Automatic driving test system and method based on digital twinning

Abstract

The invention relates to the technical field of automatic driving, and discloses an automatic driving test system and method based on digital twinning, and the automatic driving test system comprises an experimental test layer driving test subsystem and an on-site test layer driving test subsystem. Data interaction is carried out between the experimental test layer driving test subsystem and the field test layer driving test subsystem through a 4G / 5G network communication module; according to the system, a vehicle dynamics model constructed by utilizing a Lagrange equation dynamics equation has the advantages of flexible parameter configuration, capability of realizing information interaction between a physical space and a virtual space, high response timeliness, capability of accurately reflecting dynamic performance characteristics of a real vehicle in real time and the like; the problems that in the prior art, open road testing is high in cost and low in safety, and real-time data collection and processing cannot be achieved are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving, and specifically relates to an autonomous driving test system and method based on digital twins. Background Art

[0002] The Motor Vehicle Auto Driving System, also known as autonomous vehicles; Self-piloting automobile, is also called driverless cars, computer-driven cars, or wheeled mobile robots. It is an intelligent car system that achieves unmanned driving through an on-board computer system. With the development of autonomous driving technology, autonomous vehicles need to accurately and in real time perceive and understand the environment in order to make efficient driving decisions. In addition, autonomous vehicles must be able to respond to various emergencies at any time, including man-made or natural obstacles, which requires autonomous vehicles to have a complete safety design. Therefore, the accuracy and safety of the autonomous driving system are the two major challenges facing autonomous driving technology. With the continuous development of autonomous driving, it is becoming increasingly important to conduct effective testing of autonomous vehicles.

[0003] During the safety test of the autonomous driving system, open road testing has problems such as high cost, low safety, and inability to achieve real-time data collection and processing. As a simulation system that can achieve real-time data collection and transmission, analog digital twin technology is widely used in various simulation test systems, such as complex driving scene simulation, traffic scene model construction, etc., which effectively ensures the real-time nature of data collection and processing while reducing costs and ensuring test safety. Therefore, applying digital twin technology to autonomous driving test systems, providing real-time feedback of virtual simulation decisions to vehicles, and providing optimization strategies for autonomous driving technology has become an urgent problem to be solved in this field.

[0004] Therefore, the present invention proposes an autonomous driving test system and method based on digital twins to solve the problems existing in the above-mentioned prior art. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the shortcomings of the prior art, the present invention provides an autonomous driving test system and method based on digital twins, which have the advantages of flexible configuration of various parameters, ability to realize information interaction between physical space and virtual space, strong response timeliness, and ability to accurately reflect the dynamic performance characteristics of real vehicles in real time. It solves the problems of high cost, low safety, and inability to realize real-time data collection and processing in open road testing in the prior art.

[0007] (II) Technical solution

[0008] In order to achieve the above-mentioned side view purpose of being able to realize information interaction between physical space and virtual space, having strong response timeliness and being able to accurately reflect the dynamic performance characteristics of real vehicles in real time, the present invention provides the following technical solutions:

[0009] An autonomous driving test system based on digital twins, comprising an experimental test layer driving test subsystem and a field test layer driving test subsystem, wherein the experimental test layer driving test subsystem and the field test layer driving test subsystem exchange data via a 4G / 5G network communication module; the experimental test layer driving test subsystem comprises:

[0010] A high-fidelity environment model is built based on test requirements and combined with real data to simulate the external environment during autonomous driving, and the simulated environment is imported into the Carla processing module;

[0011] The high-precision sensor model is a virtual vehicle-mounted sensor model configured in the Carla processing module, which is used to perceive the autonomous driving situation and surrounding environment of the simulated autonomous driving vehicle in real time;

[0012] The Carla processing module is used to deploy high-fidelity environmental models and high-precision sensor models and connect to the MATLAB SIMULINK processing module;

[0013] MATLAB SIMULINK processing module, used to build a vehicle model based on a high-confidence vehicle dynamics model, and combine the vehicle model with the autonomous driving system to perform virtual autonomous driving system testing and real-time vehicle dynamic response.

