Automatic driving simulation test system and method based on real vehicle data

Abstract

The invention provides an automatic driving simulation test system based on real vehicle data, and the system comprises a middle interface layer which is used for receiving a drive test data packet and version information of a measured algorithm, transmitting the drive test data packet to a simulation engine module according to the version information, receiving data outputted by the simulation engine module, and transmitting the data to a measured algorithm module; the simulation engine module is used for receiving the drive test data packet, playing back the drive test data packet according to a simulation engine algorithm and positioning and chassis data information calculated by the vehicle dynamics module, and inputting the drive test data packet to the measured algorithm module through the middle interface layer; the to-be-measured algorithm module is used for performing problem restoration reproduction, repair verification and conventional scene test on the algorithm according to the positioning and chassis data information calculated by the vehicle dynamics module, forming a new version algorithm and sending the new version algorithm to the intermediate interface layer; and the vehicle dynamics module is used for calculating positioning and chassis data information.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to an autonomous driving simulation test system and method based on real vehicle data. Background Art

[0002] With the development of science and technology, autonomous driving algorithms are becoming more and more mature, but in order to ensure safety and reliability, simulation tests are needed. Existing simulation test solutions can usually only define simple and specific simulation scenarios based on scenario files. The movement trajectory of traffic vehicles in the scenario is generated by simulation software calculations, which is different from the actual movement trajectory of traffic vehicles and has a high degree of difference from human driving characteristics, resulting in poor simulation test results, low efficiency and high cost. Summary of the invention

[0003] In view of this, the present invention provides an automatic driving simulation test system and method based on real vehicle data to solve the technical problems of poor simulation test effect, low efficiency and high cost in the prior art.

[0004] The present invention provides an automatic driving simulation test system based on real vehicle data, the system comprising: an intermediate interface layer, used for receiving a road test data packet and version information of a tested algorithm, sending the road test data packet to a simulation engine module according to the version information, receiving data output by the simulation engine module, and sending it to the tested algorithm module; the simulation engine module, connected to the intermediate interface layer, used for receiving the road test data packet, playing back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputting the data to the tested algorithm module through the intermediate interface layer, the vehicle dynamics module calculation function is as follows:

[0005]

[0006] in, are the vehicle coordinates, is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle; the algorithm module under test is connected to the intermediate interface layer, and is used to restore and reproduce the problem, repair and verify the algorithm, and perform routine scenario testing based on the positioning and chassis data information calculated by the vehicle dynamics module, so as to form a new version of the algorithm and send it to the intermediate interface layer; the vehicle dynamics module is respectively connected to the intermediate interface layer and the algorithm module under test to calculate the positioning and chassis data information.

[0007] Furthermore, the system also includes: a container encapsulation module, which is used to encapsulate the intermediate interface layer, the simulation engine module and the algorithm module under test, generate a corresponding container according to the simulation requirement parameters sent by the simulation scheduling platform, implement multiple data parallel playback simulations in the container, and feed back the simulation results to the simulation scheduling platform; the simulation scheduling platform is connected to the container encapsulation module, and is used to send the simulation requirement parameters to the container encapsulation module according to the simulation task.

[0008] Furthermore, the road test data packet includes a problem packet generated when a real vehicle encounters a problem, and a scenario packet collected in a specific scenario.

[0009] The present invention also provides an automatic driving simulation test method based on real vehicle data, the method comprising: step 1, the intermediate interface layer obtains version information of the road test data packet and the version information of the algorithm under test, and sends the road test data packet to the simulation engine module according to the version information; step 2, the simulation engine module plays back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputs the positioning and chassis data information to the algorithm module under test through the intermediate interface layer, and the vehicle dynamics module calculation function is as follows:

[0010] in, are the vehicle coordinates, is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle; Step 3, the algorithm module under test performs problem restoration and reproduction, repair verification and conventional scenario testing on the algorithm under test according to the positioning and chassis data information calculated by the vehicle dynamics module, generates a new version of the algorithm under test, and feeds back to the intermediate interface layer; Step 4, the simulation engine module receives the new version of the algorithm under test through the intermediate interface layer, performs simulation tasks, and calculates and outputs simulation results.

