphics processors in the vehicle control system, reflecting the degree of computing power strain on the system when handling tasks such as perception and decision-making.

[0072] The aforementioned video frame occupancy rate refers to the proportion of memory space occupied by data used to cache video frames in the vehicle control system memory. It is used to measure the system's storage resources' capacity to handle video streams and to prevent frame loss or delays due to memory overflow.

[0073] In one optional embodiment, during virtual testing, the vehicle control system collects its network status parameters, system processing load, and video frame occupancy rate in real time. The vehicle control system comprehensively considers the different contributions of these three factors to the overall operational stability. Firstly, network status parameters are used to assess the connectivity and timing consistency of data transmission; system processing load reflects the real-time pressure on the core computing unit; and video frame occupancy rate characterizes the risk of memory cache backlog. By establishing a multi-dimensional collaborative evaluation mechanism, the parameters are dynamically fused based on their influence weights under typical operating conditions. After normalization, a weighted sum is generated to produce system status parameters. These system status parameters reflect the system's combined carrying capacity in data reception, computation processing, and resource caching.

[0074] This application embodiment accurately constructs system status parameters that reflect the overall carrying capacity of the vehicle control system by comprehensively acquiring and integrating real-time operational data from three key dimensions: network status parameters, system processing load, and video frame occupancy rate. This enables multi-dimensional, quantitative, and collaborative evaluation of resource bottlenecks, further improving the comprehensiveness and accuracy of vehicle control system status judgment, avoiding the use of only a single indicator for judgment, and providing a reliable decision-making basis for the automatic switching of subsequent video injection modes.

[0075] Optionally, based on virtual scene parameters and motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated, including: generating an initial virtual video stream of the virtual vehicle from multiple perspectives based on virtual scene parameters and motion state parameters, wherein the initial virtual video stream contains multiple initial video frames; verifying the multiple initial video frames in the initial virtual video stream to obtain a verification result, wherein the verification result is used to indicate whether there are any abnormal video frames in the multiple initial video frames; updating the initial virtual video stream based on the verification result to obtain a virtual video stream.

[0076] The aforementioned multi-view perspective can refer to the multi-image sequence generated synchronously from the simulated viewpoints of multiple cameras configured in the vehicle control system, simulating the input of real vehicle-mounted vision sensors.

[0077] The aforementioned initial virtual video stream can refer to the unprocessed original multi-channel video sequence directly rendered by the simulation engine based on virtual scene parameters and motion state parameters.

[0078] The aforementioned initial video frame can refer to each independent image frame in the initial virtual video stream, corresponding to the multi-view image output at a certain moment, and is the basic unit of the video stream.

[0079] The aforementioned verification results can refer to the judgment conclusions drawn after the initial video frames are checked for integrity, temporal consistency, image quality, and other dimensions, and are used to identify whether there are abnormal video frames.

[0080] The aforementioned abnormal video frames can refer to video frames in the initial virtual video stream that do not conform to the real sensor output specifications, such as image distortion, timestamp errors, missing data, or abnormal frame content.

[0081] The aforementioned virtual video stream can refer to a high-quality multi-channel video stream that has been verified and corrected, with abnormal video frames removed or repaired, meeting the vehicle control system's requirements for the stability, synchronization, and realism of the input video, and can be used for injection testing.

[0082] In one optional embodiment, virtual scene parameters can include road curvature, weather lighting, dynamic obstacle trajectories, and motion state parameters of the virtual vehicle, such as vehicle speed, yaw rate, and pitch angle. The vehicle control system drives a high-fidelity simulation engine based on these virtual scene parameters to generate an initial virtual video stream containing views from multiple cameras (front-view, side-view, and rear-view). Subsequently, the vehicle control system performs multi-dimensional verification on each frame. First, it compares the motion trajectories of the vehicle and environment between adjacent frames using optical flow consistency detection to ensure they conform to physical laws. If a roadside barrier exhibits non-linear jumping or an oncoming vehicle momentarily clips through it in a frame, it is considered an anomaly. Second, it can also verify whether the changes in image brightness and shadows are consistent with the simulated sun angle and cloud cover using a lighting model. If strong reflections or dark spots suddenly appear that do not conform to the light source direction under clear weather conditions, it is marked as an anomaly. Furthermore, key targets within a frame can be identified using a semantic segmentation network. These targets can include pedestrians, traffic lights, lane lines, etc. If the lane lines in a frame are discontinuous or the traffic light status conflicts with the scene's temporal logic (e.g., the green light lasts for far longer than the actual cycle), it is judged as an anomaly. Finally, interpolation analysis is performed on the timestamp sequence. If the time interval of a frame deviates from the preset frame rate, it is marked as a temporal anomaly. Through the joint verification of the above multiple aspects, once any anomaly is triggered, it is automatically marked, and the initial virtual video stream is updated to finally obtain the virtual video stream.