[0014] Preferably, the MATLAB SIMULINK processing module is connected to the Carla processing module via a Python API interface to retrieve data; the Carla processing module can perform data interaction and memory sharing with the MATLAB SIMULINK processing module.

[0015] Preferably, the field test layer driving test subsystem includes

[0016] An environmental system, including human-vehicle-road-environment conditions during autonomous vehicle testing in a real environment, used to provide environmental disturbance inputs during autonomous vehicle testing to the sensor system and vehicle multi-body system when used in field testing;

[0017] The sensing system is used to perceive the position and driving conditions of the autonomous vehicle in real-world environments during testing, and to receive the perception planning and control inputs of the vehicle's spatially coupled dynamic response system;

[0018] The chassis execution system is used to receive information from the sensor system during testing and control the steering, driving and braking of the vehicle;

[0019] Vehicle multi-body system, used to detect vehicle data during the test in real time;

[0020] The vehicle spatial coupling dynamic response system is used to determine the actual motion status of the vehicle based on the detection data of the vehicle multi-body system.

[0021] Preferably, the chassis execution system includes a steering system, a drive system and a braking system; the vehicle data includes vehicle sprung mass, suspension system, unsprung mass and tire system data;

[0022] Preferably, the actual motion condition of the vehicle includes longitudinal motion, lateral motion, vertical motion, pitch motion, roll motion and yaw motion of the vehicle.

[0023] The present invention provides a second solution: an autonomous driving test method based on digital twins, wherein the test method is implemented based on an autonomous driving test system based on digital twins, and comprises:

[0024] Step 1: Construct the experimental test layer driving test subsystem;

[0025] Step 1.1: Establish the longitudinal-lateral-vertical spatial coupled dynamic equations of the vehicle;

[0026] Step 1.2: Establish the vehicle model based on Matlab / Simulink processing module;

[0027] Step 2: Establish the interaction method between the experimental test layer driving test subsystem and the field test layer driving test subsystem;

[0028] Step 3: Load traffic flow and autonomous driving algorithm into the virtual space formed by the experimental test layer driving test subsystem and the field test layer driving test subsystem to simulate the real scene, and use the interaction method established in step 2 to interact between the experimental test layer driving test subsystem and the field test layer driving test subsystem to complete the test of the autonomous driving algorithm.

[0029] Preferably, the process of establishing the longitudinal-lateral-vertical space coupling dynamic equation of the whole vehicle in step 1.1 includes:

[0030] Step 1.1.1: In the vehicle longitudinal-lateral vehicle dynamics analysis model, the Lagrange equation dynamics equation is used to obtain:

[0031]

[0032]

[0033] Where m is the vehicle mass, x is the vehicle longitudinal displacement, and y is the vehicle lateral displacement. is the vehicle yaw angle, θ StF and θ StR are the front and rear axle steering angles, F wx 、F wy and M w are the tire longitudinal force, lateral force and self-aligning moment respectively, and P is the contact point between the tire and the ground;

[0034] Step 1.1.2: In the nonlinear vehicle vertical dynamics analysis model, simplify the suspension mechanism to equivalent suspension stiffness and damping characteristics, connecting the body and wheels; use the Lagrange equation vehicle dynamics equation to obtain:

[0035] z FL =z s +l1sinψ-l3sinθ

[0036] z FR =z s -l2sinψ-l3sinθ

[0037] z RL =z s +l1sinψ+l4sinθ

[0038] z RR =z s -l2sinψ+l4sinθ

[0039] z ML =z s +l1sinψ+l5sinθ

[0040] z MR =z s -l2sinψ+l5sinθ

[0041]

[0042]

[0043] In the formula, ψ, θ, z s are the body roll angle, pitch angle and vertical displacement, z FL , …, z MR is the vertical displacement of the connection point between the equivalent spring damper and the vehicle body, Z oi and z wi are road disturbance and wheel vertical displacement, respectively;

[0044] Combining the above vehicle dynamics equations, the longitudinal-lateral-vertical space coupling dynamics equation of the whole vehicle is obtained as follows:

[0045]

[0046] In the formula, i=1,…,n.