[0011] Furthermore, the method also includes: step 0, the container encapsulation module encapsulates the intermediate interface layer, the simulation engine module, the vehicle dynamics module and the measured algorithm module, and the simulation scheduling platform sends the simulation requirement parameters to the container encapsulation module according to the simulation task for subsequent playback simulation.

[0012] Furthermore, the method also includes: step 5, the container encapsulation module feeds back the simulation result to the simulation scheduling platform.

[0013] Furthermore, the method of sending the road test data packet to the simulation engine module according to the version information includes: step 11, the intermediate interface layer reads the version information from the road test data packet and receives the corresponding data information in real time; step 12, according to the preset simulation engine model interface protocol, converting the data information into a fixed format; step 13, sending the converted data information to the simulation engine module for communication.

[0014] Furthermore, the method for replaying the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module by the simulation engine module includes: step 21, obtaining the relative information relationship between the main vehicle and the target vehicle in the problem package and the position relationship of the main vehicle in the simulation environment; step 22, projecting the target vehicle to the position relationship under the main vehicle coordinates in the simulation, thereby retaining the target vehicle motion information in the real vehicle data packet; step 23, simulating the motion trajectory of the target vehicle in the actual environment.

[0015] Furthermore, the method for the vehicle dynamics module to calculate the positioning and chassis data information includes: step 31, collecting real vehicle motion data information; step 32, combining the kinematic model and the dynamic model, and fitting them with the collected data; step 33, comparing and calibrating the vehicle data in the simulation with the fitted data to minimize the difference between the two, so that the problem scenario can be reproduced when the data is played back.

[0016] The present invention provides an automatic driving simulation test system and method based on real vehicle data. The technical solution collects road test data packets, and forwards the data in the simulation data packets to the simulation engine module after data conversion at the intermediate interface layer. The simulation engine module forwards the positioning, chassis and other information to the algorithm under test after conversion calculation. After the algorithm under test receives the information, it updates the control instructions to the vehicle dynamics module. The vehicle dynamics module further updates the state information of the vehicle in the simulation through logical calculation, and realizes the closed-loop simulation of the entire algorithm under test. The technical solution can use this simulation system to repair the problem scenarios of the automatic driving system in the real vehicle scenario, and use conventional scenarios for test verification, accelerate the iteration of the automatic driving system, and save costs while meeting the test requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the framework of an automatic driving simulation test system based on real vehicle data provided by the present invention;

[0018] Figure 2 It is a flow chart of an automatic driving simulation test method based on real vehicle data provided by the present invention;

[0019] Figure 3 It is a flow chart of another automatic driving simulation test method based on real vehicle data provided by the present invention;

[0020] Figure 4 It is a flow chart of a method for sending a drive test data packet to a simulation engine module according to version information provided by the present invention;

[0021] Figure 5 It is a flow chart of a method for playing back positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module according to the simulation engine module provided by the present invention;

[0022] Figure 6 It is a flow chart of a method for calculating positioning and chassis data information by a vehicle dynamics module provided by the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Embodiment 1

[0025] The present invention provides an automatic driving simulation test system and method based on real vehicle data. Figure 1 As shown, the system includes an intermediate interface layer, a simulation engine module, a tested algorithm module and a vehicle dynamics module. Figure 2 As shown, the method steps are as follows.

[0026] Step 1, the intermediate interface layer obtains version information of the drive test data packet and the version information of the algorithm under test, and sends the drive test data packet to the simulation engine module according to the version information;

[0027] The intermediate interface layer is used to receive the drive test data packet and the version information of the algorithm under test, send the drive test data packet to the simulation engine module according to the version information, receive the data output by the simulation engine module, and send it to the algorithm under test module.

