Autonomous driving test system and test method

Abstract

The present invention relates to an autonomous driving software testing system and method, belonging to the field of autonomous driving technology. The system comprises: a simulation unit for constructing a virtual operating condition of an Ego vehicle based on a virtual time-domain (VTD); an autonomous driving control unit configured with the autonomous driving software under test for performing an Ego vehicle operating test based on the virtual operating condition and adjusting the virtual operating condition in real time; a test unit for coupling driver intervention information during the Ego vehicle operating test; and middleware for establishing a communication connection between the simulation unit, the autonomous driving control unit, and the test unit. The present invention addresses the problem of achieving an automatic and closed-loop testing and verification solution for in-the-loop simulation testing and functional verification of autonomous driving software.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving, and in particular to an autonomous driving test system and an autonomous driving test method. Background Art

[0002] As intelligent systems become increasingly sophisticated, fully autonomous driving is becoming increasingly possible. With the vigorous development of a new round of technological revolution and industrial transformation, the integration of automobiles with fields such as artificial intelligence and communications is accelerating, driving the development of intelligent vehicles. While some autonomous vehicles currently exist, incidents of personal injury and property damage caused by them are common. According to research by the RAND Corporation, autonomous vehicles require at least 11 billion miles (17.7 billion kilometers) of mileage testing to prove that autonomous driving systems are more reliable than human drivers. While real-world testing can effectively demonstrate the randomness, complexity, and typicality of tests, it is characterized by long lead times, high investment costs, and low development efficiency. Therefore, to more efficiently test and verify the safety of autonomous driving systems, software-in-the-loop (SIL) simulation testing is necessary to uncover unknown vulnerabilities, improve system safety, and reduce the accident rate of autonomous vehicles to zero. Currently, there is no automated, closed-loop testing and verification solution for SIL simulation testing and functional verification. To address this issue, a new approach to autonomous driving testing is needed. Summary of the Invention

[0003] The purpose of the present invention is to provide an autonomous driving test system and test method to solve the problem of in-loop simulation testing and functional verification of autonomous driving software, where there is no automatic and closed-loop testing and verification solution.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A first aspect of the present invention provides an autonomous driving test system, which includes: a simulation unit, which constructs a virtual operating condition of an Ego vehicle based on VTD; an autonomous driving control unit, which is configured with a tested autonomous driving software, and is used to perform an Ego vehicle operation test based on the virtual operating condition and adjust the virtual operating condition in real time; a test unit, which is used to couple driver intervention information during the Ego vehicle operation test; and middleware, which serves as a medium to establish a communication connection between the simulation unit, the autonomous driving control unit and the test unit.

[0006] Optionally, the virtual operating conditions of the Ego vehicle include: dynamic scene information and static information; wherein the dynamic scene information includes at least: Ego vehicle information and vehicle speed information; the static information includes at least: lane number information, lane width information and obstacle information.

[0007] Optionally, the dynamic scene information is constructed based on a ScenarioEditor tool; and the static information is constructed based on a ROD tool.

[0008] Optionally, the autonomous driving control unit is configured to: judge the real-time operating status of the Ego vehicle based on the virtual operating conditions, and execute autonomous driving algorithm calculations based on the real-time judgment results to obtain autonomous driving control instructions; and feed back the autonomous driving control instructions to the simulation unit through the middleware, so that the simulation unit can simulate the Ego vehicle to execute the autonomous driving control instructions in the virtual operating conditions it constructs.

[0009] Optionally, the test unit is configured to: preset a driving script and collect the driver's operation trigger instructions; and simultaneously feed back the operation trigger instructions to the simulation unit and the automatic driving control unit through the middleware, so that the simulation unit and the automatic driving control unit can couple the driver's operation instructions when controlling the vehicle.

[0010] Optionally, the middleware and the simulation unit transmit data based on the TCP protocol; the middleware and the automatic driving control unit transmit data through the ZMQ protocol; the middleware and the test unit transmit data through the UDP protocol.

[0011] Optionally, the test unit is also configured with a test script for executing the initial test state, variable injection and pass judgment based on UserCase conversion; the system also includes: a data playback unit for executing offline data playback, identifying and outputting error information when the test script reports an error.