[0047] Preferably, the interaction method process between the experimental test layer driving test subsystem and the field test layer driving test subsystem in step 2 includes:

[0048] First, the real sensor data in the sensor system of the driving test subsystem of the field test layer is collected and sent to the cloud server through the industrial computer;

[0049] The data is then sent to the simulation server of the driving test subsystem of the experimental test layer through the simulation server. The simulation server sends the data to the virtual space, makes decisions based on the sensor data, and feeds back the decision information to the real vehicle of the driving test subsystem of the field test layer.

[0050] Preferably, the process of performing the interaction between the experimental test layer driving test subsystem and the field test layer driving test subsystem by using the interaction method established in step 2 in step 3 includes:

[0051] First, the factors affecting the road environment, traffic conditions, traffic participants, and environmental sensors are considered, and complex road scenes are modeled to perform 1:1 digital twins and generate test scenarios for the driving test subsystem of the experimental test layer.

[0052] Then, the corresponding scene information is sent in real time to the real vehicle of the driving test subsystem of the experimental test layer through the 4G / 5G network communication module;

[0053] The simulation platform analyzes the dynamic characteristics of the vehicle in the virtual space and generates a result report.

[0054] (III) Beneficial effects

[0055] Compared with the prior art, the present invention provides an autonomous driving test system and method based on digital twins, which has the following beneficial effects:

[0056] 1. The vehicle dynamics model constructed by the present invention using the Lagrange equation dynamics equation can flexibly configure various parameters and add a mixed traffic flow model for autonomous driving simulation. The dynamic analysis results are almost the same as those of CarSim, effectively ensuring the accuracy of autonomous driving software testing;

[0057] 2. When the driving test subsystem of the experimental test layer interacts with the driving test subsystem of the field test layer, this system uses the vehicle communication module and the cellular communication link to realize the information interaction between the physical space and the virtual space. By considering the response delay of the vehicle controller, it effectively makes up for the lack of timeliness of the network transmission path;

[0058] 3. The present invention constructs an integrated virtual-real simulation system for autonomous driving, which can accurately reflect the dynamic performance characteristics of a real vehicle in real time, solving the problems of high cost, low safety, and inability to realize real-time data collection and processing in open road testing in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a system structure diagram of the digital twin-based autonomous driving test system of the present invention;

[0060] Figure 2 This is a model diagram of the longitudinal-lateral dynamics analysis of the whole vehicle of the present invention;

[0061] Figure 3 This is the vertical dynamics analysis model diagram of the whole vehicle of this launch;

[0062] Figure 4 The vehicle dynamics model architecture diagram of the present invention;

[0063] Figure 5 is a steering angle input map of the present invention;

[0064] Figure 6 This is a Carla-Matlab / Simulink joint simulation result diagram of the present invention;

[0065] Figure 7 It is the outline diagram of the speed bump of the present invention;

[0066] Figure 8 It is a road surface interference situation diagram of the present invention;

[0067] Fig. 9 It is the vertical displacement diagram of the vehicle body of the present invention;

[0068] Fig.10 This is a diagram showing the vertical speed variation of the vehicle body of the present invention;

[0069] Fig.11 This is a diagram showing the change in vertical acceleration of the vehicle body of the present invention;

[0070] Fig.12 It is the real-time dynamic response diagram of the vehicle of the present invention;

[0071] Fig.13 This is a data transmission architecture diagram of the experimental test layer driving test subsystem and the field test layer driving test subsystem of the present invention;

[0072] Fig.14 This is a diagram of the virtual-real integrated simulation system of the autonomous driving of the present invention;

[0073] Fig.15 This is a normal driving simulation diagram of the automatic driving virtual-real integrated simulation system of the present invention;

[0074] Fig.16 This is an emergency avoidance simulation diagram of the automatic driving virtual-real integrated simulation system of the present invention;

[0075] Fig.17 This is a simulation diagram of emergency braking of the virtual-real integrated simulation system for autonomous driving of the present invention.

