Low-latency transmission method and test system based on digital twin vehicle-in-the-loop testing

Abstract

The present invention relates to the field of vehicle-in-the-loop testing technology, and discloses a low-latency transmission method and test system based on digital twin vehicle-in-the-loop testing, including a real body and a twin; the real body includes a monitoring device, a data interaction device and a real test group; the monitoring device is used to collect test vehicle basic information and test vehicle operation information; the data interaction device is used to realize real-time information interaction between the real body and the twin; the real test group includes a real-scene test field, a HIL test terminal and a test vehicle; an ADAS domain controller is provided in the HIL test terminal; the twin is a simulation platform, and a virtual test field and a scene management module are provided in the simulation platform; a variety of test scenes are provided in the scene management module; and before the test starts, the ADAS domain controller is pre-stored with test scene data through a data interaction device, and the test scene data includes target object information. The present invention can achieve lower data transmission delay, higher test efficiency and test accuracy.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-in-the-loop testing technology, and in particular to a low-latency transmission method and testing system based on digital twin vehicle-in-the-loop testing. Background Art

[0002] As autonomous driving systems continue to gain popularity in the automotive world, there has been a lot of discussion about whether their functions are robust in the real world. The most important of these is the driving safety test of autonomous driving, which requires a lot of verification and confirmation.

[0003] In addition, the safety verification test of autonomous driving requires a long test process and a large test scene, and there is an obvious long-tail effect. Traditional testing methods are difficult to meet its testing requirements. At present, in order to meet the testing and verification needs of automotive intelligent connected driving systems and autonomous driving systems, a digital twin vehicle-in-the-loop test solution has been proposed, which is combined with virtual simulation for safety verification testing. Specifically, by building real-world and digital twin scenes, covering a wide range of scenes and high test authenticity, the training needs of artificial intelligence algorithms in intelligent driving scenarios are met, and the training efficiency and safety are greatly improved. The on-board equipment uses wireless communication to upload data to the server through the cloud network. The server creates a virtual world based on the received data, processes the data with a digital twin model, and then sends the processing results back to the connected vehicle, thereby realizing the vehicle's whole-vehicle-in-the-loop test, effectively solving the problems of difficult test scene coverage and high testing costs.

[0004] However, in the existing digital twin testing solutions, the interactive process involves extremely large amounts of data, the algorithm capacity is easily insufficient, and there is a high latency for complex testing scenarios, which greatly affects the real-time performance, efficiency and accuracy of the test. Summary of the invention

[0005] The present invention aims to provide a low-latency transmission method and test system based on digital twin whole-vehicle in-the-loop testing, which is used to solve the technical problems of high data transmission delay, poor test real-time performance, low test efficiency and low test accuracy in existing digital twin testing schemes.

[0006] To achieve the above object, the present invention provides the following basic scheme:

[0007] Solution 1

[0008] A low-latency test system based on digital twin vehicle-in-the-loop testing, comprising a real entity and a twin entity with established communication connection; the real entity comprises a monitoring device, a data interaction device and a real test group; the monitoring device is used to collect basic information of the test vehicle and the test vehicle operation information; the data interaction device is used to realize real-time information interaction between the real entity and the twin entity; the real test group comprises a real-scene test field, a HIL test terminal and a test vehicle; the HIL test terminal is provided with an ADAS domain controller;

[0009] The twin is a simulation platform, which is provided with a virtual test field and a scene management module; the virtual test field is a digital twin corresponding to the real test group; the scene management module is provided with a variety of test scenes, and the scene management module is used to construct and adjust the test scenes;

[0010] Before the test begins, the ADAS domain controller pre-stores test scenario data through a data interaction device, and the test scenario data includes target object information.