[0083] The aforementioned optical flow consistency detection refers to calculating the pixels in adjacent video frames and comparing whether the motion trajectories of the vehicle and environmental objects maintain physical consistency with the motion state parameters of the virtual vehicle (such as vehicle speed and yaw rate), thereby identifying abnormal artifacts such as non-realistic displacement, jitter, or local motion breakage caused by simulation errors.

[0084] The aforementioned nonlinear jump refers to a non-physical, instantaneous change in position of a static or slowly changing object within consecutive frames.

[0085] The aforementioned instantaneous clipping refers to the momentary geometric overlap of two objects that should be mutually exclusive in a virtual scene, violating physical collision constraints.

[0086] This application embodiment verifies the initial video frame to obtain the verification result of whether there are abnormal video frames in the initial video frame, thereby improving the reliability of virtual test input, avoiding control algorithm misjudgment and perception module jitter caused by abnormal video frames, and enabling subsequent control decisions to be generated based on reliable video input, which greatly enhances the accuracy and effectiveness of test results.

[0087] Optionally, updating the initial virtual video stream based on the verification result to obtain a virtual video stream includes: if the verification result shows that there are abnormal video frames among the multiple initial video frames, removing the abnormal video frames from the initial virtual video stream to obtain a virtual video stream; if the verification result shows that there are no abnormal video frames among the multiple initial video frames, determining that the initial virtual video stream is a virtual video stream.

[0088] In one optional embodiment, when the vehicle control system detects an abnormal frame in the initial virtual video stream, it directly deletes the frame from the sequence and fills the temporal gap with the previous frame or a replacement frame generated through motion interpolation to ensure the continuity of the video stream. Optionally, after removing the abnormal frame, the vehicle control system can automatically generate semantically consistent replacement frames based on the vehicle's motion state and scene parameters, so that the repaired video is visually and logically consistent with the real driving environment. The aforementioned semantically consistent replacement frames refer to image frames that are re-rendered by the simulation engine to conform to real semantic relationships, combining scene logic (such as traffic light cycles, lane line directions, and obstacle trajectories) and lighting models, ensuring that there are no logical conflicts in target categories, states, and spatiotemporal relationships (such as green lights not abruptly turning into red lights, and lane lines not being interrupted), rather than simply performing visual interpolation.

[0089] In another optional embodiment, if there are no abnormal frames, the original stream is directly retained without any modification. The motion interpolation generation described above refers to estimating the position and orientation of the abnormal frame based on the vehicle motion state parameters and image pixel stream of the two frames before and after the abnormal frame through optical flow estimation or rigid body motion model, and synthesizing an intermediate frame image that conforms to physical continuity, thus achieving smooth temporal filling.

[0090] This application embodiment ensures that the injected video stream is stable, continuous, and distortion-free by verifying and eliminating abnormal video frames, thereby improving the input quality of the control system.

[0091] Optionally, a virtual video stream is injected into the vehicle control system based on a video injection mode so that the vehicle control system can generate control decision results. This includes: performing time calibration on the virtual video stream and motion state parameters to obtain calibration results; updating the timestamp of the virtual video stream based on the calibration results to obtain an updated virtual video stream; and injecting the updated virtual video stream into the vehicle control system based on the video injection mode so that the vehicle control system can generate control decision results.

