WO2020238097A1

WO2020238097A1 - Testing method for autonomous vehicle, device, and system

Info

Publication number: WO2020238097A1
Application number: PCT/CN2019/120934
Authority: WO
Inventors: 李慧云; 石印洲; 林定方; 潘仲鸣
Original assignee: 中国科学院深圳先进技术研究院
Priority date: 2019-05-31
Filing date: 2019-11-26
Publication date: 2020-12-03
Also published as: CN110160804A; CN110160804B

Abstract

A testing method for an autonomous vehicle comprises: an actual vehicle control data acquisition module (101) used to acquire operation control data; a testbed (102) used to generate a reaction force according to the operation control data and scenario information, and acquire, according to the reaction force, state data of the autonomous vehicle; a scenario model establishment module (103) used to establish a traffic scenario model, a traffic flow model, an autonomous vehicle kinetic model, and an environment sensing sensor model, and acquire position data of the autonomous vehicle; and an evaluation module (104) used to calculate, according to the position data and the state data, a report value for evaluating a traveling state of the autonomous vehicle. The invention combines a scenario model comprising traffic flow data and the testbed, and accordingly enables realistic testing of the capabilities of an autonomous vehicle, such as environment identification, identification response time and response manner for dangerous situations, and the safety and stability of an independent recognition capability in a complex traffic environment.

Description

Test method, device and system for automatic driving vehicle

This application claims the priority of the Chinese patent application filed at the Chinese Patent Office on May 31, 2019, with the application number 201910473051.9 and the invention title "A test method, device and system for an autonomous vehicle", and its entire contents Incorporated in this application by reference.

Technical field

This application relates to the technical field of automated driving vehicle testing, and in particular to a testing method, device and system for automated driving vehicles.

Background technique

Intelligent driving is an inevitable trend in the development of autonomous vehicles in the future, and an effective way to avoid human driving errors and improve traffic efficiency. The rapid changes in existing communications, electronics and computer technologies have laid a solid foundation for the development of intelligent driving technology. The Institute of Electrical and Electronics Engineers (IEEE) predicts that by 2040, 75% of autonomous vehicles will be intelligent driving autonomous vehicles. The market growth rate of intelligent driving autonomous vehicles will be 10 times that of other autonomous vehicles, and the emergence of intelligent driving autonomous vehicles will reduce the traffic accident rate to 10%.

In the self-driving vehicle industry, all technologies need to be verified from the laboratory to production. In order for traditional autonomous vehicles to move toward autonomous driving, in addition to the efforts of various technical solutions companies, including but not limited to OEMs (original equipment manufacturers) and autonomous driving companies, it is also necessary to continuously test experimental results and conduct symmetrical debugging and optimization.

Before self-driving vehicles are officially on the road, targeted tests are needed to prove their operational safety. Road testing is undoubtedly the most direct method, but due to the weight and speed of the autonomous vehicle, testing in actual scenarios has major safety hazards, especially before the technology is mature. However, if there is no actual drive test, it is very difficult to update the technology. Therefore, governments, scientific research institutes, and enterprises in various countries have vigorously launched the establishment of the standard system, the construction of the automatic driving test assessment site and the exploration of related test methods.

In the process of testing autonomous vehicles, basic traffic factors such as road pedestrians, participating autonomous vehicles, road infrastructure and traffic lights should be considered. Especially in the open public road test process, due to the mixture of traditional driving self-driving vehicles and other traffic participants, it will increase the complexity of the road traffic of self-driving vehicles, resulting in the safe driving of self-driving vehicles on the test road. Uncertainties have risen, increasing the risk of road traffic accidents. Moreover, autonomous vehicles require approximately 275 million miles of driving distance to prove the safety of the system. In order to verify the performance of self-driving vehicles better than humans, the test mileage of self-driving vehicles needs to reach billions of miles. The traditional road test method not only has a long test period, but also has a large test cost, which cannot meet the needs of the market.

With the development of virtual reality technology, the use of computer three-dimensional modeling methods to build virtual streets, urban and rural areas, highways, etc. as test environments, and add required test cases to the virtual environment, this virtual test method can reduce automatic The development cycle of driving technology. Test scenario use cases are mainly derived from traffic accidents and natural driving data of manned autonomous vehicles, as well as experimental data from human takeover cases and special scenarios simulated in previous tests to verify the operational safety of autonomous vehicles.

However, the current model-in-the-loop and software-in-the-loop simulation tools cannot test the actual execution effects, especially the actual effects including kinematics and dynamics, traffic rules, and vehicle-road coordination. It is not possible to test the self-driving vehicle's ability to recognize the environment, the recognition response time and processing methods of dangerous situations, and the safety and stability of the autonomous cognitive ability in a complex traffic environment.

technical problem

One of the objectives of the embodiments of the present application is to provide a test method, device, and equipment for an automatic driving vehicle, so as to solve the problem that the environment recognition ability of the automatic driving vehicle cannot be tested when the automatic driving vehicle is simulated in the prior art. The identification and response time and processing methods of dangerous situations, and the safety and stability of autonomous cognitive ability and multi-system collaborative work in complex traffic environments.