[0011] Solution 2

[0012] The low-latency transmission method based on digital twin vehicle-in-the-loop testing includes the following steps:

[0013] Step 1: Before the test begins, the simulation platform constructs a digital twin corresponding to the real test group; the digital twin includes a virtual vehicle and a virtual test scene; the real test group includes a real-scene test site, a HIL test terminal, and a test vehicle;

[0014] Step 2: pre-store test scenario data in the ADAS domain controller through a data interaction device, wherein the test scenario data includes target object information; the ADAS domain controller is set in the HIL test terminal and serves as the object under test;

[0015] Step 3: Start executing vehicle testing; during the test, the test vehicle executes corresponding command actions according to the control instructions issued by the real-time processor; the real-time processor is set in the simulation platform; the real-time processor is used to calculate the required parameters according to the expected movement to control the test vehicle to achieve the expected movement.

[0016] The working principle and advantages of the present invention are: before the test starts, by pre-storing the test scenario data in the ADAS domain controller of the test vehicle, the amount of data transmission during the test process is greatly reduced. It is no longer necessary to transmit the test scenario data online like the conventional test scheme. It is only necessary to transmit the dynamic response data of the test vehicle, the necessary scenario trigger signals and trajectory matching information in real time. The amount of data transmitted by these information is relatively small, which can reduce the data delay in the existing digital twin test from 15ms-20ms to within 8ms, which can greatly reduce the delay and help improve the test efficiency.

[0017] Moreover, compared with the conventional digital twin test scheme, the conventional scheme tends to be a complete real-time interactive scheme, that is, the test equipment is placed on the test vehicle, so that the test scene information and the dynamic response information of the test vehicle interact in real time. Although this test method is relatively convenient, the amount of real-time interactive data in this method is so large that there is a high delay and sufficient vehicle space is required to install and configure the test equipment. Under normal test conditions, the impact of this information transmission delay is relatively unobvious. For example, when the vehicle speed is low, the difference between the digital virtual scene and the real test scene caused by the cloud real-time transmission delay is relatively low, so that people often ignore the impact of data transmission delay. But in fact, the higher the speed of the autonomous driving vehicle, the greater the impact of the information transmission delay. For example, when the vehicle under test passes point A at a speed of 60km / h, the virtual test scene is triggered to send the action, the information transmission delay is 100ms, and the virtual test scene reaches the vehicle under test. The vehicle has exceeded point A by about 1.7m, which makes the virtual scene and the real scene have a large position difference at the same time. For the autonomous driving safety function test, the test error caused by this position difference is unacceptable.

[0018] This solution discovered this problem and pre-processed the ADAS domain controller of the test vehicle to reduce the amount of data transmitted in the test scene. By reducing the amount of real-time data transmission and using a smaller amount of wireless signal transmission, the delay was effectively reduced. The position difference between the virtual scene and the real scene at the same time was small, which helped to improve the test accuracy. In addition, the structural setting did not bring any additional structural setting burden and was easy to operate. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a system structure diagram of embodiment 1 of the low-latency transmission method and test system based on digital twin vehicle-in-the-loop testing of the present invention. DETAILED DESCRIPTION

[0020] The following is a further detailed description through specific implementation methods:

[0021] Embodiment 1

[0022] The embodiment is basically as shown in the attached Figure 1 Shown: A low-latency test system based on digital twin vehicle-in-the-loop testing, including a real entity and a twin entity that establish a communication connection.

[0023] The real body includes monitoring equipment, data interaction equipment and a real test group. The monitoring equipment is used to collect basic information of the test vehicle and test vehicle operation information. Specifically, the monitoring equipment includes dynamic measurement equipment and video monitoring equipment; wherein the dynamic measurement equipment is used to collect test vehicle operation information; the test vehicle operation information includes real vehicle state data and control data such as vehicle acceleration, braking and steering actions, as well as vehicle position information including posture data and GPS data. Moreover, in this embodiment, the basic information of the test vehicle is also collected before the test, and the basic information of the test vehicle includes vehicle model, basic body size, etc. The video monitoring equipment is arranged inside the vehicle and is used to monitor the actions of the test vehicle driver. The video monitoring equipment can effectively confirm whether the driver's operating behavior is safe, which helps to improve the safety of the overall test.

