Method and device for acquiring automatic driving data of man-machine co-driving vehicle

Abstract

The application provides a kind of automatic driving data acquisition method and device of man-machine co-pilot vehicle, method includes: in man-machine co-pilot mode, determine the starting time of the trigger shadow mode of obtaining vehicle data;Obtain the pose information and the information of the obstacle around the vehicle of real vehicle;With the pose information and the information of the obstacle around the vehicle obtained generate simulation vehicle information, and determine the termination time of shadow mode;Determine whether the simulation vehicle information from starting time to termination time meets the first condition of pre-set, if yes, to the information of real vehicle and simulation vehicle information between starting time and termination time is collected.The application is for the acquisition of simulation vehicle information corresponding to automatic driving test model of man-machine co-pilot mode of vehicle, data credibility is high, can carry out the simulation vehicle information acquisition of automatic driving model with long time state window scene, improve the recognition ability of complex scene of automatic driving test model.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a method and apparatus for acquiring autonomous driving data in a human-machine co-driving vehicle. Background Technology

[0002] Autonomous driving, also known as driverless driving, computer-controlled driving, or wheeled mobile robots, is a cutting-edge technology that relies on computer and artificial intelligence to complete safe and efficient driving without human intervention. However, driverless driving also faces many challenges, such as inclement weather and complex traffic environments, so its performance in certain scenarios may not match that of human drivers. To address this issue, more testing and iterative updates of autonomous driving models are needed to improve their handling capabilities in complex scenarios. Testing and updating autonomous driving models requires mounting them on vehicles to acquire and collect autonomous driving data. Typically, the testing and use of autonomous driving employs a human-machine co-driving approach, where the autonomous driving model controls the vehicle while a human observes the environment and takes over the vehicle as needed based on changes in the environment.

[0003] In current autonomous driving model testing and updates, open-loop testing is often used. In this form, the position within the autonomous driving model algorithm is always consistent with the actual position of the vehicle. This leads to inconsistencies between the position state within the algorithm and the expected state of the vehicle within the algorithm. It can only be applied to autonomous driving models with instantaneous scenarios, but not to autonomous driving models with long-term state window scene recognition functions. For example, it cannot effectively obtain long-term autonomous driving data in human-machine co-driving scenarios before and after the human takes over the vehicle. Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention provides a method and apparatus for acquiring autonomous driving data in human-machine co-driving vehicles.

[0005] This invention provides a method for acquiring autonomous driving data in a human-machine co-driving vehicle, comprising:

[0006] In human-machine co-driving mode, the time when the autonomous driving mode switches to the driver takeover mode is obtained, and the start time of triggering the shadow mode is determined based on the time.

[0007] Obtain real vehicle position and orientation information and information about obstacles around the vehicle;

[0008] The acquired pose information and obstacle information around the vehicle are used to generate simulated vehicle information, and the termination time of the shadow mode is determined.

[0009] Determine whether the simulated vehicle information from the start time to the end time meets the preset first condition. If so, collect the information of the real vehicle and the simulated vehicle between the start time and the end time.

[0010] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle is provided, wherein the first condition is at least one preset danger determination condition for determining whether the simulated vehicle is in danger.

[0011] According to the present invention, a method for acquiring autonomous driving data of a human-machine co-driving vehicle includes, in human-machine co-driving mode, acquiring the time when the autonomous driving mode switches to the driver takeover mode, and determining the start time for acquiring vehicle data based on the time, comprising:

[0012] Based on the vehicle's human-machine co-driving mode, obtain the vehicle scenario corresponding to the time when it switches from autonomous driving mode to driver takeover mode;

[0013] The start time for acquiring vehicle data is determined based on the vehicle's scenario.

[0014] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle includes the step of generating simulated vehicle information using the acquired pose information and surrounding obstacle information, comprising:

[0015] The autonomous driving test model on the vehicle is used to process the pose information of the real vehicle and the obstacle information around the vehicle at the starting moment to determine the simulated vehicle information at the starting moment.

[0016] Real-time acquisition of obstacle information around the real vehicle at multiple times after the starting time, and determination of simulated vehicle information at multiple times after the starting time, including: using the autonomous driving test model to determine the simulated vehicle information at each current time based on the obstacle information around the real vehicle at the current time and the simulated vehicle information at the previous time.