[0028] Step 2: The simulation engine module plays back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputs the positioning and chassis data information to the algorithm module under test through the intermediate interface layer. The main calculation functions of the vehicle dynamics module are as follows: in, is the vehicle coordinate, φ is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle;

[0029] The simulation engine module is connected to the intermediate interface layer and is used to receive the road test data packets, play back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and input it to the algorithm module under test through the intermediate interface layer.

[0030] The vehicle dynamics module is connected to the intermediate interface layer and the measured algorithm module respectively, and is used to calculate the positioning and chassis data information.

[0031] Step 3: The algorithm module under test performs problem restoration and reproduction, repair verification, and routine scenario testing on the algorithm under test based on the positioning and chassis data information calculated by the vehicle dynamics module, generates a new version of the algorithm under test, and feeds it back to the intermediate interface layer;

[0032] The algorithm module under test is connected to the middle interface layer and is used to restore and reproduce the problem, verify the repair, and conduct routine scenario testing on the algorithm based on the positioning and chassis data information calculated by the vehicle dynamics module, and form a new version of the algorithm and send it to the middle interface layer. The algorithm under test refers to the planning and control algorithm in the autonomous driving algorithm module.

[0033] Step 4: The simulation engine module receives the new version of the algorithm under test through the intermediate interface layer, performs the simulation task, and calculates and outputs the simulation results.

[0034] The present invention provides an automatic driving simulation test system and method based on real vehicle data. The technical solution collects road test data packets, and forwards the data in the simulation data packets to the simulation engine module after data conversion at the intermediate interface layer. The simulation engine module forwards the positioning, chassis and other information to the algorithm under test after conversion calculation. After the algorithm under test receives the information, it updates the control instructions to the vehicle dynamics module. The vehicle dynamics module further updates the state information of the vehicle in the simulation through logical calculation, and realizes the closed-loop simulation of the entire algorithm under test. The technical solution can use this simulation system to repair the problem scenarios of the automatic driving system in the real vehicle scenario, and use conventional scenarios for test verification, accelerate the iteration of the automatic driving system, and save costs while meeting the test requirements.

[0035] Embodiment 2

[0036] The present invention provides an automatic driving simulation test system and method based on real vehicle data. Figure 1 As shown, the system includes an intermediate interface layer, a simulation engine module, a tested algorithm module and a vehicle dynamics module. Figure 3 As shown, the method steps are as follows.

[0037] Step 0: The container encapsulation module encapsulates the intermediate interface layer, the simulation engine module, the vehicle dynamics module and the measured algorithm module. The simulation scheduling platform sends the simulation requirement parameters to the container encapsulation module according to the simulation task for subsequent playback simulation.

[0038] The container encapsulation module is used to encapsulate the intermediate interface layer, the simulation engine module and the algorithm module under test, generate the corresponding container according to the simulation requirement parameters sent by the simulation scheduling platform, implement multiple data parallel playback simulation in the container, and feed back the simulation results to the simulation scheduling platform. The container encapsulation module mainly encapsulates each module into an image, generates the corresponding container according to the parameters sent by the platform, and implements multiple data parallel playback simulation in the container. The container encapsulation module is used to quickly deploy and concurrently use each module in the entire simulation. Each module in the above simulation is ultimately an executable file. Encapsulating it together with the environment and executable file into the image can realize the rapid deployment of the entire system on the server, and according to the protocol parameters passed in by the simulation scheduling platform, multiple simulation requirements can be concurrently implemented to improve the efficiency of simulation testing.

