An interactive adversarial testing method and system for an intelligent driving system

Abstract

This application relates to the field of intelligent driving function testing and evaluation technology, specifically, to a method and system for interactive adversarial testing of an intelligent driving system. The method includes: setting up a test scenario containing adversarial targets and traffic lights according to test requirements; installing a first control device and a first signal acquisition device on the adversarial targets, and installing a second control device on the traffic lights; the vehicle under test is equipped with an intelligent driving system and has the second signal acquisition device installed; driving the adversarial targets and traffic lights to generate adversarial actions against the vehicle under test; and testing whether the vehicle under test passes the test based on the state data of the vehicle under test, the state data of the adversarial targets, weather environment data, and road environment data. This application proposes testing tools, testing procedures, testing data, and a test result evaluation scheme, which can test and verify the safe driving capabilities of a vehicle in the presence of dynamic traffic interaction.

Description

Technical Field

[0001] This application relates to the field of intelligent driving function testing and evaluation technology, and more specifically, to an interactive confrontation testing method and system for an intelligent driving system. Background Technology

[0002] With the development of intelligent connected vehicle technology, the operational safety of autonomous driving systems has become a core concern in the industry. Specifically, when activated, autonomous driving systems must ensure they do not pose unreasonable traffic safety risks to vehicle occupants or other road users, while strictly adhering to road traffic rules and meeting relevant road traffic safety management requirements. This necessitates that autonomous driving systems possess the ability to interact efficiently and safely with road users, traffic lights, and other relevant entities.

[0003] Currently, most real-vehicle testing technologies for intelligent driving vehicles focus on testing the functions and performance of a single vehicle, such as verifying individual modules like perception, decision-making, and control. However, driving behavior in real-world road environments often involves complex interactions and confrontations between multiple vehicles. Existing testing methods struggle to fully simulate and evaluate the capabilities of autonomous driving systems in multi-vehicle interaction scenarios, resulting in a significant mismatch between testing scenarios and real-world interaction and confrontation requirements. Therefore, there is an urgent need for a method that can effectively test and evaluate the performance of intelligent driving systems in interactive and confrontational scenarios to comprehensively verify their ability to recognize and respond to traffic participants, traffic lights, obstacles, and other factors. Summary of the Invention

[0004] The purpose of this application is to provide an interactive confrontation test method and system for intelligent driving systems, which clearly proposes test tools, test procedures, test data and test result evaluation schemes, and can test and verify the safe driving ability of vehicles in the presence of dynamic traffic interaction.

[0005] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a method for interactive adversarial testing of an intelligent driving system, including: Arrange a test scenario containing adversarial targets and traffic lights according to the test requirements; A first control device and a first signal acquisition device are installed on the target object, and a second control device is installed on the traffic light; the vehicle under test is equipped with an intelligent driving system and a second signal acquisition device is installed. The first control device and the second control device respectively send countermeasure commands that match the state of the vehicle under test, so as to drive the countermeasure target and the traffic lights to generate countermeasure actions against the vehicle under test. During the confrontation, the status data of the tested vehicle is collected in real time by the second signal acquisition device; the status data of the confrontation target is collected in real time by the first signal acquisition device. Based on the status data of the vehicle under test, the status data of the target object, the weather environment data, and the road environment data, the test determines whether the vehicle under test passes the test.

[0006] Secondly, this application provides an interactive confrontation testing system for an intelligent driving system, comprising: The power supply unit includes the power supply inside the vehicle under test and the power supply inside the countermeasure target. The signal extraction unit includes a visual sensor, a sound sensor, a vehicle perception signal extraction module, an inertial navigation differential positioning module, and a real-time communication module configured on the vehicle under test; and an inertial navigation differential positioning module and a real-time communication module configured on the adversarial target. The data analysis unit includes an acoustic-optical analysis module and a real-time computing module; The synchronous control and storage unit includes a driving robot configured on the adversarial target, an industrial control computer and a test personnel operation module configured on the vehicle under test, and a second control device configured in a traffic light. The interactive confrontation test system for intelligent driving systems is used to execute interactive confrontation test methods for intelligent driving systems.