Technical solutions

In order to solve the above technical problems, the technical solutions adopted in the embodiments of this application are:

The first aspect of the embodiments of the present application provides a test system for an autonomous vehicle. The in-loop test system for an autonomous vehicle includes a real-vehicle control data acquisition module, a test bench, a scenario model establishment module, and an evaluation module. :

The real-vehicle data collection module is used to collect the control data of the control components of the tested autonomous vehicle;

The test bench is used to generate a scene feedback instruction according to the control data of the tested autonomous vehicle and the scene information of the autonomous vehicle, and control the motor of the test bench to generate a response to the vehicle according to the scene feedback instruction. The reaction force of the tested self-driving vehicle, and obtaining state data of the self-driving vehicle according to the reaction force;

The scene model establishment module is used to establish a traffic scene model, a traffic flow model, an autonomous vehicle dynamics model, and an environment perception sensor model, and the environment perception sensor model is used according to the autonomous vehicle dynamics model and the traffic scene model. And a traffic flow model to generate control instructions to control the control components of the autonomous vehicle, and combine the state data to obtain position data of the autonomous vehicle dynamics model in the traffic scene model;

The evaluation module is used to calculate the return value for evaluating the driving state of the autonomous vehicle based on the position data of the autonomous vehicle dynamics model in the traffic scene model and the simulation data of the tested autonomous vehicle .

With reference to the first aspect, in a first possible implementation of the first aspect, the actual vehicle data collection module includes an angle sensor for collecting the rotation angle of the steering wheel, a gear switch for different gears, and a One or more of stroke sensors for collecting pedal stroke.

With reference to the first aspect, in a second possible implementation of the first aspect, the test bench includes one or more of a simulation motor, a test bench control system, a sensor, and a steering loading system, wherein:

The simulation motor is used for speed control simulation and rolling resistance simulation according to the road in the scene;

The test bench control system is used to receive a decision-making instruction from an autonomous vehicle system, or send an instruction to the autonomous vehicle system, or receive a manual input control instruction;

The sensors include a rotational speed sensor and a torque sensor, which are used to detect the rotational speed of the hub of the autonomous vehicle and the transmitted torque;

The steering loading system is used when the automatic driving vehicle is steering, and the steering loading motor generates a torque that prevents the wheels from turning to verify the automatic steering of the automatic driving vehicle.

With reference to the first aspect, in a third possible implementation manner of the first aspect, the evaluation module is specifically configured to, according to the formula:

Calculate the return value used to evaluate the driving state of the autonomous vehicle, where: I[[·]] represents the indicator function, and the value is 1 when the internal conditions of the function are met, otherwise the value is 0, and v is the longitudinal speed of the autonomous vehicle , D ₁ is the distance between the autonomous vehicle and the edge of the lane directly ahead, d ₂ is the distance between the autonomous vehicle and the center axis of the lane, α and β are the constraint parameters when the autonomous vehicle turns, r is the return value, and x1 is the automatic driving The distance threshold between the vehicle and the edge of the lane directly ahead, x2 is the distance threshold between the autonomous vehicle and the center axis of the lane.

With reference to the first aspect, in a fourth possible implementation manner of the first aspect, the test system of the autonomous vehicle further includes a fusion module, and the fusion module is used for sensing data of multiple autonomous vehicles in the scene. , And collect the sensor data of multiple roads in the scene, and fuse the collected multiple sensor data.

A second aspect of the embodiments of the present application provides a test method for an autonomous vehicle based on the test system for an autonomous vehicle according to any one of the first aspect, and the test method for the autonomous vehicle includes:

According to the pre-established traffic scene model and traffic flow model, combined with the environment sensing sensor model, generate control instructions for the autonomous vehicle, and collect control data of the control components of the tested autonomous vehicle based on the control instructions;

Based on the manipulation data, combined with the scene information of the autonomous vehicle, a scene feedback instruction is generated, and the motor of the test bench is controlled according to the scene feedback instruction to generate a reaction force against the autonomous vehicle under test. Acquiring the state data of the autonomous vehicle by the anti-boat force;

According to the state data, determine the position data of the automatic driving vehicle dynamics model in the traffic scene model;

According to the state data and the position data, a return value for evaluating the driving state of the autonomous driving vehicle is calculated.

With reference to the second aspect, in the first possible implementation of the second aspect, the step of calculating a return value for evaluating the driving state of an autonomous vehicle based on the simulation data and the location data includes:

According to the formula:

Calculate the return value used to evaluate the driving state of the autonomous vehicle, where: I[[·]] represents the indicator function, and the value is 1 when the internal conditions of the function are met, otherwise the value is 0, and v is the longitudinal speed of the autonomous vehicle , D ₁ is the distance between the autonomous vehicle and the edge of the lane directly ahead, d ₂ is the distance between the autonomous vehicle and the center axis of the lane, α and β are the constraint parameters when the autonomous vehicle turns, r is the return value, and x1 is the automatic driving The distance threshold between the vehicle and the edge of the lane directly ahead, x2 is the distance threshold between the autonomous vehicle and the lane center axis.

In combination with the second aspect, in a second possible implementation of the second aspect, according to the pre-established traffic scene model and traffic flow model, combined with the environmental perception sensor model, the control instructions for the autonomous vehicle are generated, and based on all The step of collecting the control data of the control component of the tested autonomous vehicle by the control instruction includes:

Establish traffic scene models, traffic flow models including traffic flow data, autonomous vehicle dynamics models, and environmental sensor models;

According to the environment perception data collected by the environment perception sensor model in the traffic scene model and the traffic flow model, a control instruction for the autonomous vehicle is generated.