[0039] Step 1, the intermediate interface layer obtains version information of the drive test data packet and the version information of the algorithm under test, and sends the drive test data packet to the simulation engine module according to the version information;

[0040] The road test data packet is deployed in the actual vehicle based on the autonomous driving data algorithm. It mainly includes two parts: the problem packet generated when the actual vehicle encounters a problem and the conventional scenario packet collected in a specific scenario. The data in the data packet is subsequently replayed, and replayed through the relevant simulation engine algorithm, and input into the algorithm module under test to restore the problem, reproduce the repair, and perform conventional scenario testing on the algorithm. In the entire test process, the source of the target vehicle's motion trajectory is formed through the playback of the data packet, which is used for closed-loop testing of the algorithm. Playback is to replay the data of the recorded package. If the same algorithm as that on the actual vehicle is used, the scene is restored; if a new algorithm is used, it is verified whether the new algorithm optimizes the original scene effect. When the data packet is recorded, the version information in the data packet will be recorded, and then when the algorithm under test is tested, the message version information of the algorithm under test will also be input.

[0041] The intermediate interface layer model is mainly used for data interaction and realizes the compatibility scheme between different versions of data packets and different versions of algorithms. In the actual development process, the communication protocols required by the modules of the algorithm real vehicle at different stages are different. In order to make the collected data packets compatible with different versions of algorithms, the data packets are not directly played back to the simulation engine module, but need to pass through an intermediate interface layer. The input of the intermediate layer interface is the message version of the currently played back road test data packet and the current version of the algorithm information being tested. According to the two input version information, the data in the data packet is forwarded to the simulation engine module in real time, and the data of the simulation engine module is forwarded to the algorithm module being tested to realize data interaction; at the same time, the communication between the intermediate interface layer and the simulation engine model is fixed to reduce the excessive version iteration caused by incompatible message matching of the software. The intermediate interface layer sends the version information of the algorithm being tested to the simulation engine module because each module such as planning, control, and map communicates through msg messages (each module in the ROS system communicates through msg messages). During the iterative development of the algorithm, the messages between the modules will change. If there is no version information of the message in the recorded package, then it will not know how to match the message during simulation. However, considering that the simulation engine must be relatively stable and cannot be changed just because the message version has changed, an intermediate interface layer module has been developed. This intermediate interface layer module must be compatible with the historical version of the message in each data packet, and also perform data matching according to the message version in the algorithm under test.

[0042] Step 2: The simulation engine module plays back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputs the positioning and chassis data information to the algorithm module under test through the intermediate interface layer. The calculation function of the vehicle dynamics module is as follows:

[0043] in, is the vehicle coordinate, φ is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle;

[0044] The simulation engine module is mainly used for simulation data output after calculation. This module accepts the positioning, chassis and other data information calculated by the dynamic model, the datagram information forwarded by the intermediate interface layer, and after a series of logical processing, sends the data to the intermediate interface layer, and then to the subsequent algorithm module under test.

[0045] Since the calculation of the vehicle dynamics module is mainly based on a first-order inertia delay equation, vehicle kinematics equation, and collected real vehicle data, the vehicle dynamics model is calibrated. The autonomous driving control is generally divided into longitudinal control and lateral control. The vehicle can be based on its own coordinate system coordinates and geometric relationships. Establish the kinematic equation of a single vehicle with the center of gravity of the vehicle as the center: Under low-speed conditions, it is believed that the vehicle will not slide sideways, that is, there will be no drift, so β ​​= 0. After simplifying the kinematic equation, the kinematic equation under low speed and non-steering rear wheel conditions is obtained: Wherein, L is the wheelbase of the vehicle.

[0046] Step 3: The algorithm module under test performs problem restoration and reproduction, repair verification, and routine scenario testing on the algorithm under test based on the positioning and chassis data information calculated by the vehicle dynamics module, generates a new version of the algorithm under test, and feeds it back to the intermediate interface layer;

[0047] The dynamic model is mainly used to simulate the performance of the vehicle in the simulation. As the controlled object in the simulation, the accuracy of the dynamic model is related to the accuracy of the entire simulation. The dynamic model in this patent is implemented by combining the kinematic model and the real vehicle response parameters. The calibration target is to reproduce the real vehicle performance in the problem data packet, and the dynamic model parameters in the simulation are calibrated. Because, first of all, the problem must be restored using the same algorithm as the real vehicle to ensure that the system can restore the same problem as the real vehicle; then use the new version of the algorithm to test. If the actual performance of the test vehicle in this scenario is better than the original, then it is considered that the new algorithm is optimized relative to the original algorithm, that is, the original algorithm has been subjected to problem restoration reproduction, repair verification, and conventional scenario testing.