[0007] Compared with the prior art, this application has the following technical effects: The purpose of this application is to provide a method for interactive adversarial testing of intelligent driving systems. It clearly proposes testing tools, testing procedures, testing data, and a test result evaluation scheme, capable of testing and verifying the safe driving capabilities of vehicles under dynamic traffic interaction conditions. This method differs from traditional real-vehicle testing methods in that: (1) The interactive real-time communication design between the vehicle under test, the target object and the traffic lights ensures the real-time communication of the three status information, including position coordinates, speed, angle and other status information. It can realize the automatic execution of the adversarial behavior by controlling the target object and the traffic lights through the state machine, construct complex dynamic test scenarios, and realize simulated traffic flow test. (2) By coordinating the first signal acquisition device and the second signal acquisition device, the ability of the tested vehicle to identify traffic participants and traffic facilities can be determined; (3) This application clarifies the evaluation method for intelligent driving vehicle test results in interactive scenarios involving traffic participants and traffic facilities. This application avoids the shortcomings of general real-vehicle test schemes, such as asynchronous data recording, missing perception signal data, and evaluation indicators limited to collisions. It meets the multi-dimensional testing needs of intelligent driving vehicles and proposes a clear testing and evaluation method for intelligent driving vehicles in the presence of traffic interactions. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0009] Figure 1 This is a flowchart of an interactive confrontation testing method for an intelligent driving system provided in this embodiment; Figure 2 This is a schematic diagram of the countermeasure target, the tested vehicle, and the traffic light provided in the embodiments of this application; Figure 3 This is a structural diagram of an interactive confrontation test system for an intelligent driving system provided in an embodiment of this application. Detailed Implementation

[0010] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0011] Figure 1 This is a flowchart of an interactive confrontation test method for an intelligent driving system provided in this embodiment, which is applicable to road testing of vehicles equipped with intelligent driving systems.

[0012] See Figure 1 The method provided in this application includes: S110. Arrange a test scenario containing adversarial targets and traffic lights according to the test requirements.

[0013] The test site is set up according to the testing requirements, and the test site meets the necessary conditions for the normal activation of the intelligent driving system of the vehicle under test. For example, the test site is a concrete and asphalt surface with good adhesion. The following traffic elements are designed and arranged in the site, including: traffic signs, traffic markings, traffic lights, and adversarial targets. The traffic elements can be added or removed according to the scenario. See the description below for details: (1) Traffic sign design and layout.

[0014] The traffic signs were designed in accordance with the requirements of GB 5678.2 "Road Traffic Signs and Markings Part 2: Road Traffic Signs" regarding the color, shape, characters, size, and graphics of traffic signs, and installed on the support devices of the test site road in accordance with the requirements of GB 5678.2 "Road Traffic Signs and Markings Part 2: Road Traffic Signs" regarding the position, height, and angle of traffic signs.

[0015] (2) Design and layout of traffic markings.

[0016] Traffic markings were designed in accordance with the requirements for color, shape, characters, graphics, and size of traffic markings in GB 5678.3 "Road Traffic Signs and Markings Part 3: Road Traffic Markings". Clear and visible traffic markings were set according to the road width, road type, road geometry, and other road characteristics required for the test scenario.

[0017] (3) Design and layout of traffic lights.

[0018] Traffic lights include those for motor vehicles, non-motor vehicles, and pedestrian crossings. Traffic lights were installed on the test site road according to the requirements of GB14886 "Specifications for the Installation and Setting of Road Traffic Signals," which outlines the conditions, combinations, and installation methods for traffic lights. After installing a second control device within the traffic lights, the colors, numbers, icons, and other displayed information of the traffic lights can change in real time according to the second control command.

[0019] (4) Design and deployment of countermeasures targets.