In a third aspect, an embodiment of the present application provides a testing device for an autonomous vehicle based on the testing system for an autonomous vehicle according to any one of the first aspect, wherein the testing device for an autonomous vehicle includes:

The control data collection unit is used to generate control instructions for the autonomous vehicle based on the pre-established traffic scene model and traffic flow model, combined with the environmental sensing sensor model, and collect the control components of the tested autonomous vehicle based on the control instructions Manipulate data;

The state data collection unit is configured to generate a scene feedback instruction based on the control data, control the motor of the test bench according to the scene feedback instruction to generate a reaction force against the tested autonomous vehicle, and according to the anti-boat To obtain state data of the autonomous vehicle;

A location data determining unit, configured to determine location data of the dynamic model of the autonomous vehicle in the traffic scene model according to the state data;

The reward value calculation unit is used to calculate the reward value for evaluating the driving state of the autonomous vehicle based on the state data and the position data.

A fourth aspect of the embodiments of the present application provides a computer-readable storage medium that stores a computer program that, when executed by a processor, realizes the automatic The steps of the test method for driving a vehicle.

The beneficial effects of the automated driving vehicle test method provided in the embodiments of the present application are: by establishing a scene model including a traffic flow model, real-time detection of the driving environment of the automated driving vehicle through the environment sensing sensor model, and identifying the driving environment of the automated driving vehicle. In response, a control instruction is generated, the control instruction is executed by the autonomous vehicle to be tested, the control data of the autonomous vehicle to be tested is collected, and the test bench is based on the scene information of the autonomous vehicle and the control of the autonomous vehicle Data, generate scene feedback instructions, and control the motor of the test bench according to the scene feedback instructions to generate a reaction force to the autonomous vehicle under test, so as to buffer the momentum of the vehicle and simulate the road buffer in the actual scene The magnitude and mode of force are used to obtain the state data of the autonomous driving vehicle, the position of the automatic loading vehicle is calculated based on the state data, and the return value of the driving state of the autonomous driving vehicle is calculated based on the position data and the state data, Therefore, the test method described in the present application can be combined with a scene model including traffic flow data through a test bench to more realistically reflect the driving state of the autonomous vehicle, and can more reliably test the autonomous vehicle's ability to recognize the environment and dangerous situations The identification response time and processing methods of the system, the safety and stability of autonomous recognition ability in complex traffic environment.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings used in the description of the embodiments or exemplary technologies. Obviously, the accompanying drawings in the following description are only for the application For some embodiments, for those of ordinary skill in the art, other drawings can be obtained from these drawings without creative work.

Figure 1 is a schematic diagram of a test system for an autonomous vehicle provided by an embodiment of the present application;

Figure 2 is a schematic diagram of a real-vehicle data collection module provided by an embodiment of the present application;

Figure 3 is a schematic structural diagram of a test bench provided by an embodiment of the present application;

FIG. 4 is a schematic diagram of state information for designing a reward function provided by an embodiment of the present application;

FIG. 5 is a schematic diagram of speed and return value constraints provided by an embodiment of the present application;

FIG. 6 is a schematic diagram of an implementation process of a test method for an autonomous vehicle provided by an embodiment of the present application;

FIG. 7 is a schematic diagram of a test device for an automatic driving vehicle provided by an embodiment of the present application;

Fig. 8 is a schematic diagram of a test device for an autonomous vehicle provided by an embodiment of the present application.

Embodiments of the invention

In order to make the purpose, technical solutions, and advantages of this application clearer, the following describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not used to limit the present application.

In order to illustrate the technical solutions described in the present application, specific embodiments are used for description below.

Fig. 1 is a schematic structural diagram of a test system for an autonomous vehicle provided by an embodiment of the application, and the details are as follows:

The test system of the autonomous vehicle includes a real-vehicle control data acquisition module 101, a test bench 102, a scene model establishment module 103, and an evaluation module 104, in which:

The real-vehicle data collection module 101 is used to collect the control data of the control components of the tested autonomous vehicle;

The test bench 102 is used to generate a scene feedback instruction according to the control data of the tested autonomous vehicle and the scene information of the autonomous vehicle, and control the motor of the test bench to generate a pair according to the scene feedback instruction. Obtaining the state data of the autonomous vehicle based on the reaction force of the tested autonomous vehicle;

The scene model establishment module 103 is used to establish a traffic scene model, a traffic flow model, an autonomous vehicle dynamics model, and an environment perception sensor model, and the environment perception sensor model is used according to the autonomous vehicle dynamics model and the traffic scene. The model and the traffic flow model generate control instructions to control the control components of the autonomous vehicle, and combine the state data to obtain position data of the autonomous vehicle dynamics model in the traffic scene model;

The evaluation module 104 is configured to calculate a return for evaluating the driving state of the autonomous vehicle based on the position data of the autonomous vehicle dynamics model in the traffic scene model and the simulation data of the tested autonomous vehicle value.

Specifically, the real-vehicle data acquisition module 101 may be installed in an autonomous vehicle, and sensors such as lidar and cameras may be installed in the autonomous vehicle. As shown in FIG. 2, the actual-vehicle data acquisition module may include Angle sensor, rotational speed sensor, linear acceleration sensor, angular velocity sensor, tension sensor, etc. among them:

The vehicle controller sends a rotation command to the steering column motor controller to control the rotation of the steering column motor, and collects the rotation angle of the steering column through an angle sensor and sends it to the vehicle controller;

The vehicle controller sends a drive command to the drive motor controller or the transmitter throttle controller, and the drive motor or engine drives the wheels to rotate. The rotational speed of the wheels can be collected by the rotational speed sensor and sent to the vehicle controller;

The vehicle controller can send a brake command to the brake pedal control device. When the brake pedal actuator executes the brake command, the tension sensor can collect the tension signal, and the braking amplitude data can be calculated based on the tension signal, or the stroke sensor can also collect the pedal stroke.

