Autonomous driving test simulation system, method and local real-time machine

Abstract

This invention provides an autonomous driving test simulation system, method, and local real-time machine. The system includes: a simulation cloud platform that simulates the operation of a virtual master vehicle based on its state, obtains real-time sensor data and real-time V2X simulation data, and sends this data to the local real-time machine; updates the virtual master vehicle state based on the vehicle operation data; and obtains the autonomous driving test evaluation results after the test period ends. The local real-time machine converts the sensor data and V2X simulation data into the input protocol corresponding to the controller and transmits it to the controller; it also uploads the vehicle operation data to the simulation cloud platform. The controller analyzes the data using a preset autonomous driving algorithm, obtains control commands, and sends them to the local real-time machine. By using the local real-time machine to connect the simulation cloud platform and the controller, the utilization rate of hardware resources is greatly improved, the testing cost is reduced, reconfiguration is not required, and the testing efficiency is increased, all while testing the controller hardware.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to an autonomous driving test simulation system, method, and local real-time machine. Background Technology

[0002] There are currently two main approaches to simulation testing of autonomous driving algorithms: one is cloud-based large-scale concurrent simulation, where both the simulation and the algorithm are deployed in the cloud. The simulation cloud platform is responsible for simulating dynamic and static scenes, vehicle dynamics, sensor data, and V2X (vehicle-to-the-world) data based on high-precision maps. The autonomous driving algorithm generates vehicle control signals based on the simulation data simulated by the cloud platform, and in turn controls the vehicle dynamics model of the simulation cloud platform, thus forming a simulation closed loop. The other approach is stand-alone in-the-loop simulation, where the device under test is an ADAS (Advanced Driver Assistance System) or AD controller. The simulation workstation simulates dynamic and static scenes, dynamics, and sensor data, forming a simulation closed loop with the algorithm in the controller. Due to the elastic scaling mechanism of cloud computing resources, cloud simulation is suitable for large-scale concurrent testing of autonomous driving algorithms for the Model In-the-Loop (MIL) and System In-the-Loop (SIL) stages of the autonomous driving system development (V) process; while in-the-loop simulation is more suitable for the Hardware In-the-Loop (HIL) and Vehicle In-the-Loop (VIL) stages of the V process.

[0003] Cloud simulation solutions are mainly suitable for model development and code development stages, but they cannot test real controller hardware. In-loop simulation solutions can test controllers, but each controller under test requires a separate high-performance simulation workstation. When multiple ADAS / AD controllers need to be tested simultaneously, a large number of high-performance simulation workstations and real-time systems are required, resulting in high testing costs. Furthermore, the synchronization of high-precision maps and scene libraries between different simulation workstations is also a problem. Therefore, OEMs' HIL and VIL tests generally cannot perform a large number of concurrent tests like cloud simulations, resulting in relatively low testing efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an autonomous driving test simulation system, method, and local real-time machine that can test real controller hardware, reduce testing costs, and improve testing efficiency.

[0005] To achieve the above objectives, the present invention provides an autonomous driving test simulation system, comprising:

[0006] The simulation cloud platform is used to configure a virtual master vehicle and corresponding sensors in the cloud according to preset dynamic parameters; based on the virtual master vehicle state, it simulates the operation of the virtual master vehicle to obtain real-time sensor data and real-time V2X simulation data, and sends the real-time sensor data and real-time V2X simulation data to the local real-time machine; it receives vehicle operation data sent from the local real-time machine and updates the virtual master vehicle state according to the vehicle operation data; after the test period ends, it performs atomic determination on the virtual master vehicle state according to preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, it obtains the autonomous driving test evaluation results;

[0007] The local real-time machine is used to receive real-time sensor data and real-time V2X simulation data sent by the simulation cloud platform, convert the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller, and transmit it to the controller; it receives control commands sent by the controller, obtains vehicle operation data according to the control commands, and sends the vehicle operation data to the simulation cloud platform.

[0008] The controller is used to receive real-time sensor data and real-time V2X simulation data transmitted by the local real-time machine through the corresponding input protocol, analyze the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtain control commands for controlling vehicle driving, and send the control commands to the local real-time machine.

[0009] In specific implementation, the simulation cloud platform includes:

[0010] The vehicle and sensor configuration module is used to configure the virtual master vehicle and corresponding sensors according to preset dynamic parameters; receive vehicle operation data sent from the local real-time machine; and update the virtual master vehicle status according to the vehicle operation data.