[0076] Among them, Fig.12 In the figure, Figure (a) is the real-time dynamic response diagram of the vehicle's yaw angular velocity, Figure (b) is the real-time dynamic response diagram of the vehicle's lateral velocity, Figure (c) is the real-time dynamic response diagram of the vehicle's roll angle, Figure (d) is the real-time dynamic response diagram of the vehicle's roll angular velocity, Figure (e) is the real-time dynamic response diagram of the vehicle's longitudinal velocity, and Figure (f) is the real-time dynamic response diagram of the vehicle's lateral acceleration. DETAILED DESCRIPTION

[0077] 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.

[0078] Example 1: Please refer to Figure 1 , this embodiment provides a technical solution:

[0079] An autonomous driving test system based on digital twins, which is implemented based on digital twin technology and is used to test autonomous driving software, including an experimental test layer driving test subsystem and a field test layer driving test subsystem. The experimental test layer driving test subsystem and the field test layer driving test subsystem are connected via a 4G / 5G network communication module to perform data exchange.

[0080] The experimental test layer driving test subsystem includes a high-fidelity environment model, a high-precision sensor model, a Carla processing module, a MATLAB SIMULINK processing module (its structure is as follows Figure 1 shown); where

[0081] The high-fidelity environment model is constructed by UE4 (Unreal Engine 4) based on test requirements and combined with real data to simulate the external environment during autonomous driving, and the simulated environment is imported into the Carla processing module;

[0082] The high-precision sensor model is a virtual vehicle-mounted sensor model, and the high-precision sensor model is configured in the Carla processing module, and is used to perceive the autonomous driving situation and surrounding environment of the simulated autonomous driving vehicle in real time;

[0083] The Carla processing module is implemented based on the Carla software, and is used to deploy a high-fidelity environment model and a high-precision sensor model, and is connected to the MATLAB SIMULINK processing module through the Python API interface to retrieve data. The Carla processing module also sends the sensor data detected by the high-precision sensor model to the autonomous driving system, and uses the autonomous driving system to make decisions and plan paths, and control vehicle actions;

[0084] The MATLAB SIMULINK processing module is used to build a vehicle model based on a high-confidence vehicle dynamics model, and combine the vehicle model with the autonomous driving system to perform virtual autonomous driving system testing and real-time dynamic response of the vehicle, and can perform data interaction and memory sharing with the Carla processing module during the test;

[0085] The field test layer driving test subsystem includes an environmental system, a sensor system, a chassis execution system, a vehicle multi-body system and a vehicle space coupling dynamic response system;

[0086] The environmental system includes the human-vehicle-road-environment conditions during the autonomous driving vehicle test in the real environment, and is used to provide the sensor system and the vehicle multi-body system with environmental interference input during the autonomous driving vehicle test when used in the field test;

[0087] The sensing system is used to perceive the position and driving conditions of the autonomous driving vehicle in real environment during testing, and receive the perception planning control input of the vehicle space coupling dynamic response system; wherein the sensing system is obtained by combining millimeter radar, millimeter wave radar and camera to form a combined navigation system during the actual vehicle testing process;

[0088] The chassis execution system is used to receive information from the sensor system during testing and perform steering, driving and braking control on the vehicle, including a steering system, a driving system and a braking system;

[0089] The vehicle multi-body system is used to detect vehicle data in real time during the test process, wherein the vehicle data includes vehicle sprung mass, suspension system, unsprung mass and tire system data;

[0090] The vehicle spatial coupling dynamic response system is used to determine the actual motion state of the vehicle according to the detection data of the vehicle multi-body system, and the actual motion state of the vehicle includes the longitudinal motion, lateral motion, vertical motion, pitch motion, roll motion and yaw motion of the vehicle;

[0091] Example 2: Please refer to Figure 2-Figure 17 , this embodiment provides a technical solution:

[0092] A digital twin-based autonomous driving test method, which is implemented based on the digital twin-based autonomous driving test system described in Example 1, comprising:

[0093] Step 1: Build the experimental test layer driving test subsystem

[0094] This includes using Carla software to build road traffic scenarios based on test requirements and combined with real data; using Matlab / Simulink to build vehicle models and flexibly configure driving parameters and sensor parameters; building mixed traffic flow models; configuring autonomous driving perception, decision-making planning, and control algorithms; and conducting experimental simulations to complete virtual testing;

[0095] When constructing the vehicle model, the vehicle body and tires are assumed to be rigid bodies, and the vehicle dynamics equations are derived using Lagrangian analytical mechanics. Generalized coordinates are used, the derivation process is standardized, and internal forces are automatically eliminated, so that a vehicle model with higher accuracy can be obtained. The specific process includes:

[0096] Step 1.1: Establish the longitudinal-lateral-vertical space coupling dynamic equations of the whole vehicle

[0097] Step 1.1.1: Figure 2 As shown, in the vehicle longitudinal-lateral vehicle dynamics analysis model, the Lagrange equation dynamics equation can be obtained:

[0098]

[0099] Where m is the vehicle mass, x is the vehicle longitudinal displacement, and y is the vehicle lateral displacement. is the vehicle yaw angle, θ StF and θ StR are the front and rear axle steering angles, F wx 、F wy and M w are the tire longitudinal force, lateral force and self-aligning moment respectively, and P is the contact point between the tire and the ground;

[0100] Step 1.1.2: In the vehicle system, the vehicle model has highly nonlinear characteristics. In order to reduce the complexity of vehicle dynamics analysis, such as Figure 3 As shown in the figure, in the nonlinear vehicle vertical dynamics analysis model, the simplified suspension mechanism is equivalent to the suspension stiffness and damping characteristics, connecting the body and the wheels; the Lagrange equation vehicle dynamics equation can be obtained:

[0101] Z FL =z s +l1sinψ-l3sinθ

[0102] z FR =z s -l2sinψ-l3sinθ

[0103] z RL =z s +l1sinψ+l4sinθ

[0104] z RR =z s -l2sinψ+l4sinθ

[0105] z ML =z s +l1sinψ+l5sinθ

[0106] z MR =z s -l2sinψ+l5sinθ

[0107]

[0108] In the formula, ψ, θ, z s are the body roll angle, pitch angle and vertical displacement, z FL , …, z MR is the vertical displacement of the connection point between the equivalent spring damper and the vehicle body, z oi and z wi are road disturbance and wheel vertical displacement, respectively;

[0109] Combining the above vehicle dynamics equations, the longitudinal-lateral-vertical space coupling dynamics equation of the whole vehicle is obtained as follows:

[0110]

[0111] Where, i = 1,…,n;

[0112] Step 1.2: Build a vehicle model based on Matlab / Simulink processing module

[0113] Step 1.2.1: Build a high-confidence vehicle dynamics model using Matlab / Simulink processing modules

[0114] In the vehicle system, the steering angle, throttle opening, DNPR gear position, and brake pedal travel are open, and the steering system, drive system, and brake system of the vehicle can be controlled; based on the tire magic formula, the tire system generates tire longitudinal force, lateral force, and return torque to act on the vehicle dynamics model; at the same time, considering the environmental impact of the vehicle during driving, environmental interference is also input into the vehicle dynamics model; finally, through vehicle dynamics analysis, a vehicle dynamics model that can accurately realize the real-time dynamic response of the vehicle is obtained, such as Figure 4 As shown;

[0115] Step 1.2.2: When the vehicle is driving in the virtual space, the environmental information received by the on-board sensor (high-precision sensor model) is sent to the autonomous driving perception module of the autonomous driving system through the Carla processing module to make decisions and plan paths, and then the autonomous driving vehicle is controlled by the control module of the autonomous driving system; the vehicle dynamics model receives the control command, uses Matlab / Simulink to perform vehicle dynamics analysis, and sends the real-time dynamic response of the vehicle to the shared memory; the data in the shared memory is retrieved through Python SPI to drive the target vehicle to run in the Carla software of the Carla processing module;