[0092] The aforementioned time calibration can refer to the process of aligning the frame timestamps of the virtual video stream with the time reference of the motion state parameters, ensuring that the visual input and the vehicle's motion state are precisely synchronized on the time axis.

[0093] The calibration results mentioned above can refer to the timing calibration data output after time calibration, which reflects the relative time offset between video frames and motion parameters, and is used to guide timestamp correction.

[0094] The aforementioned timestamps can refer to the time identifier attached to each video frame, used to mark the time of its acquisition or generation. This is a key basis for vehicle control systems to perform multi-sensor fusion and decision-making timing dependence.

[0095] The aforementioned updated virtual video stream may refer to a virtual video stream that has been timestamped, with each frame's timestamp strictly aligned with the motion state parameters to ensure precise time consistency when injected into the control system.

[0096] In one optional embodiment, firstly, the vehicle control system synchronizes the video output of the virtual simulation platform with the vehicle motion state data. A timestamp recorder is used to collect the original timestamps of the virtual video frames and motion parameters, and the time deviation between the two is compared to obtain a calibration result that includes communication delays, processing jitter, and sampling asynchrony. Then, based on the calibration result, the original timestamp of each frame in the virtual video stream is uniformly offset and rewritten to a new timestamp synchronized with the local clock of the vehicle control system, ensuring that each frame corresponds to the actual motion state within the control decision cycle, resulting in an updated virtual video stream. Finally, based on the video injection mode, the updated virtual video stream is injected into the vehicle control system so that the vehicle control system can generate control decision results.

[0097] This application embodiment ensures accurate alignment between the virtual video stream and the vehicle's motion state by synchronizing time calibration and timestamps, eliminating timing deviations between simulation and on-board systems, improving the real-time performance and accuracy of control decisions, effectively avoiding false triggering caused by delays or asynchrony, and enhancing the reliability and engineering effectiveness of test results.

[0098] Figure 2 This is a general structural block diagram of an intelligent driving video injection and verification system based on adaptive multi-mode scheduling according to an embodiment of this application, as shown below. Figure 2As shown, the system mainly includes a host computer for generating test commands and initiating the test process; a scene generator for generating multi-view virtual video streams based on virtual scene parameters; a real camera for acquiring real-world image data for comparison and verification; a simulation board front-end for receiving and decoding the virtual video stream transmitted from the real-time streaming protocol network, while simultaneously performing frame buffering; a video formatter for standardizing and re-encoding the decoded video data according to the target format; a video injection card or field-programmable gate array (FPGA) for timing synchronization, buffering, and multi-channel parallel output of the converted video data; a serializer or physical layer converter for converting the video signal into a physical layer transmission format suitable for the vehicle interface; an Electronic Control Unit (ECU) for receiving the converted video signal and executing perception algorithms and control decisions; a cockpit domain controller for collaboratively processing and displaying non-driving-related visual information; a cockpit display screen for presenting vehicle status and interactive content; and a parallel Controller Area Network (Controller Area Network). The Network (CAN) injection server and data log are used to synchronously inject CAN messages and record test data throughout the system's operation. The scenario generator mentioned above can be a Carla scenario generator. These modules are connected sequentially via wired or wireless communication links according to the data flow, forming a complete closed loop from virtual scenario generation and video data transmission to ECU perception verification, achieving efficient injection and real-time verification of virtual scenario video signals in the intelligent driving domain.

[0099] Figure 3 This is a flowchart illustrating the workflow of an intelligent driving video injection and verification system based on adaptive multi-mode scheduling, according to an embodiment of this application. Figure 3As shown, the overall workflow includes the host computer issuing scene configuration and trigger commands. The scene generator then generates multiple virtual video streams based on these commands, which are transmitted over the network to the simulation board. The simulation board receives and decodes the virtual video streams, inserts timestamps, and caches them. The cached video streams then enter the video format conversion and encapsulation stage. After encapsulation, the data is output by the video injection card and undergoes adaptive injection scheduling. The serializer performs physical layer conversion, and the intelligent driving domain controller receives, perceives, and infers the data. Subsequently, the cockpit domain controller receives relevant data and completes rendering and layout, ultimately displaying the data on the cockpit display screen. After the simulation board completes receiving, decoding, inserting timestamps, and caching, the system sends back and records the results and logs. This process is synchronously performed until the intelligent driving domain controller receives, perceives, and infers the data. Simultaneously, the intelligent driving domain controller injects data into the server, issuing vehicle speed / turning angle, etc. → the intelligent driving domain controller receives, perceives, and infers the data. The aforementioned network transmission can be via Real-Time Streaming Protocol (RTSP) or Real-Time Transport Protocol (RTP). The video format conversion and encapsulation data described above can be transmitted at the physical layer using either the Mobile Industry Processor Interface (MIPI) or Low-Voltage Differential Signaling (LVDS).