[0011] The map and static scene simulation module is used to simulate the road data and static scene of the virtual master vehicle during operation based on the static high-precision map and scene required for the preset simulation scene, as well as the virtual master vehicle state, to obtain the real-time road data and real-time static scene data of the virtual master vehicle during operation, and send them to the dynamic scene simulation module.

[0012] The dynamic scene simulation module is used to perform traffic flow simulation and scene simulation based on real-time road data and real-time static scene data during the operation of the virtual master vehicle, and to obtain dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module.

[0013] The sensor and V2X simulation module is used to simulate the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, so as to obtain real-time sensor data and real-time V2X simulation data.

[0014] The test evaluation module is used to perform atomic determination on the virtual master vehicle state during the test period according to preset atomic determination conditions to obtain atomic results; analyze the virtual master vehicle state during the test period to obtain vehicle state; and combine the atomic results and the vehicle state to obtain autonomous driving test evaluation results.

[0015] Furthermore, to enrich the test cases, in a specific embodiment, the user can configure or manage the maps and scenes used for testing. Accordingly, the simulation cloud platform also includes:

[0016] The map and scene library management module is used to configure the static high-precision maps required for simulation scenes; configure and manage the scenes required for simulation testing; and transmit the configured static high-precision maps and scenes required for simulation scenes to the map and static scene simulation module.

[0017] In a specific embodiment, if the controller is located in a real vehicle, the local real-time machine communicates with the simulation cloud platform via a 5G air interface;

[0018] If the controller is not located in the actual vehicle, the local real-time machine communicates with the simulation cloud platform via Ethernet.

[0019] In this embodiment of the invention, a simulation cloud platform is set up, and a virtual master vehicle and corresponding sensors are configured in the cloud according to preset dynamic parameters; based on the virtual master vehicle state, the operation of the virtual master vehicle is simulated to obtain real-time sensor data and real-time V2X simulation data, which are then sent to the local real-time machine; vehicle operation data sent from the local real-time machine is received, and the virtual master vehicle state is updated according to the vehicle operation data; after the test period ends, the virtual master vehicle state is atomically determined according to preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation result is obtained; this is set up... The local real-time machine receives real-time sensor data and real-time V2X simulation data from the simulation cloud platform, converts this data into the corresponding input protocol for the controller, and transmits it to the controller. It also receives control commands from the controller, obtains vehicle operation data based on these commands, and uploads this data to the simulation cloud platform. The controller receives real-time sensor data and real-time V2X simulation data transmitted from the local real-time machine via the corresponding input protocol, analyzes this data using a pre-set autonomous driving algorithm, obtains control commands for vehicle operation, and sends these commands to the local real-time machine. Compared to existing technologies that separate cloud simulation and on-loop simulation testing, this local real-time machine links the simulation cloud platform and the controller. By testing the actual controller hardware, hardware-in-the-loop simulation can reuse the high-performance servers and massive test scenario library of the cloud simulation, significantly improving hardware resource utilization, reducing testing costs, eliminating the need for reconfiguration, reducing the workload of testing personnel, and thus improving testing efficiency.

[0020] This invention also provides a local real-time machine that can test real controller hardware while reducing testing costs and improving testing efficiency. This local real-time machine is applied in an autonomous driving test simulation system and includes:

[0021] The dynamics simulation module is used to receive dynamic parameters sent by the data interface layer and construct a vehicle dynamics model. During hardware-in-the-loop testing, it drives the pre-constructed vehicle dynamics model according to the control commands sent by the controller received by the data interface layer to obtain vehicle operation data. The vehicle operation data includes: the vehicle's global navigation satellite system signal, inertial measurement signal, and vehicle chassis state data.

[0022] The data interface layer is used to obtain dynamic parameters from the simulation cloud platform and send them to the dynamics simulation module; receive real-time sensor data and real-time V2X simulation data from the simulation cloud platform, convert the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller, and transmit them to the controller; receive control commands sent by the controller, and during hardware-in-the-loop testing, send the control commands to the dynamics simulation module, receive vehicle operation data determined by the dynamics simulation module, and send it to the simulation cloud platform; during vehicle-in-the-loop testing, send the control commands to the real vehicle, obtain the vehicle operation data of the real vehicle, and send it to the simulation cloud platform.

[0023] In a specific embodiment, to support concurrent simulation of multiple in-the-loop controllers and ensure that different controllers do not interfere with each other when simulating the virtual master vehicle simultaneously on the simulation cloud platform, the local real-time machine further includes:

[0024] The cloud registration module is used to set unique identifiers for the dynamics simulation module and the data interface layer using a daemon process. Using these unique identifiers, the dynamics simulation module and the data interface layer are registered into the simulation cloud platform to bind the local real-time machine, the virtual master vehicle, and the corresponding sensors.