[0012] The second solution of the present invention provides an autonomous driving test method, which is implemented based on the above-mentioned autonomous driving test system. The autonomous driving test method includes: constructing a virtual operating condition of an Ego vehicle based on VTD; conducting an Ego vehicle operation test in the virtual operating condition based on the tested autonomous driving software, and adjusting the virtual operating condition in real time; coupling driver intervention information during the Ego vehicle operation test; and monitoring the Ego vehicle simulated operation process in real time to perform in-loop function verification.

[0013] Optionally, the real-time monitoring of the Ego vehicle simulation operation process and the in-loop function verification include: monitoring the running status of the test script, and when the test script reports an error, performing offline data playback, identifying and outputting error information.

[0014] Beneficial effects of the present invention: The present invention proposes an autonomous driving test system and test method. The present invention constructs an Ego vehicle (test vehicle) and the operating environment of the Ego vehicle based on VTD, and then performs simulated autonomous driving of the Ego vehicle in the simulated working conditions based on the tested autonomous driving software, and couples the driver's operation information in real time to model the actual vehicle operating conditions, and performs in-loop simulation and functional verification of the tested autonomous driving software. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A system structure diagram of an autonomous driving test system provided in one embodiment of the present invention;

[0016] Figure 2 A flowchart of the steps of an autonomous driving testing method provided in one embodiment of the present invention;

[0017] Figure 3 A flowchart of the steps for implementing an autonomous driving test method provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0019] As intelligent driving becomes increasingly sophisticated, fully autonomous driving is becoming increasingly possible. With the vigorous development of a new round of technological revolution and industrial transformation, the integration of automobiles with fields such as artificial intelligence and communications is accelerating, driving the development of intelligent vehicles. Although some autonomous vehicles currently exist, incidents of human and property damage caused by autonomous vehicles occur frequently. According to research by the Rand Corporation, autonomous vehicles require at least 11 billion miles (17.7 billion kilometers) of mileage testing to prove that autonomous driving systems are more reliable than human drivers. While real-world testing can effectively demonstrate the randomness, complexity, and typicality of tests, it is characterized by long time periods, high investment costs, and low development efficiency.

[0020] Therefore, in order to more efficiently test and verify the safety of the autonomous driving system, it is necessary to conduct in-the-loop simulation testing of the autonomous driving software to discover unknown defects, improve the safety of the autonomous driving system, and reduce the accident rate of autonomous vehicles to zero.

[0021] At present, there is no automatic and closed-loop testing and verification solution for in-loop simulation testing and functional verification of autonomous driving software. Based on this defect, the present invention proposes an autonomous driving software testing system and testing method. The present invention constructs an Ego vehicle (test vehicle) and the operating environment of the Ego vehicle based on VTD, and then performs Ego vehicle simulated autonomous driving in the simulated working conditions based on the tested autonomous driving software, and couples the driver's operating information in real time to model the actual vehicle operating conditions to perform in-loop simulation and functional verification of the tested autonomous driving software.

[0022] Figure 1 This is a system structure diagram of an autonomous driving software testing system provided by one embodiment of the present invention. Figure 1 According to a first aspect of the present invention, there is provided an autonomous driving software testing system, which includes: a simulation unit, which constructs a virtual operating condition of an Ego vehicle based on VTD; an autonomous driving control unit, which is configured with a tested autonomous driving software, and is used to perform an Ego vehicle operation test based on the virtual operating condition and adjust the virtual operating condition in real time; a test unit, which is used to couple driver intervention information during the Ego vehicle test; and middleware, which serves as a medium to establish a communication connection between the simulation unit, the autonomous driving control unit and the test unit.

[0023] In an embodiment of the present invention, VTD is a toolkit for creating, configuring, simulating and evaluating virtual environments, involving road and rail traffic simulation. It is used in the development of advanced driver assistance systems and autonomous driving systems, and is also used as a training simulator. It covers the simulation process from the creation of a three-dimensional virtual world to complex traffic scenes, including simple or physical-level sensor simulation. It is applied to the software-in-the-loop, driver-in-the-loop, vehicle-in-the-loop and hardware-in-the-loop stages, and can be collaboratively simulated with third-party or customer-owned applications. Its open and modular design concept makes it easy to use interfaces and integrate. The present invention makes full use of the advantages of VTD to construct a test condition for the autonomous driving software under test, so as to facilitate autonomous driving simulation in the constructed virtual scene.