[0020] The countermeasure target refers to targets other than the vehicle being tested, including motor vehicles, non-motor vehicles, pedestrians, and obstacles. Motor vehicles and non-motor vehicles should be mass-produced vehicles or flexible targets with surface characteristics representative of such vehicles and capable of being fitted with the first control device. Pedestrians should be flexible adults and children of average human size, capable of being fitted with the first control device. After the first control device is installed, the lateral and longitudinal movements of the countermeasure target can change in real time according to the first countermeasure command.

[0021] S120. Install a first control device and a first signal acquisition device on the target object, and install a second control device on the traffic light; the vehicle under test is equipped with an intelligent driving system and a second signal acquisition device is installed.

[0022] The first control device is used to control the movement of the target object, such as a vehicle drive system, motor, hydraulic cylinder, etc. This embodiment does not limit the specific structure of the first control device. The second control device (e.g., a control chip) is used to control the display of traffic light colors, icons, etc.

[0023] Figure 2 This is a schematic diagram illustrating the countermeasures against the target object, the vehicle under test, and traffic lights provided in an embodiment of this application. Taking a motor vehicle as an example, the first signal acquisition device includes the following modules: The power supply is installed in the trunk of the vehicle. A driving robot is installed in the driver's seat; an inertial navigation differential positioning module and a real-time communication module are installed at the geometric center of the vehicle. The driving robot, the inertial navigation differential positioning module, and the real-time communication module are all connected to the power supply via wiring harnesses.

[0024] The second signal acquisition device is installed inside the vehicle under test and includes the following modules: The power supply is installed in the trunk of the vehicle under test. A front-facing vision sensor is installed in the center of the inner wall of the windshield; a rear-facing vision sensor is installed in the center of the inner wall of the rear windshield; a left-facing vision sensor is installed above the center of the left front wheel on the upper fender; a right-facing vision sensor is installed above the center of the right front wheel on the upper fender; when the icon being tested (e.g., the intelligent driving status icon) is displayed on the instrument panel, an instrument panel vision sensor is installed in front of the instrument panel; when the icon being tested is displayed on the central control screen, a central control vision sensor is installed in front of the central control screen. Figure 2 The instrument panel / central control vision sensor represents the aforementioned two display modes; a driver's vision sensor is installed at the center armrest, a sound sensor is installed at the driver's speaker, a sensing signal extraction module is connected to the CAN or Ethernet channel of the On-Board Diagnostics (OBD) interface, an inertial navigation differential positioning module, a real-time communication module (some inertial navigation differential positioning modules include a real-time communication module, so a separate real-time communication module is not required), and a real-time computing module (some industrial control computers include a real-time computing module, so a separate real-time computing module is not required) are installed at the vehicle's geometric center; an acoustic-optical analysis module and an industrial control computer (including a data synchronization acquisition module and a data storage module) are installed in the trunk of the vehicle under test; and a test operator's operation module is installed in the passenger seat.

[0025] The visual sensor, sound sensor, sensing signal extraction module, inertial navigation differential positioning module, real-time communication module, real-time computing module, acoustic-optical analysis module, industrial control computer, and test personnel operation module are all connected to the power supply via wiring harnesses. The real-time communication module of the vehicle under test is wirelessly connected to the real-time communication module of the countermeasure target.

[0026] The above signal acquisition devices are not only applicable to motor vehicles, but also to other types of traffic targets such as non-motor vehicles, pedestrians, and obstacles, enabling real-time signal interaction between the tested vehicle and various adversarial targets, and providing a basis for judging adversarial actions of the adversarial targets.

[0027] A second control device is installed inside the traffic light to receive a second countermeasure command issued by the test personnel operating the module.

[0028] Optionally, in addition to the first and second signal acquisition devices mentioned above, tools such as measuring tapes, visibility meters, lux meters, and weather monitors should be used during the test to record the attribute data of the target and weather environment data. The first / second signal acquisition devices should be used to record the state of the vehicle under test, the perception state of the vehicle under test, the state of the target, and the relative state data between the vehicle under test and the target.