[0024] The data interaction device is used to realize real-time information interaction between the real body and the twin. Specifically, the data interaction device includes a vehicle-mounted industrial computer and an injection device. In this embodiment, the vehicle-mounted industrial computer is used to collect and record monitoring equipment information. The injection device adopts a conventional data injection system, which is used to accept the virtual simulation content generated by the twin, i.e., the simulation platform, and inject it into the vehicle-mounted industrial computer, as well as accept the information collected and recorded by the vehicle-mounted industrial computer and feed it back to the twin.

[0025] The reality test group includes a real-life test field, a HIL test terminal and a test vehicle; an ADAS domain controller is provided in the HIL test terminal. In this embodiment, the ADAS domain controller is used for driving assistance. HIL stands for Hardware In the Loop. The HIL test terminal refers to the hardware-in-the-loop test terminal, which is used for simulation testing of the controller. The test vehicle is equipped with a positioning system, which is composed of a GNSS module, an IMU module, etc. Before the test starts, the ADAS domain controller pre-stores test scenario data through a data interaction device. In this embodiment, the test scenario data specifically refers to the virtual test field data and virtual test scenario data pre-built in the simulation platform. The test scenario data includes target object information. In this embodiment, the target object information specifically refers to the point cloud, image and ADC perception data sent by the fusion processing data center, as well as the traffic target location and category information generated based on this.

[0026] The twin is a simulation platform. The simulation platform is provided with a virtual test field, a scenario management module, a map construction module, a site construction module, a vehicle construction module, a trajectory simulation module, a controller module and a real-time processor. During the test, the twin and the real body transmit data in real time; the transmitted real-time data includes the dynamic response data of the test vehicle, the scenario trigger signal and the trajectory benchmarking information, and does not include the test scenario data. Among them, the scenario trigger signal is used to trigger different test scenario data pre-stored in the ADAS domain controller to realize the corresponding call of the scenario. The controller module is used to generate corresponding control decisions according to the expected motion; and the controller module acts as the object under test.

[0027] The virtual test field is a digital twin corresponding to the real test group; in this embodiment, the virtual test field is a virtual test field with a virtual vehicle corresponding to the test vehicle. The scenario management module is provided with a variety of test scenarios, and the scenario management module is used to construct and adjust the test scenarios.

[0028] The map construction module is pre-stored with a pre-built map of the virtual test site, which contains information such as road type, road length, and road direction. The site construction module is used to build a virtual test site based on the pre-built map. The vehicle construction module is used to build a virtual model of the test vehicle, including a virtual vehicle model, a dynamic model, etc. Through the cooperation of the map construction module and the site construction module, it is possible to ensure the synchronization of the position of the twin body with the real body and the construction of a geographic fence. Specifically, for the position synchronization function, the vehicle position information, here refers to the vehicle coordinates (i.e., longitude, latitude and altitude) and the pre-built map can be matched through a map matching algorithm to update their current position in the twin body and achieve position synchronization. Specifically, the map matching algorithm used in this embodiment includes: by collecting basic information of the actual test site, obtaining data and making it into a high-precision map, and then simulating the traffic environment through the map construction module and the site construction module, one or more static elements in the driving road, driving signs, traffic lights, traffic signs, traffic isolation facilities and roadside buildings can be added to calibrate the mapping relationship between the real map and the virtual traffic environment, and the mapping relationship between the test vehicle and the simulated vehicle model.

[0029] For the geo-fencing function, geo-fencing is implemented by defining a flag to check the position and status of the test vehicle in each time step. Specifically, in this embodiment, the motion information of the simulated vehicle model at each time point will be used as the analysis basis for the performance test of the controller under test (corresponding to the controller of the test vehicle) so that it can be analyzed after the case is completed.