[0024] Preferably, the virtual operating conditions of the Ego vehicle include: dynamic scene information and static information; wherein, the dynamic scene information includes at least: Ego vehicle information and vehicle speed information; the static information includes at least: lane number information, lane width information, and obstacle information.

[0025] Specifically, based on the provided OpenDrive (road standard), OpenCRG, and OpenScenario (intelligent driving and ADAS standard) open data formats, simulations of roads (geometric topology, pavement, etc.), traffic facilities (traffic lights, traffic lights), temporary roads or traffic facilities (cones, etc.), targets (cars, trucks, motorcycles, etc.), and environments (rainy days, sunny days, etc.) are realized, thereby achieving SIL (Software in the loop) simulation and functional verification.

[0026] Preferably, the dynamic scene information is constructed based on a Scenario Editor tool; and the static information is constructed based on a ROD tool.

[0027] In one possible implementation, road specifications, test case overviews, and other information are used as input to determine the functional and logical scenarios within the constructed scenario, including scenario type, road type, number of lanes, lane count, state changes of the Ego and target vehicles, and boundary values ​​of key parameters. Based on the analyzed input parameter information, a static road is constructed using the ROD (RoadDesigner) tool in the VTD software, including information on the number and width of lanes, to generate a logical road database (xodr) (a data format under OpenDrive) and a graphical road database (osgb) (a data format under the OpenSceneGraph framework). A dynamic scenario is constructed using the ScenarioEditor in the VTD software, including information on the speeds of the Ego and target vehicles, to generate a scenario file (xml). During simulation, TaskControl (project management) centrally dispatches modules such as Traffic, IG (interactive generator), ScenarioEditor, and ModuleManager, transmitting data on vehicle status, sensors, positioning, perception fusion, and high-precision maps to the middleware via TCP / RDB.

[0028] Preferably, the automatic driving control unit is configured to: judge the real-time operating status of the test vehicle based on the virtual operating conditions, and execute automatic driving algorithm calculations based on the real-time judgment results to obtain automatic driving control instructions; and feed back the automatic driving control instructions to the simulation unit through the middleware, so that the simulation unit can simulate the test vehicle to execute the automatic driving control instructions in the virtual operating conditions constructed by it.

[0029] Specifically, the autonomous driving control unit obtains the vehicle status, real sensor data, perception fusion, high-precision map and other data from the VTD software and sends it to the middleware via TCP. The middleware feeds back real-time information such as gear, acceleration, and steering wheel angle to the Ego vehicle controlled by VTD via TCP / RDB. At the same time, the middleware obtains the vehicle status, real sensor data, perception fusion, high-precision map and other data from the VTD software and sends it to the autonomous driving software algorithm via the ZMQ protocol / protobuf. The autonomous driving algorithm sends the gear information, acceleration information, and steering wheel angle data to the middleware via the ZMQ protocol / protobuf.

[0030] In one possible implementation, the middleware acts as a bridge between the VTD software and the autonomous driving algorithm. Data from the VTD software, including vehicle status (speed, acceleration, etc.), sensor data (lane markings, etc.), positioning data (lane position, etc.), sensor fusion data (object category, etc.), and high-precision maps (lane curvature 100 meters ahead, etc.), is processed by the middleware and passed as algorithm input to the autonomous driving algorithm via ZMQ / protobuf. Simultaneously, the algorithm outputs acceleration, steering wheel angle, and gear position data, which are then passed to the VTD software to control the Ego's driving behavior, such as acceleration, deceleration, and lane changes. Furthermore, the test case Python script communicates with the middleware via UDP, passing driver operations and SCP commands to the autonomous driving algorithm and VTD software, enabling functions such as autonomous driving function activation and fault variable injection. It also simulates the acceleration, deceleration, and lane changes of dynamic traffic target vehicles during simulation, enabling simulation testing of the interactive game between the Ego and traffic participants.

[0031] Preferably, the test unit is configured to: preset a driving script and collect the driver's operation trigger instructions; and simultaneously feed back the trigger operation instructions to the simulation unit and the automatic driving control unit through the middleware, so that the simulation unit and the automatic driving control unit can couple the driver's operation instructions when controlling the vehicle.