[0025] In this embodiment of the invention, a dynamics simulation module is set up to receive dynamic parameters sent by the data interface layer and construct a vehicle dynamics model. During hardware-in-the-loop testing, the constructed vehicle dynamics model is driven according to the control commands sent by the controller received by the data interface layer to obtain vehicle operation data. By setting up the data interface layer, dynamic parameters are obtained from the simulation cloud platform and sent to the dynamics simulation module. Real-time sensor data and real-time V2X simulation data sent by the simulation cloud platform are received, and the real-time sensor data and real-time V2X simulation data are converted into the input protocol corresponding to the controller and transmitted to the controller. Control commands sent by the controller are received, and during hardware-in-the-loop testing, the control commands are sent to the dynamics simulation module. The vehicle operation data determined by the dynamics simulation module is received and sent to the simulation cloud platform. During whole-vehicle-in-the-loop testing, the control commands are sent to the real vehicle to obtain the vehicle operation data of the real vehicle and send it to the simulation cloud platform. By combining the dynamics simulation module and the data interface layer, the simulation cloud platform and the controller are linked together. This not only enables the testing of real controller hardware, but also allows hardware-in-the-loop simulation to reuse the high-performance server and massive test scenario library of the cloud simulation, greatly improving the utilization of hardware resources, reducing testing costs, eliminating the need for reconfiguration, reducing the workload of testers, and thus improving testing efficiency.

[0026] This invention also provides an autonomous driving test simulation method, applied to an autonomous driving test simulation system, comprising:

[0027] The simulation cloud platform configures the virtual master vehicle and corresponding sensors in the cloud according to the preset dynamic parameters;

[0028] The simulation cloud platform simulates the operation of the virtual master vehicle based on the virtual master vehicle state, obtains real-time sensor data and real-time V2X simulation data, and sends the real-time sensor data and real-time V2X simulation data to the local real-time machine.

[0029] The local real-time machine receives real-time sensor data and real-time V2X simulation data, converts the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller, and transmits it to the controller.

[0030] The controller receives the real-time sensor data and real-time V2X simulation data, analyzes the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtains control commands for controlling vehicle driving, and sends the control commands to the local real-time machine.

[0031] The local real-time machine receives control commands sent by the controller, obtains vehicle operation data according to the control commands, and uploads the vehicle operation data to the simulation cloud platform.

[0032] The simulation cloud platform receives vehicle operation data, updates the virtual master vehicle status based on the vehicle operation data, and performs simulation of virtual master vehicle operation based on the updated virtual master vehicle status until the end of the test period.

[0033] After the test period ends, the simulation cloud platform performs atomic determination on the virtual master vehicle state according to the preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation results are obtained.

[0034] In practice, the simulation cloud platform simulates the operation of the virtual master vehicle based on its state, obtaining real-time sensor data and real-time V2X simulation data, including:

[0035] The map and static scene simulation module simulates the road data and static scene during the operation of the virtual master vehicle based on the static high-precision map and scene required by the preset simulation scene, as well as the virtual master vehicle status, to obtain the real-time road data and real-time static scene data during the operation of the virtual master vehicle, and sends them to the dynamic scene simulation module.

[0036] The dynamic scene simulation module performs traffic flow simulation and scene simulation based on real-time road data and real-time static scene data during the operation of the virtual master vehicle, and obtains dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module.

[0037] The sensor and V2X simulation module simulates the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, and obtains real-time sensor data and real-time V2X simulation data.

[0038] In a specific embodiment, after the test period ends, the simulation cloud platform performs atomic determination on the virtual master vehicle state according to preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation results are obtained, including:

[0039] The test evaluation module performs atomic determinations on the virtual master vehicle state during the test period based on preset atomic determination conditions to obtain atomic results; it analyzes the virtual master vehicle state during the test period to obtain the vehicle state; and it combines the atomic results and the vehicle state to obtain the autonomous driving test evaluation results.

[0040] In another specific embodiment, in order to improve the adaptability of the test simulation by adding simulation scenarios, the simulation cloud platform simulates the operation of the virtual master vehicle based on the virtual master vehicle state, and obtains real-time sensor data and real-time V2X simulation data, and also includes:

[0041] The map and scene library management module configures the static high-precision maps required for the simulation scene; it configures and manages the scenes required for simulation testing, and transmits the configured static high-precision maps and scenes required for the simulation scene to the map and static scene simulation module.