[0017] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle includes the step of determining the termination time of the shadow mode as follows:

[0018] Determine whether the simulated vehicle information at the current moment meets the preset second condition. If yes, continue to acquire the simulated vehicle information at the next moment. Otherwise, take the current moment as the termination moment for vehicle data acquisition and stop acquiring simulated vehicle information at multiple moments after the current moment.

[0019] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle is provided, wherein the second condition is that the simulated vehicle position information in the simulated vehicle information at the current moment is within a preset reliable perception range.

[0020] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle is provided, wherein the simulated vehicle information includes simulated vehicle behavior information and simulated vehicle pose information, and the simulated vehicle behavior information includes simulated vehicle planning information and simulated vehicle control information.

[0021] According to the present invention, a method for acquiring autonomous driving data of a human-machine co-driving vehicle includes the step of processing the pose information and surrounding obstacle information of the real vehicle at the initial moment using an autonomous driving test model mounted on the vehicle to determine the simulated vehicle information at the initial moment, comprising:

[0022] By using the autonomous driving test model on the vehicle, the obstacle information and pose information of the real vehicle at the start time are analyzed to obtain the simulated vehicle planning information at the start time.

[0023] The simulated vehicle planning information at the start time is calculated using the aforementioned autonomous driving test model to obtain the simulated vehicle control information at the start time.

[0024] The autonomous driving test model is used to simulate the vehicle control information at the start time to obtain the simulated vehicle pose information at the start time.

[0025] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle is provided, which utilizes the autonomous driving test model to determine the simulated vehicle information at each current moment based on the obstacle information around the real vehicle at the current moment and the simulated vehicle information at the previous moment, including:

[0026] The autonomous driving test model is used to analyze the obstacle information around the real vehicle at the current moment and the simulated vehicle information at the previous moment to obtain the simulated vehicle planning information at the current moment.

[0027] The simulated vehicle planning information at the current moment is calculated using the aforementioned autonomous driving test model to obtain the simulated vehicle control information at the current moment.

[0028] The autonomous driving test model is used to simulate the control information of the simulated vehicle at the current moment to obtain the pose information of the simulated vehicle at the current moment.

[0029] According to the present invention, an autonomous driving data acquisition method for a human-machine co-driving vehicle, the step of determining the start time and the simulated vehicle information at each current time includes: determining the coordinates of obstacles in the obstacle information around the vehicle using a world coordinate system.

[0030] The present invention also provides an autonomous driving data acquisition device for a human-machine co-driving vehicle, comprising:

[0031] The start time acquisition unit is used to acquire the time when the autonomous driving mode switches to the driver takeover mode in the human-machine co-driving mode, and determine the start time of triggering the shadow mode based on the time.

[0032] The information acquisition unit is used to acquire the actual vehicle's pose information and surrounding obstacle information;

[0033] The simulated vehicle information generation unit is used to generate simulated vehicle information using the acquired pose information and vehicle surrounding obstacle information, and to determine the termination time of the shadow mode.

[0034] The data acquisition unit is used to determine whether the simulated vehicle information at each time after the start time meets the preset first condition. If so, it collects the information of the real vehicle and the simulated vehicle between the start time and the end time.

[0035] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the autonomous driving data acquisition method for human-machine co-driving vehicles as described above.

[0036] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the autonomous driving data acquisition method for human-machine co-driving vehicles as described above.

[0037] The present invention provides a method and apparatus for acquiring autonomous driving data of human-machine co-driving vehicles. It can acquire simulated vehicle information corresponding to the autonomous driving test model for human-machine co-driving mode of autonomous driving vehicles. It has strong applicability. The present invention is based on shadow mode so that the simulated vehicle position of the simulated vehicle information formed by the autonomous driving test model can form a gap with the real vehicle. This enables the planning of the algorithm and control strategy in the autonomous driving test model to complete a self-consistent closed loop. It can be applied to the acquisition of simulated vehicle information of autonomous driving models with long-term state window scene recognition function, and improves the recognition ability of autonomous driving test models for complex scenes. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0039] Figure 1 This is a flowchart illustrating the method for acquiring autonomous driving data in a human-machine co-driving vehicle provided by the present invention.