[0100] In this embodiment, the host computer serves as the unified management and control center of the vehicle control system, used to configure simulation scenarios, select injection modes, and issue control commands. The host computer sends scenario parameters and control instructions to the scenario generator via Ethernet, such as Transmission Control Protocol (TCP), User Datagram Protocol (UDP), Hypertext Transfer Protocol (HTTP), or a dedicated communication channel. Configuration files or trigger frames are provided, such as those in the lightweight data exchange format JavaScript Object Notation (JSON) and Extensible Markup Language (XML). Simultaneously, it can send the video file to be injected or a real-time stream address (e.g., RTSP / Real-Time Message Transfer Protocol) as needed. Video file formats include, for example, those suitable for offline video injection of MPEG-4 Part 14 (MP4), QuickTime file format supporting multi-track and high-precision timestamps, and raw image data (RAW) sequences. The host computer also has trigger control and mode switching functions, which can adjust the video injection method and synchronization strategy in real time during the test.

[0101] The Carla scene generator operates on a virtual simulation platform, capable of generating diverse and configurable road traffic scenes within a computer environment. Its simulation content includes factors such as vehicles, pedestrians, traffic facilities, road topology, ambient lighting, and weather changes. The scene generator generates video streams from multiple virtual cameras based on instructions from the host computer. Each video stream can be independently configured with resolution (e.g., 720p, 1080p, 4K), frame rate (30fps, 60fps, 120fps), and viewing angle (forward, rearward, side, panoramic). The generated video streams are output in luminance and chromatic aberration color space, Red Green Blue Color Space (RGB), or RAW format, and can be transmitted directly to the simulation board via RTSP / RTP network streaming, file transfer (MP4, RAW), or a high-speed bus. The aforementioned high-speed bus can be the Peripheral Component Interconnect Express (PCIe) standard high-speed bus.

[0102] The simulation board serves as the receiver and preprocessor for video data, decoding, verifying, and buffering video streams from the Carla scene generator or host computer. Internally, the board integrates an RTSP receiver module, a network stream receiver module, a video decoding module, a frame integrity detection module, and a high-speed Double Data Rate Random Access Memory (DDR) buffer. The received video data first undergoes frame header comparison and cyclic redundancy check (CRC) verification to ensure data transmission integrity. Subsequently, timestamps are added, and optional image preprocessing, including cropping, scaling, and distortion correction, is performed according to a preset strategy to adapt to the resolution and optical characteristics requirements of the vehicle control system's input interface. Finally, the processed video frames are buffered in the DDR storage area as a stable, low-jitter data source for subsequent video injection modules to access as needed.

[0103] Depending on the test scenario, the buffer depth of the simulation board can be flexibly adjusted: in trigger mode, only 1 frame is buffered to reduce end-to-end latency; in continuous mode, 4 to 8 frames can be buffered to ensure the stability and continuity of the data stream. The board maintains high-speed interconnection with the video injection card or host computer through PCIe or high-speed Ethernet interfaces to achieve high-bandwidth, low-latency data transmission.

[0104] The video format conversion / encoding unit is located in the middle layer of the data link. It is responsible for converting standard video formats output by the simulation board (such as luminance and chromatic aberration analog color spaces, RGB, and RAW) into interface and protocol formats that the ECU can directly recognize, such as Mobile Industry Processor Interface (MIPI) Camera Serial Interface (CSI), LVDS, and Gigabit Multimedia Serial Link (GMSL). The conversion process includes pixel format rearrangement, layered packaging, frame rate or resolution adjustment, and packetized output. This module can be implemented based on the FPGA's internal processing core or a dedicated video encoding chip, and provides a low-latency path and configurable format interfaces to adapt to ECU input standards from different manufacturers.