[0042] This invention also provides an autonomous driving test simulation method, applied to a local real-time machine, comprising:

[0043] The data interface layer receives control commands sent by the controller;

[0044] During hardware-in-the-loop testing, the data interface layer obtains the dynamic parameters from the simulation cloud platform, sends the control commands and the dynamic parameters to the dynamic simulation module, and the dynamic simulation module constructs a vehicle dynamics model based on the dynamic parameters. The vehicle dynamics model is driven by the control commands to obtain vehicle operation data. The data interface layer receives the vehicle operation data determined by the dynamic simulation module and sends it to the simulation cloud platform.

[0045] During the vehicle-in-the-loop test, the data interface layer sends the control commands to the real vehicle, obtains the vehicle operation data of the real vehicle, and uploads it to the simulation cloud platform.

[0046] The vehicle operation data includes: the vehicle's global navigation satellite system signal, inertial measurement signal, and vehicle chassis status data.

[0047] Furthermore, to enable concurrent simulation of multiple in-the-loop controllers, the autonomous driving test simulation method provided in the specific embodiment also includes:

[0048] The cloud registration module uses a daemon process to set a unique identifier for the dynamics simulation module and the data interface layer. Using this unique identifier, the dynamics simulation module and the data interface layer are registered into the simulation cloud platform to bind the local real-time machine, the virtual master vehicle, and the corresponding sensors.

[0049] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described autonomous driving test simulation method.

[0050] This invention also provides a computer-readable storage medium storing a computer program that executes the above-described autonomous driving test simulation method. Attached Figure Description

[0051] The following figures are intended only to illustrate and explain the present invention and do not limit the scope of the invention. Wherein:

[0052] Figure 1 This is a schematic diagram of the structure of the autonomous driving test simulation system according to an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of the structure of the simulation cloud platform 01 in a specific embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of the structure of the local real-time machine 02 according to an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of the working logic of an autonomous driving test simulation system constructed in a specific embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the structure of the local real-time machine 02 in a specific embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of an autonomous driving test simulation method applied to a driving test simulation system in an embodiment of the present invention;

[0058] Figure 7 This is a schematic diagram illustrating the implementation process of step 602 in a specific embodiment of the present invention;

[0059] Figure 8 This is a schematic diagram illustrating the implementation process of step 602 in another specific embodiment of the present invention;

[0060] Figure 9This is a schematic diagram of an autonomous driving test simulation method applied to a local real-time machine in an embodiment of the present invention;

[0061] Figure 10 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0062] The present application will be further described in detail below with reference to the accompanying drawings and embodiments. Through these descriptions, the features and advantages of the present application will become clearer and more apparent. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present invention.

[0063] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments. Although various aspects of embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated otherwise.

[0064] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0065] While existing cloud simulation technologies facilitate large-scale concurrent testing, in-the-loop simulation solutions require a large number of high-performance simulation workstations and real-time systems when testing multiple ADAS / AD controllers simultaneously. Combining these two approaches could achieve low-cost, high-efficiency large-scale in-the-loop concurrent testing. However, the inventors discovered that cloud simulation solutions are primarily suitable for model and code development phases and cannot test real controller hardware. While in-the-loop simulation solutions can test controllers, each controller under test requires a separate high-performance simulation workstation, and synchronizing high-precision maps and scene libraries between different workstations is also problematic. Furthermore, cloud simulation server clusters are typically located in computer rooms (private clouds), while ADAS / AD controllers are located in laboratories or autonomous vehicles, making their geographical separation even more difficult to integrate.

[0066] Based on the aforementioned difficulties in integration, this embodiment of the invention designs an architecture that integrates autonomous driving cloud simulation and in-loop simulation, binding cloud computing resources with local controller hardware, so that cloud computing resources can be used by local in-loop simulation, thereby unifying the management and scheduling of in-loop simulation scene resources and reducing configuration complexity.