[0030] The trajectory simulation module is used to construct the motion trajectory of the test vehicle and generate the expected motion of the test vehicle; in this embodiment, the twin sends the acceleration, braking and steering instructions after the decision to the real vehicle actuator (i.e., the actuator of the test vehicle), and the test vehicle moves to the next position according to the instructions. The motion corresponding to this step is the expected motion, and the operating data generated by the test vehicle in this process is the dynamic response data. Specifically, the trajectory simulation module can display the real-time moving trajectory of the test vehicle on the pre-built map according to the position information measured by the GNSS module and the IMU module on the test vehicle, and the trajectory is the trajectory benchmarking information. The real-time processor is used to calculate the required parameters according to the expected motion to control the test vehicle to achieve the expected motion. In this embodiment, the real-time processor uses the HIL device real-time processing unit module. The scene management module pre-stores the traffic signal model and the sensor model; the traffic signal model and the sensor model are used to construct dynamic scene information. Specifically, the scene management module is provided with a scene library. When constructing the test scene, the scene management module extracts the scene reference elements from the scene library and virtually simulates and constructs other traffic participants such as road conditions, weather, trees, pedestrians and vehicles, and then renders a high-fidelity test scene. In addition, traffic signal models and sensors can be dynamically added to the test scenarios to build richer and more diverse test scenarios, which can effectively improve the richness of scene coverage of autonomous driving data.

[0031] This embodiment also provides a low-latency transmission method based on digital twin vehicle-in-the-loop testing, comprising the following steps:

[0032] The method operation in this embodiment can adopt a low-latency testing system based on digital twin vehicle in-loop testing as described above.

[0033] Step 1: Before the test begins, the simulation platform constructs a digital twin corresponding to the real test group; the digital twin includes a virtual vehicle and a virtual test scene. Specifically, the simulation platform here pre-constructs a virtual test scene according to the actual test requirements and forms a virtual test field. A virtual vehicle model corresponding to the test vehicle is set in the virtual test field. The real test group includes a real-scene test field, a HIL test terminal and a test vehicle.

[0034] Step 2: Pre-store test scenario data in the ADAS domain controller through the data interaction device, and the test scenario data includes target object information; the ADAS domain controller is set in the HIL test terminal and serves as the object under test. That is, the test scenario-related data pre-constructed in step 1 is pre-imported into the ADAS domain controller. After the test vehicle starts the test, the ADAS domain controller directly extracts the test scenario information for testing.

[0035] Step 3: Start the vehicle test; during the test, the test vehicle executes the corresponding command action according to the control command issued by the real-time processor; the real-time processor is set in the simulation platform; the real-time processor is used to calculate the required parameters according to the expected motion to control the test vehicle to achieve the expected motion. In addition, during the test, the twin body and the real body perform real-time data transmission; the transmitted real-time data includes the dynamic response data of the test vehicle, the scene trigger signal and the trajectory benchmarking information, but does not include the test scene data.

[0036] The step 3 also includes the following sub-steps:

[0037] Sub-step 1: The virtual vehicle in the digital twin operates in a virtual test scenario; the test scenario data is injected into the ADAS domain controller at the HIL test end, and the ADAS domain controller performs perception, recognition, and decision control.

[0038] Sub-step 2: The test vehicle operates in the real-scene test field; The test vehicle operates in the real-scene test field; The test vehicle receives the data information conveyed by the simulation platform and makes corresponding action responses.

[0039] Specifically, the data information conveyed by the simulation platform includes the data of each sensor and the scene trigger signal. The ADAS domain controller at the HIL test end in the real test group receives the data information. Based on the scene trigger signal, the ADAS domain controller extracts information from the pre-stored test scene information and perceives the test scene. Based on the data of each sensor, after calculation by the intelligent driving algorithm, it outputs it as the corresponding acceleration, braking, steering and other operating instructions, and transmits it to the on-board industrial computer of the test vehicle to control the test vehicle to make corresponding action responses.