[0105] The video injection card / Field-Programmable Gate Array (FPGA) is a core module of the vehicle control system, used to achieve high-speed transmission, timing control, and mode management of video signals. The video injection card features a high-bandwidth interface, high-speed double data rate (DFR) memory buffer, synchronous clock management circuitry, and a multi-channel mobile industry processor interface and camera serial interface output unit. Based on testing requirements, the injection card supports three typical injection modes: the first is trigger mode, where the injection card immediately outputs one or more frames of video from the buffer after a trigger signal is issued by the host computer or an external trigger, suitable for low-latency scenarios; the second is continuous mode, where the injection card outputs a stable video stream at a fixed frame rate, simulating the working state of an onboard camera in master mode; and the third is slave mode, where the injection card responds as a slave device and outputs corresponding video frames when the ECU issues a timing trigger signal, simulating the input of a camera in slave mode.

[0106] Figure 4 This is a block diagram of the internal structure of a video injection card / field-programmable gate array according to an embodiment of this application, such as... Figure 4 As shown, virtual test data is received from the host computer or scene generator via an input interface. The input interface uses a high-speed bus or Ethernet interface, a standard for peripheral component interconnection, to achieve high-bandwidth, low-latency data transmission. The simulation board is equipped with a frame buffer unit, which is based on a high-speed double data rate memory and integrates a queue management module. The synchronization and delay management unit performs spatiotemporal alignment between the timestamp of the virtual video stream and the motion state parameters of the virtual vehicle. The adaptive injection scheduling template dynamically calls bandwidth, video frame rate, and feedback detection from the intelligent driving domain controller. The mode control unit has three modes: trigger mode, continuous mode, and slave mode. The status monitoring collects the operating status information of the vehicle control system and transmits the status information back via the host computer feedback interface. The camera serial interface of the mobile industry processor interface sends data with the target address to the vehicle control system. Finally, the data is output via the serial interface.

[0107] To further enhance the adaptability of vehicle control systems in complex testing environments, an adaptive multi-mode video injection strategy can be adopted. This strategy is executed by the intelligent scheduling module within the injection card, which monitors network bandwidth, frame rate stability, and ECU processing feedback signals in real time. Based on this, it automatically switches the injection working mode or adjusts the buffering strategy to ensure the continuity and real-time performance of video transmission. When a decrease in network bandwidth or an increase in transmission latency is detected, it automatically switches from continuous mode to triggered mode to reduce instantaneous load and ensure stable frame output. When the ECU reports frame loss or buffer overflow signals, it automatically adjusts the frame buffer depth or output frame rate to achieve adaptive flow control. During test scenario or algorithm mode switching, the scheduling module can also automatically perform delay calibration and clock synchronization correction to ensure that video and control signals maintain timing consistency. This adaptive injection mechanism is implemented within the FPGA through programmable logic and a soft-core controller, possessing both hardware real-time response and software strategy management capabilities, ensuring low latency while maintaining good flexibility and scalability.

[0108] Figure 5 This is a flowchart of an adaptive multi-mode video injection strategy according to an embodiment of this application, such as... Figure 5 As shown, the current operating status can be monitored, including bandwidth, frame rate jitter, intelligent driving domain controller feedback signals, and cache utilization. If the status is stable, continuous mode output is maintained. If the status is unstable and a bandwidth drop or latency anomaly is detected, it automatically switches to trigger mode to reduce instantaneous load. If no bandwidth drop or latency anomaly is detected, but the intelligent driving domain controller reports frame drops or overflow, the cache depth or output frame rate is adjusted. If the intelligent driving domain controller does not report frame drops or overflow, but a scene or algorithm change, latency calibration and clock synchronization correction are performed, and the monitoring status is returned to continue executing closed-loop scheduling. If the scene or algorithm does not change, continuous mode output is maintained.