Of course, as shown in Figure 2, it may also include a linear acceleration sensor to collect the autonomous driving acceleration of the autonomous vehicle, and an angle sensor to collect the turning speed of the autonomous vehicle, or it may also include the vehicle controller according to the automatic driving An automatic lighting system for lighting control of the vehicle's driving state.

Fig. 3 is a schematic diagram of a test bench provided by this application. As shown in Fig. 3, the test bench 102 is aimed at testing and testing requirements of driving, steering, and braking systems of an unmanned autonomous vehicle. This application proposes A set of multifunctional axle coupling test platform, which can be used for comprehensive testing and verification of the whole vehicle, or a sub-system can be tested and verified separately. By importing traffic flow data and scene information, according to the specific road information in the scene information, such as The slope of the road, the road surface material, etc., combined with the pre-statistical friction force corresponding to the vehicle under different driving conditions, and the friction force action method, to control the reaction force of the test bench on the autonomous vehicle, so as to This makes the test of autonomous vehicles more consistent with the actual situation. By importing traffic flow data from real scenes, the autonomous vehicles can be trained more effectively for emergencies in the actual scene, and can test the environment of the autonomous vehicles. Recognition ability, recognition response time and processing methods of dangerous situations, and autonomous cognition ability in complex traffic environment and the safety and stability of multi-system collaborative work. The test bench can make the simulated test environment close to the real environment. It can be applied to the fields of vehicle performance testing, durability, transmission system testing, ADAS and driverless development testing.

The test bench may include a drive motor, which can be connected to an industrial power grid, or can also be connected to a vehicle-mounted power battery. The drive motor can be a DC, AC or permanent magnet motor, and the drive motor needs to meet the requirements of overload, For low inertia requirements, the power can be 80-150kw, and the drive motor can be equipped with an inverter and an AC/DC AC/DC converter. According to the model to be tested, a 2-drive (front-drive/rear-drive) or 4-drive version can be provided to meet the test requirements. It can simulate road rolling resistance and speed control according to the scene model to meet the test cycle requirements of different working conditions.

The test bench may include a test bench control system, a speed sensor, a torque sensor, a steering loading system, etc., where:

The test bench control system can adopt PLC, which can be used to receive and execute decision instructions issued by the vehicle control system, can manually input and execute driving, steering, and brake instructions, and can issue instructions to the vehicle control system.

The speed sensor and torque sensor can be equipped with sensors according to the torque test requirements of different accuracy levels, used to test the speed of the autonomous driving vehicle hub, and pass flanges, universal joints, spline shafts, spline sleeves, and universal joints The maximum speed and torque can be detected and recorded during the brake test with the torque transmitted by the accessories. Or it can also include a dynamometer, which can be connected to the hub via a bearing support and a flywheel, and can include a direction angle detection system and a loading system for detecting the direction angle of the hub.

The steering loading system can be used to test EPS (electronically controlled electric power steering system). When the autonomous vehicle is turning, the control system of the test bench sends a command to the steering loading motor, and the loading motor gives a torque that prevents the wheels from turning (simulating the resistance torque between the wheels and the ground), verifying the automatic steering column steering effect of the autonomous vehicle .

The scene model establishment module 103 described in this application can be used to establish a traffic scene model, a traffic flow model, an autonomous vehicle dynamics model, and an environment sensing sensor model, among which:

The traffic scene model can be based on a simulator such as Prescan to establish a complex scene autonomous driving traffic scene modeling, covering road models, environment models, road user models, weather lighting models, etc. Based on the import of external map data, the road slope, curvature, inclination and other information can be obtained to complete the parameterized pavement model with the same height of the real road, which can be used for automated testing; complex road network structure modeling, such as three-way intersections, overpasses, etc. The traffic scene model may include information such as road surface and roadside facilities, traffic signs, buildings, and green belts. Weather and illumination models can include day and night and sunlight, rain, snow and fog weather, and car lights and street lights models.

The traffic flow model can generate high-fidelity background traffic flow information by adopting city-level macroscopic traffic flow data, regression or Generative Adversarial Networks (GAN) learning methods. Import the data into the microscopic traffic flow tool Vissim, etc. to generate a microscopic traffic flow data environment, and inject the microscopic traffic flow data into the self-built autonomous vehicle dynamics simulation platform Matlab Simulink to obtain a traffic flow model.

The autonomous driving vehicle dynamics model may be performed by 2D or 3D dynamic modeling of the autonomous driving vehicle, and the autonomous driving vehicle dynamics model may include dynamic simulation models of the braking system, steering system, and suspension system It is established to realize the longitudinal acceleration/deceleration and lateral movement of the autonomous vehicle. It can be described by the dynamics of the autonomous vehicle based on Model Predictive Control (MPC) or PID controller (proportional-integral-derivative controller).

The environmental sensing sensor model can be used to collect signals. After the collected signals are fused, they are transmitted to the decision-making layer, and finally to the entire actuator. The types of sensors can include cameras, millimeter wave radars, ultrasonic radars, lidars, and DGPS+IMU sensors that require positioning. These sensors can be used to detect some surrounding traffic scenes, such as self-driving vehicles and pedestrians. It can also be used for path planning and positioning, confirming the position of the self-driving vehicle on the road, and combining the detected driving area to perform local Path planning, and finally control the power brake steering system of the entire autonomous vehicle.