[0029] S130: Send countermeasure commands matching the state of the vehicle under test to the first control device and the second control device respectively, so as to drive the countermeasure target and traffic lights to generate countermeasure actions against the vehicle under test.

[0030] When performing countermeasure control on traffic lights, the following steps are taken: First, the type of traffic light is obtained, including: motor vehicle traffic lights, pedestrian crossing traffic lights, direction indicator traffic lights, lane indicator traffic lights, and flashing warning traffic lights. Based on the distance between the vehicle being tested and the traffic light, as well as the status of the vehicle being tested, a second countermeasure command is generated. The second countermeasure command is then sent to the second control device to drive changes in the display of the traffic light.

[0031] Specifically, the traffic light countermeasure state machine issues a second countermeasure command when the following trigger conditions are met, automatically changing the display time, color, and indicator icons of the traffic lights to verify the real-time performance of the tested vehicle. The trigger settings for the traffic light countermeasure state machine are as follows: 1) When the traffic lights in the scene are for motor vehicles: When the vehicle being tested approaches the traffic light to a distance of 50m, the traffic light changes color from green to yellow, and then to red after 3 seconds; when the vehicle being tested stops 5 seconds before the stop line, the traffic light changes color from red to green.

[0032] 2) When the traffic lights in the scene are pedestrian crossing lights: When the vehicle being tested approaches the pedestrian crossing to a distance of 50m, the pedestrian crossing signal light changes from red to green; when the vehicle being tested stops in front of the pedestrian crossing for 5 seconds, the pedestrian crossing signal light changes from red to green.

[0033] 3) When the traffic lights in the scene are directional indicator lights: When the vehicle under test approaches the directional indicator light to a distance of 50m, the display color of the directional indicator light in the direction of travel of the vehicle under test changes from green arrow to yellow arrow, and then to red arrow after 3 seconds; when the vehicle under test is stationary 5 seconds before the stop line, the display color of the directional indicator light changes from red to green.

[0034] 4) When the traffic lights in the scene are lane indicator lights: When the test vehicle approaches the lane indicator light to a distance of 50m and the lane indicator light icon is a green arrow, the lane indicator light icon changes to a red cross; when the test vehicle approaches the lane indicator light to a distance of 50m and the lane indicator light icon is a red cross, the lane indicator light icon changes to a green arrow.

[0035] 5) When the traffic lights in the scene are flashing warning lights: When the tested vehicle approaches the flashing warning light to a distance of 50m, the flashing warning light's display icon changes from no pattern to a flashing yellow light.

[0036] When performing countermeasure control on a target object, firstly, the type of the target object is acquired, including: motor vehicles, non-motorized vehicles, pedestrians, animals, and obstacles; the states of the target object and the vehicle under test are acquired through a first signal acquisition device and a second signal acquisition device; the industrial control computer determines whether the target object's countermeasure state machine has reached the trigger condition based on the states of the target object and the vehicle under test; if the trigger condition is reached, the industrial control computer generates a first countermeasure command based on the type of the target object and the state of the vehicle under test; the first countermeasure command is sent to a first control device to drive the movement changes of the target object.

[0037] Specifically, when the triggering condition is met, the control state machine of the adversarial target issues the first adversarial command, completes the active interactive adversarial trajectory setting of the adversarial target, and verifies the driving performance of the tested vehicle under adversarial conditions. The parameters involved in the triggering condition (i.e., the parameters that need to be collected and calculated) include at least: relative distance (…). ), the speed of the vehicle being measured ( ), speed against the target ( ), the steering angle of the vehicle being tested ( Following distance () ), pre-collision time ( ) etc. Among them, following distance ( ), pre-collision time ( The calculation formula is: ; ; The first countermeasure command is used to automatically update the motion parameters of the countermeasure target, including at least: motion time, motion distance, motion speed, motion acceleration, motion angle, and motion angular acceleration.