[0048] Step 4: The simulation engine module receives the new version of the algorithm under test through the intermediate interface layer, performs the simulation task, and calculates and outputs the simulation results.

[0049] Step 5: The container encapsulation module feeds back the simulation results to the simulation scheduling platform.

[0050] The simulation scheduling platform is connected to the container encapsulation module and is used to send simulation requirement parameters to the container encapsulation module according to the simulation task. The simulation scheduling platform mainly encapsulates the background algorithm and sends it to the container encapsulation module in the background server according to the corresponding simulation test to realize the parallel simulation of multiple tasks and the rapid test and verification of the algorithm. The simulation scheduling platform is mainly used to pass the parameters to the backend server, execute the corresponding simulation tasks, and obtain the corresponding simulation results. The simulation scheduling platform transmits the simulation task parameters to the background server in the form of a network protocol. The server receives the corresponding simulation parameters and executes the script according to the corresponding simulation parameters to start the image encapsulated in the server to generate a container. In the container, each module in the system is started, and the data in the simulation is recorded and transmitted back to the platform. The transmitted simulation parameters can easily realize multiple simulation tasks in parallel and improve the simulation efficiency.

[0051] The present invention provides an automatic driving simulation test system and method based on real vehicle data. The technical solution collects road test data packets, and forwards the data in the simulation data packets to the simulation engine module after data conversion at the intermediate interface layer. The simulation engine module forwards the positioning, chassis and other information to the algorithm under test after conversion calculation. After the algorithm under test receives the information, it updates the control instructions to the vehicle dynamics module. The vehicle dynamics module further updates the state information of the vehicle in the simulation through logical calculation, and realizes the closed-loop simulation of the entire algorithm under test. The technical solution can use this simulation system to repair the problem scenarios of the automatic driving system in the real vehicle scenario, and use conventional scenarios for test verification, accelerate the iteration of the automatic driving system, and save costs while meeting the test requirements.

[0052] Embodiment 3

[0053] The present invention provides an automatic driving simulation test system and method based on real vehicle data. Figure 1 As shown, the system includes an intermediate interface layer, a simulation engine module, a tested algorithm module and a vehicle dynamics module. Figure 2 As shown, the method steps are as follows.

[0054] Step 1, the intermediate interface layer obtains version information of the drive test data packet and the version information of the algorithm under test, and sends the drive test data packet to the simulation engine module according to the version information;

[0055] like Figure 4 As shown, the method steps of sending the drive test data packet to the simulation engine module according to the version information are as follows.

[0056] Step 11, the intermediate interface layer reads the version information from the drive test data packet and receives the corresponding data information in real time;

[0057] Step 12, converting the data information into a fixed format according to a preset simulation engine model interface protocol;

[0058] Step 13, sending the converted data information to the simulation engine module for communication.

[0059] The middle-layer model receives the corresponding data packet information in real time based on the version information read from the data packet, and then communicates with the simulation model layer according to the interface protocol specific to the simulation engine model, converts the data packet data into a fixed format and sends it to the simulation engine module, receives information of the simulation engine model specific protocol, and forwards the required information to the algorithm module under test according to the version of the algorithm under test.

[0060] Step 2, the simulation engine module plays back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputs the positioning and chassis data information to the algorithm module under test through the intermediate interface layer;

[0061] like Figure 5 The method of playing back the positioning and chassis data information calculated by the simulation engine module according to the simulation engine algorithm and the vehicle dynamics module includes the following steps.