The evaluation module can be used to test the performance of single-vehicle decision-making control, vehicle-road collaborative testing, large-scale autonomous vehicle wireless communication technology and capabilities, and vehicle networking communication congestion control strategy testing. The state information on the design state of the evaluation function may include: The angle between the driving direction of the vehicle and the longitudinal axis of the lane coordinate system

The speed of the autonomous vehicle along the horizontal axis of the vehicle body coordinate system, that is, the forward speed ν, the distance d ₁ of the autonomous vehicle from the edge of the lane within 200m, and the distance between the origin of the autonomous vehicle body coordinate system and the lane axis The lateral offset distance d ₂ . d ₂ =0 indicates that the autonomous vehicle is on the lane axis, |d ₂ |=1 indicates that the autonomous vehicle is on the edge of the lane, and |d ₂ |>1 indicates that the autonomous vehicle is outside the lane.

Figure 4 shows a schematic diagram of the state information used for the design of the reward function. The thick lines on the upper and lower sides in the figure indicate the edge of the lane, and the dotted line in the middle indicates the central axis of the lane. The angle between the longitudinal speed v of the autonomous vehicle and the central axis of the lane is

The distance d ₁ indicated by the green line is the distance between the autonomous vehicle and the edge of the lane directly ahead, and the distance d ₂ indicated by the blue line is the distance from the origin of the vehicle body coordinate system to the center axis of the lane.

The design of the reward function described in this application includes the following considerations: (a) It is hoped that the longitudinal speed v of the autonomous driving vehicle can be as large as possible; (b) It is hoped that the forward direction of the autonomous vehicle is as consistent as possible with the direction of the lane, that is, angle

Try to approach zero as much as possible; (c) hope that the distance d ₁ between the autonomous vehicle and the edge of the lane ahead is as large as possible; (c) hope that the autonomous vehicle is located as close to the center axis of the lane as possible during driving, that is, d ₂ tends to zero; (d) It is hoped that the self-driving vehicle can predict the curve before the curve, brake in advance and slow down, so as to ensure safe cornering without breaking out of the track. Based on the above expectations for the performance of autonomous vehicles, this article gives a reward function in the following form:

Calculate the return value used to evaluate the driving state of the autonomous vehicle, where: I[[·]] represents the indicator function, and the value is 1 when the internal conditions of the function are met, otherwise the value is 0, and v is the longitudinal speed of the autonomous vehicle , D ₁ is the distance between the autonomous vehicle and the edge of the lane ahead, d ₂ is the distance between the autonomous vehicle and the center axis of the lane, α and β are the constraint parameters of the autonomous vehicle when turning, r is the return value, and x1 is the automatic driving The distance threshold between the vehicle and the edge of the lane ahead, x2 is the distance threshold between the autonomous vehicle and the central axis of the lane. For example, X1 can be set to 10, and X2 is set to 50. It can be adjusted according to the width of the road and the safety level of the road. The values of X1 and X2 can be adapted to the test requirements of different roads. The constraint parameters α and β can be determined based on the reference initial value and the continuous optimization method combined with the actual measurement result.

The reward function described in this application adopts the form of product instead of the form of summation. Through the form of product, the reward function is made to have super-linear properties and is more sensitive to changes in the behavior of autonomous vehicles, thereby enabling autonomous vehicles to be driven more quickly The behavior meets expectations.

The product terms of the reward function respectively correspond to the five expectations for the behavior of autonomous vehicles described above. Because it is hoped that the longitudinal speed of the self-driving vehicle can be as large as possible, v is used as a product term. Under the premise that it is not affected by other product terms, the larger v is, the greater the reward value is obtained.

with

The term constrains the angle between the forward direction of the autonomous vehicle and the direction of the lane

Tends to zero when

The item is 1, it will not affect the result; when

Will reduce the return value. The distance between the autonomous vehicle and the edge of the lane directly ahead is constrained by the following product terms:

Wherein, the distance threshold x2 between the autonomous vehicle and the edge of the front lane can be set to 50 meters, when d ₁ >50, the value of this item is 1, which will not affect the result of the return value; and when d ₁ ≤50, the value of this item is less than 1, which will reduce the reward value, and the smaller d _{1 is, the} smaller the reward value will be.

(1-|d ₂ |) constrains the autonomous vehicle to be located in the lane (refers to the lane in the direction of travel, for the lane in the one-way direction, it refers to the entire lane, for the lane in the two-way direction, it refers to the lane in the current direction of the autonomous vehicle ) Near the central axis, when d ₂ = 0 meters, that is, the autonomous vehicle is located at the central axis of the lane, and the (1-|d ₂ |) term is 1, which will not affect the result. d ₂ ≠ 0 meters, the autonomous vehicle Deviating from the position of the center axis of the lane, (1-|d ₂ |)<1, will reduce the return value or even make the return value negative.

The behavior constraints of the autonomous vehicle when cornering are realized by the compound term shown in the following formula:

Among them, α and β are the turning constraint parameters of the autonomous vehicle that need to be adjusted according to the experimental results. This item comprehensively considers the speed constraints of the automatic driving vehicle before and when entering the curve, and d ₁ =10 can be set as the limit value of whether or not the curve is encountered. When d ₁ ＞10, the self-driving vehicle does not encounter a curve, and the compound term degenerates into a speed term. At this time, the greater the speed, the greater the return value; when d ₁ ≤10, the term is the second order of the speed v function.