[0040] Figure 2 This is a schematic diagram of the structure of the autonomous driving test model provided by the present invention;

[0041] Figure 3 This is a schematic diagram of the structure of the autonomous driving data acquisition device for human-machine co-driving vehicles provided by the present invention;

[0042] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0044] like Figure 1 As shown in the figure, an embodiment of the present invention provides a flowchart of an autonomous driving data acquisition method, which includes:

[0045] S1: In human-machine co-driving mode, obtain the time when the autonomous driving mode switches to the driver takeover mode, and determine the start time of triggering the shadow mode based on the time.

[0046] S2: Obtain the actual vehicle's pose information and surrounding obstacle information;

[0047] S3: Generate simulated vehicle information using the acquired pose information and obstacle information around the vehicle, and determine the termination time of the shadow mode;

[0048] S4: Determine whether the simulated vehicle information from the start time to the end time meets the preset first condition. If so, collect the information of the real vehicle and the simulated vehicle between the start time and the end time.

[0049] The vehicle described in this invention is a vehicle with human-machine co-driving capability. Human-machine co-driving refers to the driver and the intelligent system simultaneously sharing control of the vehicle, with driving tasks completed through a combination of human and machine capabilities, while both humans and machines possess the ability to independently complete driving tasks. Specifically, the vehicle itself is equipped with an autonomous driving model for actual autonomous driving control. Passengers in the vehicle can take over at any time based on the vehicle's environment and driving conditions.

[0050] The vehicle is equipped with multiple sensors for collecting information about obstacles around the vehicle, such as one or more of visual sensors, infrared cameras, lidar, millimeter-wave radar, and ultrasonic radar, as well as sensing devices for collecting the vehicle's position, such as a GPS positioning device. Based on the above devices and other control data of the vehicle, the vehicle's pose information and information about obstacles around the vehicle can be obtained.

[0051] Furthermore, after confirming the termination time, it is determined whether the simulated vehicle information from the start time to the termination time meets the preset first condition. If the first condition is met, the information of the real vehicle and the simulated vehicle between the start time and the termination time is collected, and then the process returns to step S1 to re-acquire the start time of the new vehicle data acquisition and enter the next simulated vehicle information acquisition. If the first condition is not met, the process directly returns to step S1 to redetermine the start time of the new vehicle data acquisition and enter the next simulated vehicle information acquisition.

[0052] The acquisition is a data storage or transmission method. For example, the real information of the vehicle and the simulated vehicle information between the start time and the end time can be stored in a pre-set memory on the vehicle as autonomous driving data; or the real information of the vehicle and the simulated vehicle information between the start time and the end time can be stored in a pre-set temporary storage on the vehicle as autonomous driving data, and the real information of the vehicle and the simulated vehicle information between the start time and the end time can be transmitted to a cloud server for processing through a remote communication module.

[0053] The information of the real vehicle refers to the actual data information of the vehicle, such as the real vehicle pose information, real vehicle control information, real vehicle sensor information, and vehicle surrounding obstacle information at each time between the start and end times. The information of the simulated vehicle refers to the simulated vehicle behavior information and simulated vehicle pose information generated by the autonomous driving test model. The above information can be selected and combined as needed.

[0054] In this embodiment, by determining whether to record and collect data between the start and end times based on a preset first condition, the acquisition and collection range of autonomous driving data can be confirmed, effectively reducing the amount of autonomous driving data that needs to be collected. Collecting only data that meets the first condition can improve data reliability and reduce data transmission costs.

[0055] The autonomous driving data acquisition method for human-machine co-driving vehicles provided by this invention can acquire simulated vehicle information corresponding to the autonomous driving test model for human-machine co-driving mode of autonomous driving vehicles. It has strong applicability. Based on the shadow mode, this invention enables the simulated vehicle position of the simulated vehicle information formed by the autonomous driving test model to form a gap with the real vehicle, so that the planning and control strategies of the algorithm in the autonomous driving test model can complete a self-consistent closed loop. It can be applied to the acquisition of simulated vehicle information of autonomous driving models with long-term state window scene recognition function, and improves the recognition ability of autonomous driving test models for complex scenes.