[0040] Sub-step 3: The test vehicle feeds back real-time operating information to the simulation platform (i.e., the real-time operating information of the action response in sub-step 2). The virtual vehicle in the digital twin operates according to the real-time operating information fed back by the test vehicle, thereby completing the vehicle-in-the-loop test.

[0041] The present embodiment provides a low-latency transmission method and test system based on digital twin whole-vehicle in-the-loop testing. By pre-importing test scenario data into the ADAS domain controller of the test vehicle, the amount of data transmission during the test process is greatly reduced. It is no longer necessary to transmit the test scenario data online like conventional test schemes. It is only necessary to transmit the dynamic response data of the test vehicle, necessary scenario trigger signals and trajectory matching information in real time. The amount of data transmitted by these information is relatively small, and the data delay in the existing digital twin test can be reduced from 15ms-20ms to within 8ms, which can greatly reduce the delay and help improve the test efficiency.

[0042] Moreover, what is special is that this solution breaks through the conventional thinking that simulation data information needs to rely on vehicle-mounted test equipment for full real-time interaction, and finds that there is a data transmission delay problem in this conventional solution, especially in the scenario where the test vehicle is running at high speed. This delay problem, even a small delay, will cause a large test error. In response to this, this solution separates the test scene data through certain pre-operations and pre-stores it in the ADAS domain controller. By setting the scene trigger signal to correspond to the call scene, the previous large amount of online transmission is transferred to a very small amount of wireless signal transmission, while avoiding the problem of vehicle model simulation accuracy. In addition, the setting in the ADAS domain controller also avoids the addition of additional components, does not cause additional burden on the test space inside the vehicle, and can achieve better test results.

[0043] Embodiment 2

[0044] The low-latency testing system based on the digital twin vehicle-in-the-loop testing is based on the first embodiment and a scene switching module and a scene calling module are added.

[0045] The scene switching module is arranged in the simulation platform; the scene calling module is arranged in the ADAS domain controller. Specifically, in this embodiment, when constructing the test scene, the scene construction module pre-constructs several different test scenes and sets corresponding labels for different test scenes. The scene switching module is used to quickly control the simulation platform to switch the test scene according to the actual test requirements, and transmit the switched scene label information to the scene calling module.

[0046] In this embodiment, a plurality of test scenario data and corresponding label information are pre-stored in the ADAS domain controller. The scenario calling module is used to immediately call the ADAS domain controller to load the test scenario with the scenario label information when the scenario label information is received. Moreover, when there is overlapping information between the pre-switched test scenario and the test scenario before the switch, the scenario calling module also controls that when calling the pre-loaded test scenario data, only the part that does not overlap with the test scenario before the switch is called and updated.

[0047] This embodiment also provides a low-latency transmission method based on digital twin vehicle in-loop testing, which is the same as the method described in Embodiment 1, so it is not repeated here.

[0048] The present embodiment provides a low-latency transmission method and test system based on digital twin vehicle in-loop testing, which can achieve rapid switching between different test scenarios and does not repeatedly load overlapping test scenarios, thereby helping to further improve test efficiency.