[0116] (1) Assume that the autonomous vehicle is traveling at a speed of 0.5 m / s and the steering angle is a sinusoidal function. (like Figure 5 ), input to a high-confidence vehicle dynamics model; e.g. Figure 6 , and obtain the vehicle virtual simulation results; the vehicle model parameters of the commercial vehicle dynamics software CarSim are all identified and calibrated through the actual data collected from the experimental vehicle; therefore, in order to verify the accuracy of the vehicle model, the vehicle model is changed to a four-wheeled sedan and compared with the dynamic analysis results of the vehicle dynamics analysis software CarSim;

[0117] (2) Figure 7 , assuming that the vehicle passes a speed bump, the road disturbance is input to the vehicle tire (such as Figure 8 ), causing the vehicle to move in the vertical direction; comparing the vehicle dynamics model in this embodiment with the commercial vehicle dynamics software CarSim, Fig. 9 , Fig.10 , Fig.11 are the vertical displacement, vertical velocity and vertical acceleration of the vehicle body respectively; according to the comparison results, it can be obtained that the vehicle dynamics model used in this embodiment has only a slight difference compared with CarSim;

[0118] (3) When the vehicle is turning, such as Fig.12 , which respectively displays the six state variables of the vehicle: yaw rate, lateral velocity, roll angle, roll rate, longitudinal velocity, and lateral acceleration; the results show that the vehicle dynamics model used in this embodiment is very consistent with the CarSim vehicle model. Therefore, in this embodiment, the experimental test layer driving test subsystem constructed can accurately reflect the dynamic performance characteristics of the real vehicle in real time; Fig.12 It can be seen that the virtual-real integrated simulation system for autonomous driving constructed by the present invention can accurately reflect the dynamic performance characteristics of a real vehicle in real time.

[0119] Step 2: Establish the interaction method between the experimental test layer driving test subsystem and the field test layer driving test subsystem:

[0120] Then, the real sensor data in the sensor system of the driving test subsystem of the field test layer is collected and sent to the cloud server through the industrial computer (the industrial computer is used for data collection, and the transmission and sending of data rely on 4G / 5G), and then sent to the simulation server of the driving test subsystem of the experimental test layer through the simulation server. The simulation server sends the data to the virtual space, makes decisions based on the sensor data, and feeds back the decision information to the real vehicle of the driving test subsystem of the field test layer;

[0121] The data transmission architecture that can realize data transmission between the experimental test layer driving test subsystem and the field test layer driving test subsystem is as follows: Fig.13 As shown in the figure, the on-board sensors collect real information about the vehicle, and the surrounding vehicles and roadside equipment such as radars and cameras collect information about the vehicle's driving environment. LTE-V2X is used to collect traffic environment information, such as roads, surrounding vehicles, pedestrians, etc.; and then the connection between the physical space and the virtual space is established through the 4G / 5G network communication module for information exchange;

[0122] Fig.13 It specifically describes the information interaction between physical space and virtual space. Fig. 9 , Fig.10 , Fig.11 All of them are comparison result diagrams of Example 1 of the present invention and the algorithm proposed in this embodiment and the carsim vehicle model, which are basically the same as the simulation results, and even have faster response times for some states, which indirectly confirms the real-time performance of the system. It shows that when the driving test subsystem of the experimental test layer interacts with the driving test subsystem of the field test layer, the system uses the vehicle communication module and the cellular communication link to realize the information interaction between the physical space and the virtual space, and effectively makes up for the lack of timeliness of the network transmission path by considering the response delay of the vehicle controller.