When the constraint parameter α=120, β=180, draw the curve of the quadratic function, that is, the constraint curve of speed and return value, as shown in Figure 5. At this time, the curve of the quadratic function has a maximum value. v = 181/2, which means that when the autonomous vehicle encounters a curve, if the value of the composite item is to be as large as possible, the speed of the autonomous vehicle should be as close as possible to 90.5km/h, thus limiting the automatic driving vehicle in The maximum speed when passing a curve prevents the autonomous vehicle from rushing out of the track due to excessive speed. In addition, the speed of the self-driving vehicle before entering the curve may be much greater than 90.5km/h. Once d ₁ reaches the set boundary value, the self-driving vehicle will automatically brake and decelerate, thus realizing the automatic driving of predicting the curve and braking control. In the actual situation, the constraint parameters α and β can be optimized according to the actual driving data of the autonomous vehicle.

In an embodiment of the present application, when the sensor data is fused, the fusion module may also be included. The fusion module is used to integrate the sensor data of multiple autonomous vehicles in the scene and the multiple sensors in the scene. The sensor data of the road is collected, and the collected multiple sensor data are merged. Multi-sensor information fusion improves the effectiveness of the system by coordinating, combining, and complementing the information obtained by multiple sensors, and achieves better performance than a single sensor. Multi-sensor information can be fused by methods such as multi-source information distributed parallel fusion method, Kalman filter method, dynamic and static filter technology, interactive adaptive, factor graph method, etc., to reduce the amount of high-precision sensor data and improve the accuracy of perception fusion.

FIG. 6 is a schematic diagram of the implementation process of an automatic driving vehicle test method based on the above-mentioned automatic driving vehicle test system provided by an embodiment of the application, and the details are as follows:

In step S601, according to the pre-established traffic scene model and traffic flow model, combined with the environmental sensing sensor model, a control instruction for the autonomous vehicle is generated, and the control data of the control component of the tested autonomous vehicle is collected based on the control instruction ；

Wherein, the control instruction can inject traffic flow data into the autonomous vehicle simulation platform through the established traffic scene model, traffic flow model including traffic flow data, autonomous driving vehicle dynamics model, and environment sensing sensor model. The driving vehicle simulation platform establishes a traffic scene model including traffic flow data, detects the environment perception data collected by the traffic scene model through the environment perception sensor model, and generates control instructions for the autonomous vehicle.

In step S602, based on the manipulation data and combined with the scene information of the autonomous vehicle, a scene feedback instruction is generated, and the motor of the test bench is controlled according to the scene feedback instruction to generate a signal for the autonomous vehicle under test. The reaction force, obtaining the state data of the autonomous vehicle according to the reaction force;

According to the manipulation data, combined with the road information of the traffic scene model, the autonomous vehicle is tested on the test bench, and the road parameters of the vehicle are determined according to the road data included in the scene information, including Such as the slope of the road, the material of the road, the weather conditions (rainy, tomorrow, snow), etc., according to the road parameters, combined with the control data of the vehicle, determine the reaction force acting on the autonomous vehicle, that is, the action Due to the friction of the autonomous vehicle, the autonomous vehicle can simulate the actual test scenario more realistically and generate more realistic state data of the autonomous vehicle, including the vehicle's speed, acceleration, angular velocity, etc.

In step S603, according to the state data, determine the position data of the autonomous vehicle dynamics model in the traffic scene model;

According to the state data, the displacement information of the autonomous vehicle dynamics model in the traffic scene can be determined, and the position data of the autonomous vehicle driving the autonomous vehicle can be determined according to the displacement information, including the autonomous vehicle mentioned above The angle between the direction of travel and the axis of the lane

The distance d ₁ between the autonomous vehicle and the edge of the lane ahead, the distance d ₂ from the origin of the vehicle body coordinate system to the center axis of the lane, etc.

In step S604, according to the state data and the position data, a reward value for evaluating the driving state of the autonomous driving vehicle is calculated.

According to the vehicle speed, acceleration, angular velocity and other information included in the state data, as well as the position data, the return value of a single vehicle can be calculated. One of the calculation methods can be:

with

Item constraint the angle between the forward direction of the autonomous vehicle and the direction of the lane

Tends to zero when

The item is 1, it will not affect the result; when

Wherein, the distance threshold x2 between the autonomous vehicle and the edge of the front lane can be set to 50 meters. When d ₁ >50, the value of this item is 1, which will not affect the result of the return value; and when d ₁ ≤50, the value of this item is less than 1, which will reduce the reward value, and the smaller d _{1 is, the} smaller the reward value will be.

When the constraint parameter α=120, β=180, the graph of the quadratic function is drawn as shown in Figure 5. At this time, there is a maximum point in the graph of the quadratic function, which is obtained at v=181/2, which means When the autonomous vehicle encounters a curve, if the value of the compound item is to be as large as possible, the speed of the autonomous vehicle should be as close as possible to 90.5km/h, thus limiting the maximum speed of the autonomous vehicle when passing the curve, and avoid The autonomous vehicle drove off the track due to excessive speed. In addition, the speed of the self-driving vehicle before entering the curve may be much greater than 90.5km/h. Once d ₁ reaches the set boundary value, the self-driving vehicle will automatically brake and decelerate, thus realizing the automatic driving of predicting the curve and braking control.

Of course, the calculation formula of the above evaluation function can also be modified according to the weight of speed or smoothness, so as to obtain the calculation result of the evaluation function that meets actual use requirements.