[0049] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement the scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for the ordinary technicians in the relevant field to implement this application. It should be pointed out that for the technicians in this field, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A low-latency test system based on digital twin vehicle-in-the-loop testing, characterized by: It includes a real body and a twin body that establish a communication connection; the real body includes a monitoring device, a data interaction device and a real test group; the monitoring device is used to collect basic information of the test vehicle and the test vehicle operation information; the data interaction device is used to realize real-time information interaction between the real body and the twin body; the real test group includes a real-scene test field, a HIL test terminal and a test vehicle; the HIL test terminal is provided with an ADAS domain controller; The twin is a simulation platform, which is provided with a virtual test field and a scene management module; the virtual test field is a digital twin corresponding to the real test group; the scene management module is provided with a variety of test scenes, and the scene management module is used to construct and adjust the test scenes; Before the test begins, the ADAS domain controller pre-stores test scenario data through a data interaction device, wherein the test scenario data includes target object information; During the test, the twin body and the real body transmit data in real time; the transmitted real-time data includes the dynamic response data of the test vehicle, the scene trigger signal and the trajectory benchmarking information, but does not include the test scene data; It also includes a scene switching module and a scene calling module; The scene switching module is arranged in the simulation platform; the scene calling module is arranged in the ADAS domain controller; when constructing the test scene, the scene construction module pre-constructs several different test scenes and sets corresponding tags for different test scenes; The scene switching module is used to control the simulation platform to switch the test scene, and transmit the switched scene label information to the scene calling module; the ADAS domain controller correspondingly pre-stores a variety of test scene data and corresponding label information; The scenario calling module immediately calls the ADAS domain controller to load the test scenario with the scenario tag information upon receiving the scenario tag information.

2. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 1 is characterized in that: The monitoring equipment includes a dynamic measurement device and a video monitoring device; the dynamic measurement device is used to collect the test vehicle operation information; the video monitoring device is arranged inside the vehicle and is used to monitor the actions of the test vehicle driver.

3. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 1 is characterized in that: The data interaction device includes a vehicle-mounted industrial computer and an injection device.

4. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 1 is characterized in that: The simulation platform is also provided with a map construction module, a site construction module and a vehicle construction module; the map construction module stores a pre-built map of the virtual test site; the site construction module is used to build a virtual test site according to the pre-built map; and the vehicle construction module is used to build a virtual model of the test vehicle.

5. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 1 is characterized in that: The simulation platform also includes a trajectory simulation module and a controller module; the trajectory simulation module is used to construct a motion trajectory of the test vehicle and generate an expected motion of the test vehicle; the controller module is used to generate corresponding control decisions according to the expected motion; and the controller module serves as the object under test.

6. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 4 is characterized in that: The simulation platform also includes a real-time processor, which is used to calculate required parameters according to the expected motion to control the test vehicle to achieve the expected motion.

7. The low-latency test system based on digital twin vehicle-in-the-loop testing according to claim 1 is characterized in that: The scene management module pre-stores a traffic signal model and a sensor model; the traffic signal model and the sensor model are used to construct dynamic scene information.

8. A low-latency transmission method based on digital twin vehicle-in-the-loop testing, characterized in that: The low-latency test system based on digital twin vehicle-in-the-loop testing as described in any one of claims 1 to 7 is used for testing, comprising the following steps: Step 1: Before the test begins, the simulation platform constructs a digital twin corresponding to the real test group; the digital twin includes a virtual vehicle and a virtual test scene; the real test group includes a real-scene test site, a HIL test terminal, and a test vehicle; Step 2: pre-store test scenario data in the ADAS domain controller through a data interaction device, wherein the test scenario data includes target object information; the ADAS domain controller is set in the HIL test terminal and serves as the object under test; Step 3: Start executing vehicle testing; during the test, the test vehicle executes corresponding command actions according to the control instructions issued by the real-time processor; the real-time processor is set in the simulation platform; the real-time processor is used to calculate the required parameters according to the expected movement to control the test vehicle to achieve the expected movement.

9. The low-latency transmission method based on digital twin vehicle-in-the-loop testing according to claim 8 is characterized in that: The step 3 also includes the following sub-steps: Sub-step 1: The virtual vehicle in the digital twin operates in a virtual test scenario; the test scenario data is injected into the ADAS domain controller at the HIL test end, and the ADAS domain controller performs perception, recognition, and decision control; Sub-step 2: The test vehicle runs in the real-scene test field; the test vehicle receives data information transmitted by the simulation platform and responds with corresponding actions; Sub-step 3: The test vehicle feeds back real-time operating information to the simulation platform, and the virtual vehicle in the digital twin operates according to the real-time operating information fed back by the test vehicle.

Images