[0038] The following test scenarios illustrate in detail the triggering process of the adversarial target control state machine: 1) If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is in a state of attempting to cut out (e.g., the steering angle is greater than the threshold), generate a cut-in command for the target object to the left rear or left front to block the lane change space of the tested vehicle.

[0039] 2) If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is attempting to overtake (e.g., the steering angle is greater than the threshold and it has overtaken the vehicle in front), a cut-out command is generated for the target object in front (i.e. the vehicle being overtaken), blocking the path of the tested vehicle.

[0040] 3) If the target object is a motor vehicle or a non-motor vehicle, the tested vehicle has no intention to change lanes, and the relative distance to the target object in front is less than the safety threshold, or the pre-collision time is less than the safety threshold, or the following distance is less than the safety threshold, a braking command is generated for the target object in front to test the distance perception capability of the tested vehicle.

[0041] 4) If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is in a state of attempting to cut in (e.g., the steering angle is greater than the threshold and it is facing the position behind the vehicle in front), an acceleration command is generated for the target object (i.e. the vehicle in front) to block the merging gap of the tested vehicle.

[0042] 5) If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle enters the intersection and the relative distance to the target object is less than the safety threshold, or the pre-collision time is less than the safety threshold, a cross-intersection instruction is generated for the target object to test the test vehicle's intersection passage capability.

[0043] 6) If the target object is a pedestrian or an animal, and the relative distance between the tested vehicle and the target object is less than the safety threshold or the pre-collision time is less than the safety threshold, generate a road crossing or diagonal crossing action command for the target object to test the test vehicle's response capability to pedestrians or animals.

[0044] 7) If the target object is an aerial obstacle (such as a foam block on a tree beside the road), and the relative distance between the tested vehicle and the aerial obstacle is less than the safety threshold or the pre-collision time is less than the safety threshold, a fall command is generated for the aerial obstacle. The foam block contains a vibration motor, which vibrates at a high frequency after receiving the fall command to make the foam block fall without damaging the vehicle, thus testing the test vehicle's response to aerial obstacles.

[0045] 8) If the target object is a ground obstacle (such as a foam block on the road), and the relative distance between the tested vehicle and the ground obstacle is less than the safety threshold or the pre-collision time is less than the safety threshold, a movement command is generated for the ground obstacle. The bottom of the foam block is equipped with wheels and a motor. After receiving the movement command, the motor controls the wheels to rotate, so as to move the foam block and test the test vehicle's response to the ground obstacle.

[0046] S140. During the confrontation action, the state data of the tested vehicle is collected in real time through the second signal acquisition device; the state data of the confrontation target is collected in real time through the first signal acquisition device.

[0047] S150. Based on the state data of the vehicle under test, the state data of the target object, the weather environment data, and the road environment data, test whether the vehicle under test passes the test.

[0048] Optionally, by combining weather environment data, road environment data, audible and visual warning data, instrument panel / central control icon data, vehicle status data, vehicle perception status, target status data, and relative status data between the vehicle and the target, the test vehicle can be judged to pass the test. Vehicles equipped with intelligent driving systems should comply with the safety requirements of road traffic signs, markings, traffic lights, and other traffic facilities during operation, and should not pose unreasonable safety risks to other road users.

[0049] Optional, the specific determination criteria are as follows: 1) The tested vehicle did not collide with any opposing target, meaning the relative lateral distance was always greater than zero; 2) The tested vehicle should meet the indication requirements when it arrives at the coordinate position of road signs, road markings, and traffic lights; 3) The tested vehicle did not straddle the lane lines; 4) The tested vehicle did not perform emergency braking or emergency steering measures without cause; 5) The vehicle under test shall properly control its turn signals, hazard warning signals, brake lights, and other lighting and light signal devices; 6) The vehicle being tested must use its horn correctly; 7) The vehicles being tested shall follow the principles of right-of-way, yielding to straight-going vehicles when turning, yielding to left-turning vehicles when turning right, and yielding to main roads when on side roads; 8) The tested vehicle was not parked in no-parking areas such as pedestrian crossings or grid areas; 9) If the intelligent driving function of the vehicle under test is disengaged during the test, the driver should be alerted by sound and light signals.