[0056] In an alternative embodiment, the first condition is at least one preset danger determination condition for determining whether the simulated vehicle is in danger.

[0057] In this embodiment, a dangerous behavior condition library can be formed through multiple preset danger judgment conditions to determine whether a danger has occurred. Specifically, the dimensions of the danger judgment conditions include safety conditions, such as whether a collision has occurred or whether there is a risk of collision; tactile conditions, such as whether there is heavy braking, sudden turns, or sharp turns; traffic regulation conditions, such as whether there is speeding or running a red light; and unreasonable acceleration, deceleration, or lane changing. For example, one danger judgment condition is that the vehicle has collided. If the simulated vehicle has collided, then this danger judgment condition is met, and the simulated vehicle is determined to be in danger. Another example is that another danger judgment condition is that the vehicle is speeding. If the simulated vehicle speeds, then this danger judgment condition is met, and the simulated vehicle is determined to be in danger. To avoid redundancy, not all examples will be listed. Those skilled in the art can determine whether the simulated vehicle is in danger based on the possible dangerous situations that the vehicle may encounter, thus forming danger judgment conditions. In addition, pre-trained or preset judgment models can also be used to determine whether the simulated vehicle meets the danger judgment conditions.

[0058] In this embodiment, a first condition is used to determine whether a dangerous situation has occurred in the simulated vehicle behavior information acquired by the autonomous driving test model. If a dangerous situation occurs in the simulated vehicle, it indicates that the human intervention time was a reasonable intervention, which can avoid the occurrence of dangerous behavior, avoid collecting invalid intervention data, improve data reliability, and reduce data collection volume and cost. In an optional embodiment, in the human-machine co-driving mode, the time when the autonomous driving mode switches to the driver intervention mode is acquired, and the start time for acquiring vehicle data is determined based on the time, including:

[0059] Based on the vehicle's human-machine co-driving mode, obtain the vehicle scenario corresponding to the time when the autonomous driving mode switches to the driver takeover mode;

[0060] The start time for acquiring vehicle data is determined based on the vehicle's scenario.

[0061] Specifically, in this embodiment, a pre-trained scene analyzer can be used to analyze the radar information and / or visual information acquired by the vehicle to obtain the vehicle scene corresponding to the moment when the autonomous driving mode switches to the driver takeover mode.

[0062] Specifically, in this embodiment, the starting time for acquiring vehicle data can be determined using a pre-trained starting time acquisition model based on the vehicle's scenario.

[0063] For example, the input data for the start time acquisition model is the vehicle's scene information, and the output is the starting offset. Since the start time may not be the same as the current scene's current time, the start time acquisition model will offset the current scene's current time based on the output starting offset to form the start time, and then use this start time to generate simulated vehicle information. For example, if the time of switching from autonomous driving mode to driver takeover mode is time T, the scene analyzer processes the vehicle information at time T to obtain the scene corresponding to time T. Then, the pre-trained start time acquisition model processes the scene corresponding to time T to obtain the starting offset t, so the start time is T+t, where t can be a negative time value. In this way, different start times are generated for different switching scenarios to simulate vehicle information generation.

[0064] In this embodiment, since the positions of the real vehicle and the simulated vehicle may become disconnected, the positions of the real vehicle and the simulated vehicle are synchronized by generating a start time. Data is specifically acquired at the moment when the autonomous driving mode switches to the driver takeover mode, which improves the effectiveness of the data and facilitates the acquisition of subsequent simulated vehicle information.

[0065] In an optional embodiment, the step of generating simulated vehicle information using the acquired pose information and vehicle-surrounding obstacle information includes:

[0066] The autonomous driving test model on the vehicle is used to process the pose information of the real vehicle and the obstacle information around the vehicle at the starting moment to determine the simulated vehicle information at the starting moment.

[0067] Real-time acquisition of obstacle information around the real vehicle at multiple times after the starting time, and determination of simulated vehicle information at multiple times after the starting time, including: using the autonomous driving test model to determine the simulated vehicle information at each current time based on the obstacle information around the real vehicle at the current time and the simulated vehicle information at the previous time.