The test method of the autonomous vehicle described in FIG. 6 corresponds to the test system of the autonomous vehicle described in FIG. 1.

FIG. 7 is a schematic structural diagram of a test device for an autonomous driving vehicle provided by an embodiment of the application. The test device for an autonomous vehicle includes:

The control data acquisition unit 701 is used to generate control instructions for the autonomous vehicle based on the pre-established traffic scene model and traffic flow model in combination with the environmental sensor model, and collect the control components of the tested autonomous vehicle based on the control instructions Control data;

The state data collection unit 702 is configured to generate a scene feedback instruction based on the control data, control the motor of the test bench according to the scene feedback instruction to generate a reaction force to the tested autonomous vehicle, and according to the reaction To obtain state data of the autonomous vehicle;

The position data determining unit 703 is configured to determine the position data of the dynamic model of the autonomous driving vehicle in the traffic scene model according to the state data;

The reward value calculation unit 704 is configured to calculate a reward value for evaluating the driving state of the autonomous vehicle based on the state data and the position data.

The testing device for the autonomous vehicle corresponds to the testing method for the autonomous vehicle.

Fig. 8 is a schematic diagram of a test device for an autonomous vehicle provided by an embodiment of the present application. As shown in FIG. 8, the test device 8 for an autonomous vehicle of this embodiment includes a processor 80, a memory 81, and a computer program 82 stored in the memory 81 and running on the processor 80, such as an automatic Test procedures for driving vehicles. When the processor 80 executes the computer program 82, the steps in the foregoing embodiments of the test method for autonomous vehicles are implemented. Alternatively, when the processor 80 executes the computer program 82, the functions of the modules/units in the foregoing device embodiments are realized.

Exemplarily, the computer program 82 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 81 and executed by the processor 80 to complete This application. The one or more modules/units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 82 in the test device 8 of the autonomous vehicle. For example, the computer program 82 can be divided into:

A location data determining unit, configured to determine location data of an autonomous vehicle dynamics model in a traffic scene model according to the state data;

The test equipment of the autonomous vehicle may include, but is not limited to, a processor 80 and a memory 81. Those skilled in the art can understand that FIG. 8 is only an example of the test device 8 of an autonomous driving vehicle, and does not constitute a limitation on the test device 8 of an autonomous vehicle. It may include more or less components than shown in the figure, or a combination Certain components, or different components, for example, the test device of the autonomous vehicle may also include input and output devices, network access devices, buses, and so on.

The so-called processor 80 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.

The memory 81 may be an internal storage unit of the test device 8 of the autonomous driving vehicle, such as a hard disk or memory of the test device 8 of the autonomous vehicle. The memory 81 may also be an external storage device of the test device 8 of the autonomous vehicle, such as a plug-in hard disk or a smart memory card (Smart Media Card, SMC) equipped on the test device 8 of the autonomous vehicle. Secure Digital (SD) card, Flash Card, etc. The memory 81 may also include both the internal storage unit of the test device 8 of the autonomous vehicle and an external storage device. The memory 81 is used to store the computer program and other programs and data required by the test equipment of the autonomous vehicle. The memory 81 can also be used to temporarily store data that has been output or will be output.

Those skilled in the art can clearly understand that for the convenience and conciseness of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned functions can be allocated to different functional units and modules as required. Module completion means dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist alone physically, or two or more units can be integrated into one unit. The above-mentioned integrated units can be hardware-based Formal realization can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the present application. For the specific working process of the units and modules in the foregoing system, reference may be made to the corresponding process in the foregoing method embodiment, which is not repeated here.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not detailed or recorded in an embodiment, reference may be made to related descriptions of other embodiments.

A person of ordinary skill in the art may realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software depends on the specific application and design constraint conditions of the technical solution. Professionals and technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of this application.

In the embodiments provided in this application, it should be understood that the disclosed device/terminal device and method may be implemented in other ways. For example, the device/terminal device embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units. Or components can be combined or integrated into another system, or some features can be omitted or not implemented. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.

In addition, the functional units in each embodiment of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or software functional unit.

If the integrated module/unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, this application implements all or part of the processes in the above-mentioned embodiments and methods, and can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the program is executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file, or some intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory) , Random Access Memory (RAM, Random Access Memory), electrical carrier signal, telecommunications signal, and software distribution media. It should be noted that the content contained in the computer-readable medium can be appropriately added or deleted in accordance with the requirements of the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to the legislation and patent practice, the computer-readable medium Does not include electrical carrier signals and telecommunication signals.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still implement the foregoing The technical solutions recorded in the examples are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application, and should be included in Within the scope of protection of this application.

Claims

A test system for an autonomous vehicle, characterized in that the test system for an autonomous vehicle includes a real vehicle control data acquisition module, a test bench, a scene model establishment module, and an evaluation module, wherein:

The real-vehicle data collection module is used to collect the control data of the control components of the tested autonomous vehicle;

The test bench is used to generate a scene feedback instruction according to the control data of the tested autonomous vehicle and the scene information of the autonomous vehicle, and control the motor of the test bench to generate a response to the vehicle according to the scene feedback instruction. The reaction force of the tested self-driving vehicle, and obtaining state data of the self-driving vehicle according to the reaction force;

The scene model establishment module is used to establish a traffic scene model, a traffic flow model, an autonomous vehicle dynamics model, and an environment perception sensor model, and the environment perception sensor model is used to establish a traffic scene model based on the autonomous vehicle dynamics model and the traffic scene model And a traffic flow model to generate control instructions to control the control components of the autonomous vehicle, and to obtain position data of the autonomous vehicle dynamics model in the traffic scene model in combination with the state data;