[0032] Specifically, the processed vehicle status, sensor, positioning, high-precision map and other data transmitted from the middleware are processed by the environment construction and prediction modules. The Ego vehicle then decides whether to center the vehicle, change lanes to the target lane and the vehicle's driving speed based on its vehicle status, positioning information, lane attributes, lane line attributes, surrounding target vehicle attributes, navigation information, driver intentions, etc. When the Ego vehicle decides to change lanes, the lateral planning module will plan a safe, comfortable and efficient curve to control the vehicle's lane changing behavior. At the same time, the longitudinal planning module will plan the vehicle's acceleration based on the target vehicle. The lateral and longitudinal control modules will output the precise steering wheel angle and target acceleration in real time based on the planned curve and acceleration size to achieve precise lateral and longitudinal behavior of the target vehicle.

[0033] Preferably, the middleware and the simulation unit transmit data based on the TCP protocol; the middleware and the automatic driving control unit transmit data through the ZMQ protocol; and the middleware and the test unit transmit data through the UDP protocol.

[0034] Specifically, based on the analyzed input parameters, the Road Designer (ROD) tool in the VTD software is used to build a static road, including the number and width of lanes, generating a logical road database (xodr) and a graphical road database (osgb). Dynamic scenarios are built using the Scenario Editor in the VTD software, including the Ego vehicle and target vehicle speeds, generating scenario files (xml). During simulation, TaskControl centrally schedules modules such as Traffic, IG, Scenario Editor, and ModuleManager, transmitting vehicle status, sensor data, positioning, perception fusion, and high-precision map data to the middleware via TCP / RDB. TCP is designed to adapt to a layered protocol hierarchy that supports multi-network applications. TCP provides reliable communication services between paired processes in a host computer connected to different but interconnected computer communication networks. TCP assumes that it can obtain simple, potentially unreliable datagram services from lower-level protocols. In principle, TCP should be able to operate on a variety of communication systems, from hardwired connections to packet-switched or circuit-switched networks.

[0035] Furthermore, the middleware passes data as algorithm input to the autonomous driving algorithm through ZMQ / protobuf. At the same time, it can pass the acceleration, steering wheel angle, and gear position data output by the autonomous driving algorithm to the VTD software, thereby controlling the Ego car's acceleration, deceleration, lane changing and other driving behaviors. ZMQ is a series of interfaces similar to Socket (a protocol-independent network programming interface). The difference between it and Socket is that ordinary sockets are end-to-end (1:1 relationship), while ZMQ can have an N:M relationship. People are more familiar with BSD sockets as point-to-point connections. Point-to-point connections require explicit connection establishment, connection destruction, protocol selection (TCP / UDP) and error handling, etc. ZMQ shields these details, making your network programming simpler. ZMQ is used for communication between nodes, and a node can be a host or a process.

[0036] Furthermore, the test case Python script can communicate with the middleware via UDP, passing driver operations and SCP commands to the autonomous driving algorithm and VTD software. This enables functions such as autonomous driving function activation and fault variable injection. It can also simulate the acceleration, deceleration, and lane-changing behavior of dynamic traffic target vehicles during simulation, enabling simulation testing of the interactive game between the Ego car and traffic participants. UDP provides a method for applications to send encapsulated IP packets without establishing a connection. RFC768 describes UDP.

[0037] Preferably, the test unit is also configured with: a test script, based on UserCase (use case diagram) conversion, for executing the initial test state, variable injection and pass judgment; the system also includes: a data playback unit, for executing offline data playback, identifying and outputting error information when the test script reports an error.

[0038] Specifically, the UserCase test case is converted into a test case, including the initial test state, variable injection, and pass judgment. The test case is then converted into a Python script. The Python script can communicate with the middleware via UDP. The generated Python test script is used to conduct efficient, large-scale simulation tests to verify the autonomous driving software functions. Data from the simulation test process is recorded based on the autonomous driving system framework. When the Python test script fails (reports an error), the corresponding data is found for offline data playback to discover autonomous driving software defects and achieve software-in-the-loop functional verification.