[0050] This application avoids the shortcomings of traditional real-vehicle testing schemes, such as mismatch between testing tools and key points of road traffic regulations, asynchronous data recording, missing perception signal data, and evaluation indicators that only include whether a collision occurs. It meets the special needs of traffic regulation compliance testing and proposes a clear method for real-vehicle testing and evaluation related to traffic regulations.

[0051] The following detailed description of the interactive confrontation testing method for the intelligent driving system provided in this application is illustrated through a specific implementation method.

[0052] First, complete the design and layout of the test scenario, control equipment, and signal acquisition device.

[0053] Then, the deployed equipment, the vehicle under test, and the target object are debugged to ensure the normal conduct of the test. For example, ensure that the driver is in the driver's seat and can control the activation and deactivation of the intelligent driving function and handle emergencies during the test; the tester is in the passenger seat and holds a tester operation module, which can be used to send control signals to the target object and traffic lights to pre-test whether the target object and traffic lights can respond normally. At the same time, if the state machine fails during the formal test, the target object and traffic lights can be manually controlled to ensure the smooth conduct of the test.

[0054] Next, based on the test scenario, dynamically design and execute the following test steps: 1) Testers input calibration commands into the tester operation module to complete the distance calibration of the front, rear, left, and right vision sensors, enabling them to acquire images of key elements such as adversarial targets, traffic lights, and road signs and markings, and to measure the distance between the wheels and lane lines; calibration of the intelligent driving icon of the instrument / central control vision sensor to enable it to identify the activation or deactivation of intelligent driving functions; and calibration of the alarm tone of the sound sensor to enable it to recognize various alarm prompts in the vehicle.

[0055] 2) The tester inputs the coordinate system establishment command in the tester operation module to establish the local coordinate system with the starting point of the movement of the vehicle under test as the origin, the coordinate system of the vehicle under test with the vehicle under test as the origin, and the coordinate system of the target object with the target object as the origin, so that coordinate transformation can be realized between the three.

[0056] 3) Testers use the tester operation module to collect the coordinates of traffic markings, traffic signs, and traffic lights in the local coordinate system; 4) The tester inputs the data extraction start command into the tester operation module, the driver activates the intelligent driving function of the vehicle under test and drives into the test area to start the test. During the test, the operation of the intelligent driving vehicle should not be interfered with except in an emergency.

[0057] 5) When the triggering condition is met, the traffic signal countermeasure state machine issues a second countermeasure command to complete the automatic changes of the traffic signal display time, display color, indicator icon, etc.

[0058] 6) When the aforementioned triggering conditions are met, the anti-target control state machine issues the first anti-target command to complete the active interactive anti-target trajectory setting.

[0059] To facilitate real-time data acquisition, synchronization, and computation during testing, this application also provides an interactive adversarial testing system for intelligent driving systems. (See attached document.) Figure 3 It includes a power supply unit, a signal extraction unit, a data analysis unit, and a synchronization control and storage unit.

[0060] The power supply unit includes the power source inside the vehicle under test and the power source inside the countermeasure target.

[0061] The signal extraction unit includes a visual sensor, a sound sensor, a perception signal extraction module, an inertial navigation differential positioning module, and a real-time communication module configured on the vehicle under test; and an inertial navigation differential positioning module and a real-time communication module configured on the adversarial target. The two real-time communication modules communicate with each other to achieve time alignment and interaction. Through multi-sensor integration, raw audio and video signals, vehicle status data, adversarial target status data, and the perception status of the vehicle under test can be acquired, enabling interactive real-time communication between the vehicle under test, the adversarial target, and traffic lights. Specific data collected is shown in the table below:

[0062] In the table above, weather conditions, vehicle attributes, target object attributes, traffic sign status, and traffic marking status are static data. These need to be kept constant throughout the test, recorded only once before the test begins, and stored in the industrial control computer on the vehicle under test. Traffic light status, vehicle status, vehicle perception status, target object status, relative status between the vehicle under test and the target object, and raw signals are dynamic data. Data at each sampling time will be recorded based on the test start time and stored in the industrial control computer. The traffic light status, vehicle status, target object status, and relative status between the vehicle under test and the target object can be communicated via a real-time communication module to automatically output first / second adversarial commands through a state machine, simulating complex test scenarios with adversarial characteristics.