The evaluation module is used to calculate the return value for evaluating the driving state of the autonomous vehicle based on the position data of the autonomous vehicle dynamics model in the traffic scene model and the simulation data of the tested autonomous vehicle .
The test system for an autonomous vehicle according to claim 1, wherein the actual vehicle data collection module includes an angle sensor for collecting the rotation angle of the steering wheel, for adopting gear switches of different gears, and One or more of stroke sensors for collecting pedal stroke.
The test system for an autonomous vehicle according to claim 1, wherein the test bench includes one or more of a simulation motor, a test bench control system, a sensor, and a steering loading system, wherein:

The simulation motor is used for speed control simulation and rolling resistance simulation according to the road in the scene;

The control system of the test bench is used to receive a decision-making instruction of an autonomous vehicle system, or send an instruction to the autonomous vehicle system, or receive a manual input control instruction;

The sensors include a rotational speed sensor and a torque sensor, which are used to detect the rotational speed of the hub of the autonomous vehicle and the transmitted torque;

The steering loading system is used when the automatic driving vehicle is steering, and the steering loading motor generates a torque that prevents the wheels from turning to verify the automatic steering of the automatic driving vehicle.
The test system for an autonomous vehicle according to claim 1, wherein the evaluation module is specifically configured to, according to the formula:

Calculate the return value used to evaluate the driving state of autonomous vehicles, where:
Represents the indicator function, when the internal conditions of the function are met, the value is 1, otherwise the value is 0, v is the longitudinal speed of the autonomous vehicle, d 1 is the distance between the autonomous vehicle and the edge of the lane ahead, and d 2 is the autonomous driving The distance between the vehicle and the central axis of the lane, α and β are the constraint parameters of the autonomous vehicle when turning, r is the return value, x2 is the distance threshold between the autonomous vehicle and the edge of the lane directly ahead, and x1 is the distance between the autonomous vehicle and the central axis of the lane Distance threshold.
The test system of an autonomous vehicle according to claim 1, wherein the test system of the autonomous vehicle further comprises a fusion module, and the fusion module is used for sensing data of multiple autonomous vehicles in the scene. , And collect the sensor data of multiple roads in the scene, and fuse the collected multiple sensor data.
A testing method for an autonomous vehicle based on the testing system for an autonomous vehicle according to any one of claims 1 to 5, wherein the testing method for the autonomous vehicle comprises:

According to the pre-established traffic scene model and traffic flow model, combined with the environment sensing sensor model, generate control instructions for the autonomous vehicle, and collect control data of the control components of the tested autonomous vehicle based on the control instructions;

Based on the manipulation data, combined with the scene information of the autonomous vehicle, a scene feedback instruction is generated, and the motor of the test bench is controlled according to the scene feedback instruction to generate a reaction force against the autonomous vehicle under test. Acquiring the state data of the autonomous vehicle by the anti-boat force;

According to the state data, determine the position data of the automatic driving vehicle dynamics model in the traffic scene model;

According to the state data and the position data, a return value for evaluating the driving state of the autonomous driving vehicle is calculated.
The method for testing an autonomous vehicle according to claim 6, wherein the step of calculating a return value for evaluating the driving state of the autonomous vehicle based on the simulation data and the position data comprises:

According to the formula:

Calculate the return value used to evaluate the driving status of autonomous vehicles, where:
Represents the indicator function, when the internal conditions of the function are met, the value is 1, otherwise the value is 0, v is the longitudinal speed of the autonomous vehicle, d 1 is the distance between the autonomous vehicle and the edge of the lane ahead, and d 2 is the autonomous driving The distance between the vehicle and the center axis of the lane, α and β are the constraint parameters when the autonomous vehicle turns, r is the return value, x1 is the distance threshold between the autonomous vehicle and the edge of the lane ahead, and x2 is the distance between the autonomous vehicle and the center axis of the lane Distance threshold.
The method for testing an autonomous vehicle according to claim 6, characterized in that, according to the pre-established traffic scene model and traffic flow model, combined with the environment sensing sensor model, the control instructions for the autonomous vehicle are generated and based on all The step of collecting the control data of the control component of the automated driving vehicle under test by the control instruction includes:

Establish traffic scene models, traffic flow models including traffic flow data, autonomous vehicle dynamics models, and environmental sensor models;

According to the environment perception data collected by the environment perception sensor model in the traffic scene model and the traffic flow model, a control instruction for the autonomous vehicle is generated.
An automatic driving vehicle test device based on the automatic driving vehicle test system according to any one of claims 1 to 5, wherein the automatic driving vehicle test device comprises:

The control data collection unit is used to generate control instructions for the autonomous vehicle based on the pre-established traffic scene model and traffic flow model, combined with the environmental sensing sensor model, and collect the control components of the tested autonomous vehicle based on the control instructions Manipulate data;

The state data collection unit is configured to generate a scene feedback instruction based on the control data, control the motor of the test bench according to the scene feedback instruction to generate a reaction force against the tested autonomous vehicle, and according to the anti-boat To obtain state data of the autonomous vehicle;

A location data determining unit, configured to determine location data of an autonomous vehicle dynamic model in a traffic scene model according to the state data;

The reward value calculation unit is used to calculate the reward value for evaluating the driving state of the autonomous vehicle based on the state data and the position data.
A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, realizes the test of the autonomous vehicle according to any one of claims 6 to 8 Method steps.