[0109] Through this closed-loop adaptive strategy, the vehicle control system can automatically select the optimal injection mode under different network conditions, load states, and test environments, thereby achieving intelligent scheduling and highly stable output of the video stream.

[0110] To support long-distance and automotive-grade transmission, the MIPI signal output from the injection card is converted into LVDS or coaxial signals by a serializer before being transmitted to the ECU via an onboard cable. The serializer module also provides I2C / SPI interfaces, facilitating initialization configuration, clock synchronization, and data channel management by the host computer or ECU, thereby ensuring signal transmission stability and electromagnetic compatibility.

[0111] As the core algorithm processing unit, the ECU receives the video stream after physical layer conversion and performs perception processing on the video frames. The ECU's internal Artificial Intelligence (AI) inference module can perform tasks such as object detection, lane line recognition, obstacle detection, and traffic sign recognition, outputting corresponding perception results and control commands. During data processing, the ECU synchronously records the timestamps of the input frames and the algorithm inference time, and sends the results back to the host computer or data recording module for performance analysis and accuracy verification.

[0112] The cockpit domain controller and cockpit display screen constitute the human-machine interface of the vehicle control system. The cockpit domain controller is responsible for receiving video output and analysis results from the ECU, and rendering, layout, and switching modes of the screen according to test requirements. The display screen can simultaneously display multiple video sources and perception overlay information (such as detection boxes, recognition tags, trajectory lines, etc.), and supports multiple display modes such as split screen, surround view, and panoramic view, making it convenient for testers to observe the operational effects of Advanced Driver Assistance Systems (ADAS).

[0113] To ensure consistency between the video footage and the vehicle's dynamic behavior, this vehicle control system further incorporates a CAN injection module. This module can inject vehicle status signals, including vehicle speed, wheel speed, steering angle, and braking status, into the ECU in real time based on virtual scene data generated by the scene generator or instructions from the host computer script. Through synchronized video and CAN data injection, the vehicle's driving state can be completely reproduced in the virtual environment, achieving collaborative simulation of vision and dynamics.

[0114] To facilitate data analysis and problem tracing, data recording and playback modules are set up at key nodes (including simulation boards, injection cards, and ECUs). This module can record raw video frames, timestamps, sensing results, control commands, and CAN bus data, and supports playback and comparative analysis functions for algorithm performance evaluation, error statistics, and problem reproduction.

[0115] To further improve the stability and synchronization of the vehicle control system, this invention adopts a unified time synchronization and clock alignment mechanism. For example, it uses Precision Time Protocol (PTP) as the primary synchronization mechanism. In environments without network time synchronization or as a redundant backup, the vehicle control system switches to Network Time Protocol (NTP) to provide unified time synchronization for the host computer, simulation board, injection card, and ECU. A high-precision timestamp is added to each video frame at each node and kept consistent with the CAN bus timeline. Frame integrity detection (sequence number comparison and CRC check), dynamic cache management (adjusting buffer depth according to injection mode), and fault tolerance strategies (frame loss compensation, mode rollback, alarm logs) are implemented in the link to ensure stable operation of the vehicle control system under different network environments and hardware configurations, achieving reliable injection throughout the entire process from virtual scene generation to ECU verification.

[0116] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0117] According to an embodiment of this application, a test device for a vehicle control system is also provided. It should be noted that the device can execute the test method for the vehicle control system described above. The specific implementation method and preferred application scenarios are the same as those in the above embodiments, and will not be repeated here. Figure 6 This is a schematic diagram of a test apparatus for a vehicle control system according to an embodiment of this application, as shown below. Figure 6 As shown, the device includes: a host computer 602, a scene generator 604, a video injection module 606, and a result generation module 608.

[0118] The host computer 602 is used to acquire virtual scene parameters of the virtual test scenario, motion state parameters of the virtual vehicle, and video injection mode of the vehicle control system. The video injection mode represents the injection mode for injecting video into the vehicle control system. The scene generator 604 is used to generate virtual video streams of the virtual vehicle from multiple perspectives based on the virtual scene parameters and motion state parameters. The video injection module 606 is used to inject the virtual video stream into the vehicle control system based on the video injection mode so that the vehicle control system can generate control decision results. The result generation module 608 is used to determine the test results of the vehicle control system based on the control decision results.