[0063] The data analysis unit includes an acoustic-optical analysis module and a real-time computing module. The acoustic-optical analysis module is used to parse the raw audio and video signals to obtain readable data. The real-time computing module is used to parse the test vehicle status data, the countermeasure target status data, and the test vehicle's perception signals to obtain readable data. Then, all the parsed readable data (including the readable data parsed by the acoustic-optical analysis module) is interpolated into archived data with a unified sampling frequency, including weather environment data, road environment data, acoustic-optical warning data, instrument / central control icon data, test vehicle status data, test vehicle perception status, countermeasure target status data, and relative status data between the test vehicle and the countermeasure target.

[0064] The synchronous control and storage unit includes a driving robot (suitable for motor vehicles / non-motor vehicles, responding to the first countermeasure command to cut in, cut out, brake, accelerate, and cross intersections) configured on the adversarial target, an industrial control computer and test personnel operation module configured on the vehicle under test, and a second control device configured within the traffic light. The synchronous control and storage unit is responsible for storing all test data and also for issuing calibration commands, coordinate system establishment commands, data extraction start commands, signal extraction end commands, control signals for the adversarial target, and control signals for the traffic light, enabling synchronous signal extraction, real-time target motion control, and real-time traffic light control during the test.

[0065] 7) When the vehicle under test leaves the test area, comes to an unexpected stop, or the intelligent driving function is disengaged, the tester inputs a signal to extract the end command in the tester's operation module and ends the test.

[0066] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this application can be achieved, and this is not limited herein.

[0067] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for interactive adversarial testing of an intelligent driving system, characterized in that, include: Arrange a test scenario containing adversarial targets and traffic lights according to the test requirements; A first control device and a first signal acquisition device are installed on the target object, and a second control device is installed on the traffic light; the vehicle under test is equipped with an intelligent driving system and a second signal acquisition device is installed. The first control device and the second control device respectively send countermeasure commands that match the state of the vehicle under test, so as to drive the countermeasure target and the traffic lights to generate countermeasure actions against the vehicle under test. During the confrontation, the status data of the tested vehicle is collected in real time by the second signal acquisition device; the status data of the confrontation target is collected in real time by the first signal acquisition device. Based on the status data of the vehicle under test, the status data of the target object, the weather environment data, and the road environment data, the test determines whether the vehicle under test passes the test.

2. The method according to claim 1, characterized in that, The test scenario, which includes adversarial targets and traffic lights, is set up according to the test requirements, including: The test site is set up according to the test requirements, and the test site has the necessary conditions for the normal activation of the intelligent driving system of the vehicle under test. Traffic elements, including traffic signs, traffic markings, traffic lights, and combat targets, are arranged on the test site.

3. The method according to claim 2, characterized in that, The second signal acquisition device includes: a power supply, a visual sensor, a sound sensor, a sensing signal extraction module, an inertial navigation differential positioning module, a real-time communication module, a real-time computing module, an acoustic-optical analysis module, an industrial control computer, and a test personnel operation module connected to the power supply.

4. The method according to claim 3, characterized in that, The first signal acquisition device includes: a power supply, a driving robot connected to the power supply, an inertial navigation differential positioning module, and a real-time communication module; The real-time communication module of the vehicle under test is wirelessly connected to the real-time communication module of the target.