[0039] In one possible implementation, during the simulation test, data is recorded in real time through the autonomous driving software framework. When a large-scale test case fails, the recorded DAT format data is loaded into an offline playback tool to track internal algorithm variables, such as lane change decisions, target longitudinal distance, acceleration, lane type, and other variables. The values ​​of these variables are constantly changing as the simulation progresses, which enables rapid fault location and discovery, improving the efficiency and accuracy of autonomous driving algorithm defect detection.

[0040] Figure 2 This is a flowchart of the steps of the autonomous driving test method provided by one embodiment of the present invention. Figure 2 As shown, an embodiment of the present invention provides an autonomous driving test method, the method comprising:

[0041] Step S10: constructing a virtual operating condition of the Ego vehicle based on the VTD;

[0042] Step S20: Based on the tested autonomous driving software, performing an Ego vehicle operation test in the virtual operating condition, and adjusting the virtual operating condition in real time;

[0043] Step S30: During the Ego vehicle test, driver intervention information is coupled;

[0044] Step S40: Real-time monitoring of the Ego vehicle simulation operation process to perform in-loop function verification.

[0045] Preferably, the real-time monitoring of the Ego vehicle simulation operation process and the in-loop function verification include: monitoring the running status of the test script, and when the test script reports an error, performing offline data playback, identifying and outputting error information.

[0046] like Figure 3During a complete operation, for step ①, the functional and logical scenarios in the constructed scenario are determined using national road specifications and test case overviews as input, including scenario type, road type, number of lanes, lane count, state changes of the Ego vehicle and target vehicle, and boundary values ​​of key parameters. Based on the analyzed input parameter information, the ROD (Road Designer) tool in the VTD software is used to construct static roads, including the number of lanes and lane width, and generate the logical road database xodr and the graphical road database osgb. The ScenarioEditor in the VTD software is used to construct dynamic scenarios, including the speed of the Ego vehicle and target vehicle, and generate the scenario file xml. During simulation, TaskControl uniformly dispatches modules such as Traffic, IG, ScenarioEditor, and ModuleManager, and transmits data on vehicle status, sensors, positioning, perception fusion, and high-precision maps to the middleware via TCP / RDB.

[0047] For step ②, the middleware acts as a bridge between the VTD software and the autonomous driving algorithm. It processes data from the VTD software, including vehicle status (speed, acceleration, etc.), sensor data (lane markings, etc.), positioning data (lane position, etc.), sensor fusion data (object category, etc.), and high-precision maps (lane curvature 100 meters ahead, etc.). The data is then passed to the autonomous driving algorithm via ZMQ / protobuf as input. Simultaneously, the algorithm outputs acceleration, steering wheel angle, and gear position data, which are then passed to the VTD software to control the Ego's driving behaviors, such as acceleration, deceleration, and lane changes. Furthermore, the test case Python script communicates with the middleware via UDP, passing driver operations and SCP commands to the autonomous driving algorithm and VTD software, enabling functions such as autonomous driving function activation and fault variable injection. It also simulates the acceleration, deceleration, and lane changes of dynamic traffic target vehicles during simulation, enabling simulation testing of the interactive game between the Ego and traffic participants.

[0048] For step ③, the processed vehicle status, sensor, positioning, high-precision map and other data transmitted from the middleware are processed by the environment construction and prediction modules. The Ego vehicle then decides whether to center the vehicle, change lanes to the target lane, and maintain the vehicle's speed based on its vehicle status, positioning information, lane attributes, lane line attributes, surrounding target vehicle attributes, navigation information, and driver intention. When the Ego vehicle decides to change lanes, the lateral planning module plans a safe, comfortable, and efficient curve to control the vehicle's lane change behavior. At the same time, the longitudinal planning module plans the vehicle's acceleration based on the target vehicle. The lateral and longitudinal control modules output the precise steering wheel angle and target acceleration in real time based on the planned curve and acceleration size to achieve accurate lateral and longitudinal behavior of the target vehicle.

[0049] For step ④, based on the ASPICE process, the autonomous driving system requirements and software requirements are used as one of the sources for test case writing. However, functional abnormality scenarios that occur during actual vehicle testing are also one of the sources for case writing. Python is used to convert test cases into test scripts that can communicate with middleware, inject input into the autonomous driving algorithm, and control the behavior of the VTD target vehicle.