[0062] Step 21, obtaining the relative information relationship between the main vehicle and the target vehicle in the question package and the position relationship of the main vehicle in the simulation environment;

[0063] Step 22, projecting the target vehicle to the positional relationship of the main vehicle coordinates in the simulation, thereby retaining the target vehicle motion information in the real vehicle data packet;

[0064] Step 23, simulating the motion trajectory of the target vehicle in the actual environment.

[0065] The simulation engine module is mainly used to process the data packet information. It mainly includes reprojecting the obstacle information. According to the relative information relationship between the main vehicle and the target vehicle in the problem packet and the position relationship of the main vehicle in the simulation environment, the target vehicle is projected to the position relationship under the main vehicle coordinates in the simulation, thereby retaining the target vehicle motion information in the real vehicle data packet and simulating the target vehicle motion trajectory in the actual environment.

[0066] Step 3: The algorithm module under test performs problem restoration and reproduction, repair verification, and routine scenario testing on the algorithm under test based on the positioning and chassis data information calculated by the vehicle dynamics module, generates a new version of the algorithm under test, and feeds it back to the intermediate interface layer;

[0067] like Figure 6 As shown, the method for the vehicle dynamics module to calculate the positioning and chassis data information includes the following steps.

[0068] Step 31, collecting real vehicle motion data information;

[0069] Step 32, combining the kinematic model and the dynamic model, and fitting them with the collected data;

[0070] Step 33, compare and calibrate the vehicle data in the simulation with the fitted data to minimize the difference between the two, so that the problem scenario can be reproduced when the data is played back.

[0071] The dynamics model is mainly used for dynamics simulation. Based on the collected real vehicle motion data information, the kinematic model and the dynamics model are combined and the data is fitted. The performance in the simulation is compared and calibrated with the actual vehicle performance to minimize the difference between the two and reproduce the problem scene during data playback. In addition, the map information in the simulation environment is mainly provided by the map library and the map engine.

[0072] Step 4: The simulation engine module receives the new version of the algorithm under test through the intermediate interface layer, performs the simulation task, and calculates and outputs the simulation results.

[0073] The present invention provides an automatic driving simulation test system and method based on real vehicle data. The technical solution collects road test data packets, and forwards the data in the simulation data packets to the simulation engine module after data conversion at the intermediate interface layer. The simulation engine module forwards the positioning, chassis and other information to the algorithm under test after conversion calculation. After the algorithm under test receives the information, it updates the control instructions to the vehicle dynamics module. The vehicle dynamics module further updates the state information of the vehicle in the simulation through logical calculation, and realizes the closed-loop simulation of the entire algorithm under test. The technical solution can use this simulation system to repair the problem scenarios of the automatic driving system in the real vehicle scenario, and use conventional scenarios for test verification, accelerate the iteration of the automatic driving system, and save costs while meeting the test requirements.

[0074] In summary, the embodiments of the present invention provide an autonomous driving simulation test system and method based on real vehicle data. In response to the simulation requirements in autonomous driving vehicles, based on road test data packets, the data packets are replayed and the replayed data are reprojected to realize closed-loop simulation of the tested algorithm in the autonomous driving algorithm. The road test data can be restored and reproduced, repaired and verified, and tested in conventional scenarios, thereby improving the simulation test efficiency and saving the simulation test cost.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An automatic driving simulation test system based on real vehicle data, characterized in that: The system comprises: The intermediate interface layer is used to receive the drive test data packet and the version information of the algorithm under test, send the drive test data packet to the simulation engine module according to the version information, receive the data output by the simulation engine module, and send it to the algorithm under test module; The simulation engine module is connected to the intermediate interface layer and is used to receive the road test data packet, replay the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and input it to the algorithm module under test through the intermediate interface layer. The calculation function of the vehicle dynamics module is as follows: in, are the vehicle coordinates, is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle; The algorithm module under test is connected to the middle interface layer and is used to restore and reproduce the problem, verify the repair, and conduct routine scenario testing on the algorithm based on the positioning and chassis data information calculated by the vehicle dynamics module, and form a new version of the algorithm, which is sent to the middle interface layer. The vehicle dynamics module is connected to the intermediate interface layer and the measured algorithm module respectively, and is used to calculate the positioning and chassis data information.