[0123] Based on the above steps 1 and 2, when in use, the test layer driving test subsystem described in step 1 is introduced into the field test layer driving test subsystem to construct a high-precision full-factor virtual space; in the field test layer driving test subsystem, vehicle driving information and real driving environment information are collected by using vehicle-mounted sensors such as integrated navigation and laser radar, and sensor data collection is realized through an industrial computer, and then Socket programming is used to establish communication with the simulation server through the network transmission layer, and the test layer driving test subsystem is connected with the field test layer driving test subsystem to realize the real-time dynamic response of the real vehicle in the virtual space;

[0124] Step 3: Use a real vehicle to send vehicle information to the virtual space through the 4G / 5G network communication module; at the same time, load traffic flow and autonomous driving algorithms in the virtual space to simulate the real scene and send them back to the real autonomous driving vehicle in a timely manner; the decision results of the controller will be sent back to the control simulator to test the autonomous driving algorithm, and finally realize the virtual-real integrated simulation of autonomous driving based on digital twins; the specific process includes

[0125] First, the influencing factors such as road environment (lanes, lane lines, road surface, weather and lighting, scene elements), traffic conditions (traffic flow, pedestrian congestion, adaptive cruise control), traffic participants (vehicles, pedestrians, obstacles) and environmental sensors (radar, camera, global positioning system / map, wireless communication) are considered. At the same time, complex road scenes are modeled and 1:1 digital twins are performed to generate test scenarios for the driving test subsystem of the experimental test layer.

[0126] Then, the corresponding scene information is sent in real time to the real vehicle of the driving test subsystem of the experimental test layer through the 4G / 5G network communication module;

[0127] The simulation platform analyzes the dynamic characteristics of the vehicle in the virtual space and generates a result report, enabling the autonomous driving test system to effectively achieve real-time optimization of the autonomous driving logic strategy.

[0128] In addition, in this embodiment, three driving models are shown (such as Fig.15 , Fig.16 and Fig.17 ): Normal driving, emergency avoidance and emergency braking situations.

[0129] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0130] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An autonomous driving test system based on digital twins, characterized by: It includes an experimental test layer driving test subsystem and a field test layer driving test subsystem, and the experimental test layer driving test subsystem and the field test layer driving test subsystem exchange data through a 4G / 5G network communication module; The experimental test layer driving test subsystem includes: A high-fidelity environment model is built based on test requirements and combined with real data to simulate the external environment during autonomous driving, and the simulated environment is imported into the Carla processing module; The high-precision sensor model is a virtual vehicle-mounted sensor model configured in the Carla processing module, which is used to perceive the autonomous driving situation and surrounding environment of the simulated autonomous driving vehicle in real time; Carla processing module, which is used to deploy high-fidelity environment models and high-precision sensor models and connect to MATLAB SIMULINK processing modules; MATLAB SIMULINK processing module, which is used to build a vehicle model based on a high-confidence vehicle dynamics model and combine the vehicle model with the autonomous driving system to perform virtual autonomous driving system testing and real-time vehicle dynamic response.

2. The digital twin-based autonomous driving test system according to claim 1, characterized in that: The MATLAB SIMULINK processing module is connected to the Carla processing module through the Python API interface to retrieve data; the Carla processing module can perform data interaction and memory sharing with the MATLAB SIMULINK processing module.

3. The digital twin-based autonomous driving test system according to claim 1, characterized in that: The field test layer driving test subsystem includes: An environmental system, including human-vehicle-road-environment conditions during autonomous vehicle testing in a real environment, used to provide environmental disturbance inputs during autonomous vehicle testing to the sensor system and vehicle multi-body system when used in field testing; The sensing system is used to perceive the position and driving conditions of the autonomous vehicle in real-world environments during testing, and to receive the perception planning and control inputs of the vehicle's spatially coupled dynamic response system; The chassis execution system is used to receive information from the sensor system during testing and control the steering, driving and braking of the vehicle; Vehicle multi-body system, used to detect vehicle data during the test in real time; The vehicle spatial coupling dynamic response system is used to determine the actual motion status of the vehicle based on the detection data of the vehicle multi-body system.

4. The digital twin-based autonomous driving test system according to claim 3, characterized in that: The chassis execution system includes a steering system, a driving system and a braking system; the vehicle data includes vehicle sprung mass, suspension system, unsprung mass and tire system data.