[0068] In this embodiment, the autonomous driving test model is mounted on the vehicle. For example, it can be mounted on a dedicated testing processing module on the vehicle. This processing module is communicatively connected to various components on the vehicle and can acquire the vehicle's pose information and surrounding obstacle information. Alternatively, it can be mounted on other processing modules on the vehicle with sufficient computing power. The mounting is a general representation of a technical form, and the form of mounting the autonomous driving test model on the vehicle is not limited to the two examples mentioned above.

[0069] Furthermore, the autonomous driving test model and the aforementioned autonomous driving model can be the same autonomous driving model, or they can be two different autonomous driving models.

[0070] To facilitate understanding, the meanings of "real vehicle" and "simulated vehicle" in this embodiment are further explained: A real vehicle is the vehicle itself, the vehicle used for acquiring autonomous driving data. This vehicle can be, but is not limited to, cars, buses, and freight vehicles. "Real" specifically means that the vehicle is a physical vehicle, and the acquired pose information and surrounding obstacle information of the real vehicle are information about the actual vehicle. A simulated vehicle is a vehicle generated by the autonomous driving test model mounted on the vehicle, based on the pose information and surrounding obstacle information of the real vehicle, and after planning, control, and dynamic model analysis using autonomous driving algorithms. This simulated vehicle is not a physical vehicle, but rather an integration of vehicle information generated by the autonomous driving test model.

[0071] The autonomous driving test model described in this embodiment refers to a model capable of generating vehicle behavior planning and control based on vehicle sensor information. In this embodiment, the autonomous driving test model also incorporates a vehicle dynamics model, which can generate corresponding simulated vehicle pose information based on simulated vehicle control information. Correspondingly, since the vehicle is controlled by a human or the autonomous driving model, there will be deviations between the control of the real vehicle and the simulated vehicle. During the acquisition of simulated vehicle information, the position of the simulated vehicle will deviate from that of the real vehicle in terms of position and attitude. In addition, in this embodiment of the invention, pose information includes the vehicle's position information and attitude information.

[0072] The autonomous driving data acquisition method for human-machine co-driving vehicles provided by this invention can process the pose information and surrounding obstacle information of the real vehicle at the initial moment using the autonomous driving test model on the vehicle to obtain the simulated vehicle information at the initial moment; and obtain the simulated vehicle information at multiple moments after the initial moment in chronological order. This invention enables the simulated vehicle position in the simulated vehicle information formed by the autonomous driving test model to form a gap with the real vehicle, so that the planning and control strategies of the algorithm in the autonomous driving test model can complete a self-consistent closed loop. It can be applied to the acquisition of simulated vehicle information of autonomous driving models with long-term state window scene recognition function, and improves the recognition ability of autonomous driving test models for complex scenes.

[0073] In one optional embodiment, the step of determining the termination time of the shadow pattern includes:

[0074] Determine whether the simulated vehicle information at the current moment meets the preset second condition. If yes, continue to acquire the simulated vehicle information at the next moment. Otherwise, take the current moment as the termination moment for vehicle data acquisition and stop acquiring simulated vehicle information at multiple moments after the current moment.

[0075] In this embodiment, a stopping condition is set for the acquisition of simulated vehicle information using the second condition. This allows the simulation calculation to stop at the next moment when the simulated vehicle information does not meet the second condition, and the termination time is confirmed, thereby reducing the computing power consumption and data volume of the autonomous driving test model.

[0076] In one optional embodiment, the second condition is that the simulated vehicle location information in the simulated vehicle information at the current moment is within a preset reliable perception range.

[0077] Specifically, the reliable perception range is the reliable location range of obstacle information around the vehicle in the autonomous driving test model. This reliable perception range moves with the real vehicle, which is always within it. Specifically, it can be a spatial range where the horizontal distance from the real vehicle is less than a preset distance threshold. This spatial range can be understood as the range formed by drawing a circle with the real vehicle as the center and the preset distance threshold as the radius, or it can be a rectangular spatial range where the lateral distance from the real vehicle is less than a preset lateral distance threshold and the longitudinal distance is less than a preset longitudinal distance threshold.