[0050] For step ⑤, during the simulation test, data is recorded in real time through the autonomous driving software framework. When a large-scale test case fails, the recorded DAT format data is loaded into the offline playback tool to track the algorithm's internal variables, such as lane change decisions, target longitudinal distance, acceleration, lane type, and other variables. The values ​​of these variables change continuously as the simulation progresses. This enables rapid fault location and discovery, improving the efficiency and accuracy of autonomous driving algorithm defect detection.

[0051] An embodiment of the present invention also provides a computer-readable storage medium, which stores instructions that, when executed on a computer, enable the computer to execute the above-mentioned autonomous driving test method.

[0052] Those skilled in the art will appreciate that all or part of the steps in the methods of the aforementioned embodiments can be accomplished by instructing the relevant hardware through a program, which is stored in a storage medium and includes a number of instructions for causing a single-chip microcomputer, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0053] The above describes in detail the optional embodiments of the present invention in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept of the embodiments of the present invention, a variety of simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the scope of protection of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner unless there is any contradiction. In order to avoid unnecessary repetition, the embodiments of the present invention will no longer describe the various possible combinations separately.

[0054] In addition, the various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

[0055] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or modification made by those skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. An autonomous driving test system, characterized in that: The autonomous driving test system includes: The simulation unit builds the virtual operating conditions of the Ego vehicle based on the VTD; an autonomous driving control unit, configured with the tested autonomous driving software, for conducting an Ego vehicle operation test based on the virtual operating conditions and making real-time adjustments to the virtual operating conditions; A test unit for coupling driver intervention information during the Ego vehicle operation test; a middleware, serving as a medium to establish a communication connection between the simulation unit, the automatic driving control unit, and the test unit; The test unit is configured to: preset a driving script and collect an operation trigger instruction of the driver; and feed the operation trigger instruction back to the simulation unit and the automatic driving control unit simultaneously through the middleware, so that the simulation unit and the automatic driving control unit can couple the driver's operation instruction when performing vehicle control; The automatic driving control unit is configured to: judge the real-time operating status of the Ego vehicle based on the virtual operating condition, and execute automatic driving algorithm calculations based on the real-time judgment results to obtain automatic driving control instructions; and feed back the automatic driving control instructions to the simulation unit through the middleware, so that the simulation unit can simulate the Ego vehicle to execute the automatic driving control instructions in the virtual operating condition it constructs.

2. The autonomous driving test system according to claim 1, characterized in that: The virtual operating conditions of the Ego vehicle include: Dynamic scene information and static information; among them, The dynamic scene information includes at least: Ego vehicle information and vehicle speed information; The static information includes at least: lane number information, lane width information and obstacle information.

3. The autonomous driving test system according to claim 2, characterized in that: The dynamic scene information is built based on the ScenarioEditor tool; the static information is built based on the ROD tool.

4. The autonomous driving test system according to claim 1, wherein: The middleware and the simulation unit perform data transmission based on the TCP protocol; The middleware and the autonomous driving control unit perform data transmission via the ZMQ protocol; The middleware and the test unit perform data transmission via the UDP protocol.

5. The autonomous driving test system according to claim 1, wherein: The test unit is also configured with a test script for executing the initial test state, variable injection and pass judgment based on UserCase conversion; The system further comprises: The data playback unit is used to perform offline data playback, identify and output error information when the test script reports an error.

6. An autonomous driving testing method, characterized in that: The autonomous driving test method is implemented based on the autonomous driving test system according to any one of claims 1 to 5, and the autonomous driving test method includes: Construct the virtual operating conditions of the Ego vehicle based on VTD; Based on the tested autonomous driving software, the Ego vehicle is tested in the virtual operating conditions, and the virtual operating conditions are adjusted in real time; During the Ego vehicle operation test, driver intervention information is coupled; Monitor the Ego vehicle simulation operation process in real time and perform in-loop function verification.

7. The autonomous driving test method according to claim 6, characterized in that: The real-time monitoring of the Ego vehicle simulation operation process and the in-loop function verification include: Monitor the running status of the test script, and when the test script reports an error, perform offline data playback, identify and output error information.

8. A computer-readable storage medium having instructions stored thereon, which, when executed on a computer, enables the computer to execute the autonomous driving test method according to any one of claims 6 or 7.

Images