5. The digital twin-based autonomous driving test system according to claim 3, characterized in that: The actual motion condition of the vehicle includes the longitudinal motion, lateral motion, vertical motion, pitch motion, roll motion and yaw motion of the vehicle.

6. An autonomous driving testing method based on digital twins, characterized in that: The test method is implemented based on the digital twin-based autonomous driving test system according to claim 1, and includes: Step 1: Construct the experimental test layer driving test subsystem; Step 1.1: Establish the longitudinal-lateral-vertical spatial coupled dynamic equations of the vehicle; Step 1.2: Establish the vehicle model based on Matlab / Simulink processing module; Step 2: Establish the interaction method between the experimental test layer driving test subsystem and the field test layer driving test subsystem; Step 3: Load traffic flow and autonomous driving algorithm into the virtual space formed by the experimental test layer driving test subsystem and the field test layer driving test subsystem to simulate the real scene, and use the interaction method established in step 2 to interact between the experimental test layer driving test subsystem and the field test layer driving test subsystem to complete the test of the autonomous driving algorithm.

7. The digital twin-based autonomous driving testing method according to claim 6, characterized in that: The process of establishing the longitudinal-lateral-vertical space coupling dynamic equation of the whole vehicle described in step 1.1 includes: Step 1.1.1: In the vehicle longitudinal-lateral vehicle dynamics analysis model, the Lagrange equation dynamics equation is used to obtain: Where m is the vehicle mass, x is the vehicle longitudinal displacement, and y is the vehicle lateral displacement. is the vehicle yaw angle, θ StF and θ StR are the front and rear axle steering angles, F wx 、F wy and M w are the tire longitudinal force, lateral force and self-aligning moment respectively, and P is the contact point between the tire and the ground; Step 1.1.2: In the nonlinear vehicle vertical dynamics analysis model, simplify the suspension mechanism to equivalent suspension stiffness and damping characteristics, connect the body and wheels; use the Lagrange equation vehicle dynamics equation to obtain: With FL =from s +l1sinψ-l3sinθ With FR =from s -l2sinψ-l3sinθ With RL =from s +l1sinψ+l4sinθ With RR =from s -l2sinψ+l4sinθ With ML =from s +l1sinψ+l5sinθ With MR =from s -l2sinψ+l5sinθ In the formula, ψ, θ, z s are the body roll angle, pitch angle and vertical displacement, z FL ,…,z MR is the vertical displacement of the connection point between the equivalent spring damper and the vehicle body, z oi and z wi are road disturbance and wheel vertical displacement, respectively; Combining the above vehicle dynamics equations, the longitudinal-lateral-vertical space coupling dynamics equation of the whole vehicle is obtained as follows: In the formula, i=1,…,n.

8. The digital twin-based autonomous driving testing method according to claim 6, characterized in that: The interaction method process of the experimental test layer driving test subsystem and the field test layer driving test subsystem in step 2 includes: First, the real sensor data in the sensor system of the driving test subsystem of the field test layer is collected and sent to the cloud server through the industrial computer; The data is then sent to the simulation server of the driving test subsystem of the experimental test layer through the simulation server. The simulation server sends the data to the virtual space, makes decisions based on the sensor data, and feeds back the decision information to the real vehicle of the driving test subsystem of the field test layer.

9. The digital twin-based autonomous driving testing method according to claim 6, characterized in that: The process of using the interaction method established in step 2 to perform the interaction between the experimental test layer driving test subsystem and the field test layer driving test subsystem in step 3 includes: First, the factors affecting the road environment, traffic conditions, traffic participants, and environmental sensors are considered, and complex road scenes are modeled to perform 1:1 digital twins and generate test scenarios for the driving test subsystem of the experimental test layer. Then, the corresponding scene information is sent in real time to the real vehicle of the driving test subsystem of the experimental test layer through the 4G / 5G network communication module; The simulation platform analyzes the dynamic characteristics of the vehicle in the virtual space and generates a result report.

Images