[0078] Because the position of the simulated vehicle can deviate from that of the real vehicle, but the obstacle information around the simulated vehicle used to generate the simulated vehicle information is obtained from the real vehicle, when the simulated vehicle position information in the current moment exceeds the reliable perception range, it leads to a significant distortion in the behavior and perception output regarding obstacles on the road, resulting in unreliable simulated vehicle information generated by the autonomous driving test model. Therefore, by setting the second condition in this embodiment, using the distance relationship between the real vehicle position and the simulated vehicle position to determine whether the simulated vehicle information generated by the autonomous driving test model at this time is reliable, the reliability of the simulated vehicle information can be effectively reduced, and the amount of data and computing power consumed can be reduced.

[0079] In one optional embodiment, the simulated vehicle information includes simulated vehicle behavior information and simulated vehicle pose information, wherein the simulated vehicle behavior information includes simulated vehicle planning information and simulated vehicle control information.

[0080] The step of using an autonomous driving test model mounted on the vehicle to process the pose information and surrounding obstacle information of the real vehicle at the initial moment to determine the simulated vehicle information at the initial moment includes: analyzing the surrounding obstacle information and pose information of the real vehicle at the initial moment using the autonomous driving test model mounted on the vehicle to obtain the simulated vehicle planning information at the initial moment; calculating the simulated vehicle planning information at the initial moment using the autonomous driving test model to obtain the simulated vehicle control information at the initial moment; and simulating the simulated vehicle control information at the initial moment using the autonomous driving test model to obtain the simulated vehicle pose information at the initial moment.

[0081] The autonomous driving test model uses the obstacle information around the real vehicle at the current moment and the simulated vehicle information from the previous moment to determine the simulated vehicle information at each current moment, including:

[0082] The autonomous driving test model is used to analyze the obstacle information around the real vehicle at the current moment and the simulated vehicle information at the previous moment to obtain the simulated vehicle planning information at the current moment; the autonomous driving test model is used to calculate the simulated vehicle planning information at the current moment to obtain the simulated vehicle control information at the current moment; the autonomous driving test model is used to simulate the simulated vehicle control information at the current moment to obtain the simulated vehicle pose information at the current moment.

[0083] The autonomous driving test model in this embodiment is for ease of understanding, as follows: Figure 2 As shown, it includes a planning module, a control module, and a dynamic model. The planning module can analyze the obstacle information around the vehicle and the vehicle pose information to obtain the simulated vehicle planning information. The control module can calculate the simulated vehicle pose information to obtain the simulated vehicle control information. The dynamic model is used to obtain the simulated vehicle pose information based on the simulated vehicle control information.

[0084] The autonomous driving test model in this embodiment can process the obstacle information around the real vehicle and the real vehicle's pose information at the initial moment to obtain the simulated vehicle's pose information at the initial moment step by step. Then, it obtains the simulated vehicle information at the current moment step by step based on the obstacle information around the real vehicle at the current moment and the simulated vehicle information at the previous moment. At each moment after the initial moment, the simulated vehicle pose information generated at the previous moment can be introduced to form a position difference between the simulated vehicle and the real vehicle. This enables the identification of autonomous driving strategies and problem scenarios with long-term state windows and is applicable to data acquisition for L3 / L4+ level autonomous driving algorithms. The simulated vehicle information generated in this embodiment provides a data foundation for subsequent collection, is applicable to complex scenarios, and has a more powerful and comprehensive ability to identify problems.

[0085] In an optional embodiment, determining the start time and the simulated vehicle information at each current time includes: determining the coordinates of obstacles in the vehicle perimeter obstacle information using a world coordinate system.

[0086] Specifically, since the positions of the real vehicle and the simulated vehicle are separated, it is necessary to convert the obstacle information acquired by the real vehicle into world coordinates to facilitate the autonomous driving test model's position confirmation and processing, and to facilitate the acquisition of simulated vehicle information. In this embodiment, the coordinates of the obstacles used in the autonomous driving model used to control the operation of the real vehicle are coordinates in the vehicle body coordinate system.

[0087] The following describes the autonomous driving data acquisition device for human-machine co-driving vehicles provided by the present invention. The autonomous driving data acquisition device for human-machine co-driving vehicles described below can be referred to in correspondence with the autonomous driving data acquisition method for human-machine co-driving vehicles described above.

[0088] The present invention provides an autonomous driving data acquisition device for human-machine co-driving vehicles, such as... Figure 3 As shown, it includes:

[0089] The start time acquisition unit 31 is used to acquire the time when the autonomous driving mode switches to the driver takeover mode in the human-machine co-driving mode, and determine the start time of triggering the shadow mode based on the time.