5. The method according to claim 4, characterized in that, Sending countermeasure commands matching the state of the tested vehicle to the first control device and the second control device respectively, including: The types of traffic lights include: motor vehicle traffic lights, pedestrian crossing traffic lights, directional traffic lights, lane indicator traffic lights, and flashing warning traffic lights; Based on the distance between the tested vehicle and the traffic light, and the state of the tested vehicle, a second countermeasure command is generated. The second countermeasure command is sent to the second control device to drive the display change of the traffic signal light.

6. The method according to claim 5, characterized in that, Sending countermeasure commands matching the state of the tested vehicle to the first control device and the second control device respectively, including: The types of targets to be identified include: motor vehicles, non-motorized vehicles, pedestrians, animals, and obstacles. The status of the target object and the tested vehicle is collected through the first signal acquisition device and the second signal acquisition device. Based on the states of the target object and the tested vehicle, determine whether the target object's state machine has reached the trigger condition; If the triggering conditions are met, a first countermeasure command is generated based on the type of the countermeasure target and the state of the tested vehicle. The first countermeasure command is sent to the first control device to drive the movement changes of the countermeasure target.

7. The method according to claim 6, characterized in that, If the triggering conditions are met, a first countermeasure command is generated based on the type of the countermeasure target and the state of the tested vehicle, including: If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is in a state of attempting to cut out, a cut-in command is generated for the target object to the left rear or left front, blocking the test vehicle's lane change space. If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is attempting to overtake, a cut-out command is generated for the target object ahead, blocking the path of the tested vehicle. If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle has no intention to change lanes, and the relative distance to the target object is less than the safety threshold, or the pre-collision time is less than the safety threshold, or the following distance is less than the safety threshold, a braking command is generated for the target object to test the distance perception capability of the tested vehicle. If the target object is a motor vehicle or a non-motor vehicle, and the tested vehicle is in a state of attempting to cut in, an acceleration command is generated for the target object to block the merging gap of the tested vehicle. If the target object is a motor vehicle or a non-motor vehicle, and the vehicle under test enters the intersection and the relative distance to the target object is less than the safety threshold, or the pre-collision time is less than the safety threshold, a cross-intersection instruction is generated for the target object to test the vehicle's intersection clearance capability.

8. The method according to claim 7, characterized in that, If the triggering conditions are met, a first countermeasure command is generated based on the type of the countermeasure target and the state of the tested vehicle, including: If the target object is a pedestrian or animal, and the relative distance between the tested vehicle and the target object is less than the safety threshold or the pre-collision time is less than the safety threshold, a road crossing or diagonal crossing action command is generated for the target object to test the test vehicle's response capability to pedestrians or animals.

9. The method according to claim 8, characterized in that, If the triggering conditions are met, a first countermeasure command is generated based on the type of the countermeasure target and the state of the tested vehicle, including: If the target object is an aerial obstacle, and the relative distance between the tested vehicle and the aerial obstacle is less than the safety threshold or the pre-collision time is less than the safety threshold, a fall command is generated for the aerial obstacle to test the test vehicle’s response capability to the aerial obstacle. If the target object is a ground obstacle, and the relative distance between the tested vehicle and the ground obstacle is less than a safety threshold or the pre-collision time is less than a safety threshold, a movement command is generated for the ground obstacle to test the test vehicle's response capability to the ground obstacle.

10. An interactive confrontation testing system for an intelligent driving system, characterized in that, include: The power supply unit includes the power supply inside the vehicle under test and the power supply inside the countermeasure target. The signal extraction unit includes a visual sensor, a sound sensor, a vehicle perception signal extraction module, an inertial navigation differential positioning module, and a real-time communication module configured on the vehicle under test; and an inertial navigation differential positioning module and a real-time communication module configured on the adversarial target. The data analysis unit includes an acoustic-optical analysis module and a real-time computing module; The synchronous control and storage unit includes a driving robot configured on the adversarial target, an industrial control computer and a test personnel operation module configured on the vehicle under test, and a second control device configured in a traffic light. The intelligent driving system interaction confrontation test system is used to execute the intelligent driving system interaction confrontation test method according to any one of claims 1-9.

Images