2. The automatic driving simulation test system based on real vehicle data according to claim 1, characterized in that: The system further comprises: The container encapsulation module is used to encapsulate the intermediate interface layer, the simulation engine module and the algorithm module under test, generate the corresponding container according to the simulation requirement parameters sent by the simulation scheduling platform, implement multiple data parallel playback simulations in the container, and feed back the simulation results to the simulation scheduling platform; The simulation scheduling platform is connected to the container encapsulation module and is used to send simulation requirement parameters to the container encapsulation module according to the simulation task.

3. The automatic driving simulation test system based on real vehicle data according to claim 1, characterized in that: The road test data packets include problem packets generated when a real vehicle encounters a problem, and scenario packets collected in specific scenarios.

4. A method using the automatic driving simulation test system based on real vehicle data according to claims 1-3, characterized in that: The method comprises: Step 1, the intermediate interface layer obtains version information of the drive test data packet and the version information of the algorithm under test, and sends the drive test data packet to the simulation engine module according to the version information; Step 2: The simulation engine module plays back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module, and inputs the positioning and chassis data information to the algorithm module under test through the intermediate interface layer. The calculation function of the vehicle dynamics module is as follows: in, are the vehicle coordinates, is the yaw angle, β is the sideslip angle of the center of mass, δ f is the front wheel turning angle, δ r is the rear wheel turning angle; Step 3: The algorithm module under test performs problem restoration and reproduction, repair verification, and routine scenario testing on the algorithm under test based on the positioning and chassis data information calculated by the vehicle dynamics module through the following functions, generates a new version of the algorithm under test, and feeds it back to the intermediate interface layer; Step 4: The simulation engine module receives the new version of the algorithm under test through the intermediate interface layer, performs the simulation task, and calculates and outputs the simulation results.

5. The automatic driving simulation test method based on real vehicle data according to claim 4 is characterized in that: The method further comprises: Step 0: The container encapsulation module encapsulates the intermediate interface layer, the simulation engine module, the vehicle dynamics module and the measured algorithm module. The simulation scheduling platform sends the simulation requirement parameters to the container encapsulation module according to the simulation task for subsequent playback simulation.

6. The automatic driving simulation test method based on real vehicle data according to claim 5, characterized in that: The method further comprises: Step 5: The container encapsulation module feeds back the simulation results to the simulation scheduling platform.

7. The automatic driving simulation test method based on real vehicle data according to claim 4, characterized in that: The method for sending the drive test data packet to the simulation engine module according to the version information includes: Step 11, the intermediate interface layer reads the version information from the drive test data packet and receives the corresponding data information in real time; Step 12, converting the data information into a fixed format according to a preset simulation engine model interface protocol; Step 13, sending the converted data information to the simulation engine module for communication.

8. The automatic driving simulation test method based on real vehicle data according to claim 4, characterized in that: The method for the simulation engine module to play back the positioning and chassis data information calculated by the simulation engine algorithm and the vehicle dynamics module includes: Step 21, obtaining the relative information relationship between the main vehicle and the target vehicle in the question package and the position relationship of the main vehicle in the simulation environment; Step 22, projecting the target vehicle to the positional relationship of the host vehicle coordinates in the simulation, thereby retaining the target vehicle motion information in the real vehicle data packet; Step 23, simulating the motion trajectory of the target vehicle in the actual environment.

9. The automatic driving simulation test method based on real vehicle data according to claim 4, characterized in that: The method for the vehicle dynamics module to calculate the positioning and chassis data information includes: Step 31, collecting real vehicle motion data information; Step 32, combining the kinematic model and the dynamic model, and fitting them with the collected data; Step 33, compare and calibrate the vehicle data in the simulation with the fitted data to minimize the difference between the two, so that the problem scenario can be reproduced when the data is played back.

Images