[0090] Information acquisition unit 32 is used to acquire the actual vehicle's pose information and surrounding obstacle information;

[0091] The simulated vehicle information generation unit 33 is used to generate simulated vehicle information using the acquired pose information and vehicle surrounding obstacle information, and to determine the termination time of the shadow mode.

[0092] The data acquisition unit 34 is used to determine whether the simulated vehicle information at each time after the start time meets the preset first condition. If so, it collects the information of the real vehicle and the simulated vehicle between the start time and the end time.

[0093] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logic instructions in the memory 430 to execute an autonomous driving data acquisition method for a human-machine co-driving vehicle. This method includes: acquiring the start time of vehicle data acquisition when the autonomous driving mode switches to driver takeover mode, based on the vehicle's human-machine co-driving mode; acquiring the pose information and surrounding obstacle information of the real vehicle at the start time, and acquiring surrounding obstacle information at multiple subsequent times; processing the pose information and surrounding obstacle information of the real vehicle at the start time using an autonomous driving test model mounted on the vehicle to obtain simulated vehicle information at the start time; and acquiring simulated vehicle information at multiple times after the start time in chronological order. The step of acquiring simulated vehicle information at multiple times after the start time includes: processing the surrounding obstacle information of the real vehicle at the current time and the simulated vehicle information at the previous time using the autonomous driving test model to obtain simulated vehicle information at the current time.

[0094] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a 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 the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0095] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the autonomous driving data acquisition method for human-machine co-driving vehicles provided by the above methods. The method includes: in human-machine co-driving mode, acquiring the time when the autonomous driving mode switches to the driver takeover mode, and determining the start time of triggering the shadow mode based on the time; acquiring the pose information of the real vehicle and the obstacle information around the vehicle; generating simulated vehicle information using the acquired pose information and obstacle information around the vehicle; and determining the end time of the shadow mode; judging whether the simulated vehicle information at each time after the start time meets a preset first condition; if so, collecting the information of the real vehicle and the simulated vehicle information between the start time and the end time.

[0096] On another front, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements an autonomous driving data acquisition method for a human-machine co-driving vehicle provided by the methods described above. This method includes: in human-machine co-driving mode, acquiring the time when the autonomous driving mode switches to a driver takeover mode, determining the start time of triggering a shadow mode based on the time, acquiring the pose information of the real vehicle and surrounding obstacle information, generating simulated vehicle information using the acquired pose information and surrounding obstacle information, and determining the end time of the shadow mode; determining whether the simulated vehicle information at each time after the start time meets a preset first condition, and if so, collecting the information of the real vehicle and the simulated vehicle information between the start time and the end time. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0097] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic driving data acquisition method of a man-machine co-driving vehicle, characterized by, The method comprises the following steps: In the man-machine co-driving mode, the time when the automatic driving mode is switched to the driver takeover mode is acquired, and the starting time of triggering the shadow mode is determined based on the time; Obtain the pose information and the surrounding obstacle information of the real vehicle; Generate simulated vehicle information using the obtained pose information and surrounding obstacle information, and determine the termination time of the shadow mode; Determine whether the simulated vehicle information from the starting time to the termination time meets a preset first condition, and if so, collect the information of the real vehicle and the simulated vehicle information between the starting time and the termination time; The step of generating simulated vehicle information using the obtained pose information and surrounding obstacle information comprises: Using the automatic driving test model carried by the vehicle to process the pose information and surrounding obstacle information of the real vehicle at the starting time to determine the simulated vehicle information at the starting time; Real-time acquisition of the surrounding obstacle information of the real vehicle at multiple time points after the starting time, and determination of the simulated vehicle information at multiple time points after the starting time, comprising: using the automatic driving test model to determine the simulated vehicle information at each current time based on the surrounding obstacle information of the real vehicle at the current time and the simulated vehicle information at the previous time; The simulated vehicle information includes simulated vehicle behavior information and simulated vehicle pose information, and the simulated vehicle behavior information includes simulated vehicle planning information and simulated vehicle control information. 2.The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, wherein The first condition is at least one preset dangerous condition for determining whether the simulated vehicle is in danger. 3.The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, wherein, In the man-machine co-driving mode, the time when the automatic driving mode is switched to the driver takeover mode is acquired, and the starting time of acquiring vehicle data is determined based on the time, comprising: According to the man-machine co-driving mode of the vehicle, the scene of the vehicle corresponding to the time when the automatic driving mode is switched to the driver takeover mode is acquired; Determine the starting time of acquiring vehicle data based on the scene of the vehicle. 4.The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, wherein, The step of determining the termination time of the shadow mode comprises: Determine whether the simulated vehicle information at the current time meets a preset second condition, if so, continue to acquire the simulated vehicle information at the next time, otherwise, take the current time as the termination time of data acquisition of the vehicle, and stop acquiring the simulated vehicle information at multiple time points after the current time.