[0119] According to an embodiment of this application, a test device for a vehicle control system is also provided. It should be noted that the device can execute the test method for the vehicle control system described above. The specific implementation method and preferred application scenarios are the same as those in the above embodiments, and will not be repeated here. Figure 7 This is a schematic diagram of a test apparatus for a vehicle control system according to an embodiment of this application, as shown below. Figure 7 As shown, the device 700 includes: an acquisition module 702, a generation module 704, an injection module 706, and a determination module 708.

[0120] The acquisition module is used to acquire virtual scene parameters of the virtual test scenario, motion state parameters of the virtual vehicle, and video injection mode of the vehicle control system, wherein the video injection mode is used to represent the injection mode for injecting video into the vehicle control system; the generation module is used to generate a virtual video stream of the virtual vehicle from multiple perspectives based on the virtual scene parameters and the motion state parameters; the injection module is used to inject the virtual video stream into the vehicle control system based on the video injection mode, so that the vehicle control system can generate control decision results; and the determination module is used to determine the test results of the vehicle control system based on the control decision results.

[0121] The acquisition module is also used to determine the video injection mode as master-slave video injection mode when a master-slave video injection instruction is received. In master-slave video injection mode, a virtual video stream is injected into the vehicle control system according to the control signal of the master device, which is the vehicle control system. When no master-slave video injection instruction is received, the module acquires the system status parameters of the vehicle control system and determines the video injection mode of the vehicle control system based on the system status parameters and the video stream parameters of the virtual video stream.

[0122] The acquisition module is further configured to determine, based on system status parameters and video stream parameters, whether the virtual video stream meets the preset stability conditions of the vehicle control system, wherein the preset stability conditions represent the conditions for stable transmission of the virtual video stream to the vehicle control system; if the virtual video stream meets the preset stability conditions, the video injection mode is determined to be a continuous video injection mode, wherein the continuous video injection mode is used to inject the virtual video stream into the vehicle control system according to a preset video injection frequency; if the virtual video stream does not meet the preset stability conditions, the video injection mode is determined to be a triggered video injection mode, wherein the triggered video injection mode is used to inject the virtual video stream into the vehicle control system when a key video frame is detected.

[0123] The acquisition module is also used to acquire network status parameters, system processing load, and video frame occupancy rate of the vehicle control system. The system processing load represents the proportion of computing resources occupied in the vehicle control system, and the video frame occupancy rate represents the proportion of video frames occupied in the memory space of the vehicle control system. Based on the network status parameters, system processing load, and video frame occupancy rate, the system status parameters are determined.

[0124] The generation module is also used to generate an initial virtual video stream of the virtual vehicle from multiple perspectives based on virtual scene parameters and motion state parameters. The initial virtual video stream contains multiple initial video frames. The multiple initial video frames in the initial virtual video stream are verified to obtain a verification result, which is used to indicate whether there are any abnormal video frames among the multiple initial video frames. The initial virtual video stream is updated based on the verification result to obtain a virtual video stream.

[0125] The generation module is also used to remove abnormal video frames from the initial virtual video stream to obtain a virtual video stream if the verification result shows that there are abnormal video frames among the multiple initial video frames; and to determine that the initial virtual video stream is a virtual video stream if the verification result shows that there are no abnormal video frames among the multiple initial video frames.

[0126] The injection module is also used to perform time calibration on the virtual video stream and motion state parameters to obtain calibration results; update the timestamp of the virtual video stream based on the calibration results to obtain the updated virtual video stream; and inject the updated virtual video stream into the vehicle control system based on the video injection mode so that the vehicle control system can generate control decision results.

[0127] Embodiments of this application also provide a vehicle, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods described in various embodiments of this application when it runs.

[0128] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0129] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0130] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0131] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0132] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0133] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0134] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0135] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A test method for a vehicle control system, characterized in that, include: The virtual scene parameters of the virtual test scene, the motion state parameters of the virtual vehicle, and the video injection mode of the vehicle control system are obtained, wherein the video injection mode is used to represent the injection mode for injecting video into the vehicle control system. Based on the virtual scene parameters and the motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated; The virtual video stream is injected into the vehicle control system based on the video injection mode, so that the vehicle control system can generate control decision results; Based on the control decision results, the test results of the vehicle control system are determined.