5. The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 4, characterized in that, The second condition is that the simulated vehicle position information in the simulated vehicle information at the current time is within a preset trusted perception range. 6.The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, wherein, The step of using the automatic driving test model carried by the vehicle to process the pose information and surrounding obstacle information of the real vehicle at the starting time to determine the simulated vehicle information at the starting time comprises: Using the automatic driving test model carried by the vehicle to analyze the surrounding obstacle information and the pose information of the real vehicle at the starting time to obtain the simulated vehicle planning information at the starting time; Using the automatic driving test model to calculate the simulated vehicle planning information at the starting time to obtain the simulated vehicle control information at the starting time; Using the automatic driving test model to simulate the simulated vehicle control information at the starting time to obtain the simulated vehicle pose information at the starting time.

7. The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, characterized in that, Determine the simulation vehicle information of each current time based on the real vehicle's vehicle surrounding obstacle information of the current time and the simulation vehicle information of the last time by using the automatic driving test model, including: Analyze the real vehicle's vehicle surrounding obstacle information of the current time and the simulation vehicle information of the last time by using the automatic driving test model, and obtain the simulation vehicle planning information of the current time; Calculate the simulation vehicle planning information of the current time by using the automatic driving test model, and obtain the simulation vehicle control information of the current time; Simulate the simulation vehicle control information of the current time by using the automatic driving test model, and obtain the simulation vehicle pose information of the current time. 8.The automatic driving data acquisition method of a man-machine co-driving vehicle according to claim 1, wherein, The determination of the starting time and the simulation vehicle information of each current time respectively includes: determining the coordinates of the obstacles in the vehicle surrounding obstacle information by using the world coordinate system.

9. An automatic driving data acquisition device for a vehicle with human co-pilot, characterized by, Including: A starting time acquisition unit is configured to acquire a time when the automatic driving mode is switched to the driver takeover mode in the man-machine co-driving mode, and determine a starting time of the shadow mode based on the time; An information acquisition unit is configured to acquire the pose information and the vehicle surrounding obstacle information of the real vehicle; A simulation vehicle information generation unit is configured to generate the simulation vehicle information by using the acquired pose information and vehicle surrounding obstacle information, and determine a termination time of the shadow mode; A data acquisition unit is configured to determine whether the simulation vehicle information of each time after the starting time meets a preset first condition, and if so, acquire the information of the real vehicle and the simulation vehicle information between the starting time and the termination time; The step of generating the simulation vehicle information by using the acquired pose information and vehicle surrounding obstacle information includes: Processing the pose information and the vehicle surrounding obstacle information of the real vehicle at the starting time by using the automatic driving test model carried by the vehicle to determine the simulation vehicle information of the starting time; Real-time acquisition of the vehicle surrounding obstacle information of the real vehicle at multiple times after the starting time, and determination of the simulation vehicle information of the multiple times after the starting time, including: determining the simulation vehicle information of each current time based on the real vehicle's vehicle surrounding obstacle information of the current time and the simulation vehicle information of the last time by using the automatic driving test model; The simulation vehicle information includes simulation vehicle behavior information and simulation vehicle pose information, and the simulation vehicle behavior information includes simulation vehicle planning information and simulation vehicle control information.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the automatic driving data acquisition method of the man-machine co-driving vehicle according to any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the automatic driving data acquisition method of the man-machine co-driving vehicle according to any one of claims 1 to 8.

Images