2. The method according to claim 1, characterized in that, The method further includes: Upon receiving a master-slave mode video injection instruction, the video injection mode is determined to be a master-slave video injection mode, wherein the master-slave video injection mode is used to inject the virtual video stream into the vehicle control system according to the control signal of the master device, and the master device is the vehicle control system. If the master-slave mode video injection instruction is not received, the system status parameters of the vehicle control system are obtained, and the video injection mode of the vehicle control system is determined based on the system status parameters and the video stream parameters of the virtual video stream.

3. The method according to claim 2, characterized in that, Based on the system state parameters and the video stream parameters of the virtual video stream, the video injection mode of the vehicle control system is determined, including: Based on the system state parameters and the video stream parameters, it is determined whether the virtual video stream meets the preset stability conditions of the vehicle control system, wherein the preset stability conditions are used to represent the conditions for stable transmission of the virtual video stream to the vehicle control system; When the virtual video stream meets the preset stability condition, the video injection mode is determined to be a continuous video injection mode, wherein the continuous video injection mode is used to inject the virtual video stream into the vehicle control system according to a preset video injection frequency; If the virtual video stream does not meet the preset stability condition, the video injection mode is determined to be a triggered video injection mode, wherein the triggered video injection mode is used to inject the virtual video stream into the vehicle control system when a key video frame is detected.

4. The method according to claim 2, characterized in that, Obtaining the system state parameters of the vehicle control system includes: The network status parameters, system processing load, and video frame occupancy rate of the vehicle control system are obtained, wherein the system processing load is used to represent the occupancy ratio of computing resources in the vehicle control system, and the video frame occupancy rate is used to represent the occupancy ratio of video frames in the memory space of the vehicle control system. The system status parameters are determined based on the network status parameters, the system processing load, and the video frame occupancy rate.

5. The method according to claim 1, characterized in that, Based on the virtual scene parameters and the motion state parameters, a virtual video stream of the virtual vehicle from multiple perspectives is generated, including: Based on the virtual scene parameters and the motion state parameters, an initial virtual video stream of the virtual vehicle from the multiple perspectives is generated, wherein the initial virtual video stream contains multiple initial video frames; The plurality of initial video frames in the initial virtual video stream are verified to obtain a verification result, wherein the verification result is used to indicate whether there are any abnormal video frames in the plurality of initial video frames; The initial virtual video stream is updated based on the verification result to obtain the virtual video stream.

6. The method according to claim 5, characterized in that, The initial virtual video stream is updated based on the verification result to obtain the virtual video stream, including: If the verification result indicates that there is an abnormal video frame among the plurality of initial video frames, the abnormal video frame in the initial virtual video stream is removed to obtain the virtual video stream. If the verification result indicates that there is no abnormal video frame among the plurality of initial video frames, the initial virtual video stream is determined to be the virtual video stream.

7. The method according to any one of claims 1-6, characterized in that, The virtual video stream is injected into the vehicle control system based on the video injection mode, so that the vehicle control system generates control decision results, including: The virtual video stream and the motion state parameters are time-calibrated to obtain calibration results; The timestamp of the virtual video stream is updated based on the calibration results to obtain the updated virtual video stream; The updated virtual video stream is injected into the vehicle control system based on the video injection mode, so that the vehicle control system can generate control decision results.

8. A testing device for a vehicle control system, characterized in that, include: The host computer is used to generate virtual scene parameters and motion state parameters of the virtual vehicle for the virtual test scenario. A scene generator is used to generate a virtual video stream of the virtual vehicle from multiple perspectives based on the virtual scene parameters and the motion state parameters. A video injection module is used to inject the virtual video stream into the vehicle control system according to a video injection mode; The vehicle control system is used to generate control decision results based on the virtual video stream; The result generation module is used to generate test results for the vehicle control system based on the control decision results.

9. A vehicle, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 7.

Images