[0067] Specifically, embodiments of the present invention provide an autonomous driving test simulation system that can both test real controller hardware and reduce testing costs while improving testing efficiency, such as... Figure 1 As shown, the system includes:

[0068] Simulation cloud platform 01 is used to configure a virtual master vehicle and corresponding sensors in the cloud according to preset dynamic parameters; based on the virtual master vehicle state, it simulates the operation of the virtual master vehicle, obtains real-time sensor data and real-time V2X simulation data, and sends the real-time sensor data and real-time V2X simulation data to the local real-time machine 02; it receives vehicle operation data sent by the local real-time machine 02, and updates the virtual master vehicle state according to the vehicle operation data; after the test period ends, it performs atomic determination on the virtual master vehicle state according to preset atomic determination conditions, and obtains atomic results; combining the atomic results and the virtual master vehicle state, it obtains the autonomous driving test evaluation results;

[0069] Local real-time machine 02 is used to receive real-time sensor data and real-time V2X simulation data sent by simulation cloud platform 01, convert the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to controller 03, and transmit it to controller 03; it also receives control commands sent by controller 03, obtains vehicle operation data according to the control commands, and sends the vehicle operation data to simulation cloud platform 01.

[0070] The controller 03 is used to receive real-time sensor data and real-time V2X simulation data transmitted by the local real-time machine 02 through the corresponding input protocol, analyze the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtain control commands for controlling vehicle driving, and send the control commands to the local real-time machine 02.

[0071] As can be seen from the functions of the above components, in this embodiment of the invention, by setting up a simulation cloud platform 01, a virtual master vehicle and corresponding sensors are configured in the cloud according to preset dynamic parameters; based on the virtual master vehicle state, the operation of the virtual master vehicle is simulated to obtain real-time sensor data and real-time V2X simulation data, which are then sent to the local real-time machine 02; the vehicle operation data sent by the local real-time machine 02 is received, and the virtual master vehicle state is updated according to the vehicle operation data; after the test period ends, the virtual master vehicle state is atomically determined according to preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation result is obtained; and this is set up... The local real-time machine 02 receives real-time sensor data and real-time V2X simulation data from the simulation cloud platform 01, converts the data into the corresponding input protocol for the controller 03, and transmits it to the controller 03. It also receives control commands from the controller 03, obtains vehicle operation data based on these commands, and uploads the vehicle operation data to the simulation cloud platform 01. The controller 03 receives real-time sensor data and real-time V2X simulation data transmitted by the local real-time machine 02 through the corresponding input protocol, analyzes the data using a preset autonomous driving algorithm, obtains control commands for controlling vehicle driving, and sends the control commands to the local real-time machine 02. Compared to existing technologies that separate cloud simulation and on-loop simulation testing, this approach, by using the local real-time machine 02 to connect the simulation cloud platform 01 and the controller 03, allows for the reuse of the high-performance server and massive test scenario library of the cloud simulation in hardware-in-the-loop simulation, significantly improving hardware resource utilization, reducing testing costs, eliminating the need for reconfiguration, reducing the workload of testing personnel, and thus improving testing efficiency.

[0072] In a specific embodiment, the structure of the simulation cloud platform 01 is as follows: Figure 2 As shown, it includes:

[0073] The vehicle and sensor configuration module 201 is used to configure the virtual master vehicle and corresponding sensors according to preset dynamic parameters; receive vehicle operation data sent from the local real-time machine 02, and update the virtual master vehicle status according to the vehicle operation data;

[0074] The map and static scene simulation module 202 is used to simulate the road data and static scene of the virtual master vehicle during operation based on the static high-precision map and scene required for the preset simulation scene, as well as the virtual master vehicle status, to obtain the real-time road data and real-time static scene data of the virtual master vehicle during operation, and send them to the dynamic scene simulation module.

[0075] The dynamic scene simulation module 203 is used to perform traffic flow simulation and scene simulation based on real-time road data and real-time static scene data during the operation of the virtual master vehicle, and to obtain dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module.

[0076] The sensor and V2X simulation module 204 is used to simulate the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, so as to obtain real-time sensor data and real-time V2X simulation data.

[0077] The test evaluation module 205 is used to perform atomic determination on the virtual master vehicle state during the test period according to the preset atomic determination conditions to obtain atomic results; analyze the virtual master vehicle state during the test period to obtain vehicle state; and combine the atomic results and vehicle state to obtain autonomous driving test evaluation results.

[0078] The preset dynamic parameters are configured in advance by petroleum testing personnel according to testing needs. Accordingly, the simulation cloud platform 01 in the specific embodiment also includes a vehicle and sensor configuration module, which is mainly used to receive configuration requirements from users (testing personnel) and extract preset dynamic parameters based on these requirements. In the specific embodiment, users can manually configure various dynamic parameters on the front-end interface of the simulation cloud platform 01, including parameters such as transmission, braking, steering, suspension, and tires, as well as related parameters such as sensors.

[0079] Specifically, atomic judgment is a relatively independent and indivisible judgment of autonomous vehicle behavior, such as crossing the line, changing lanes, collision, and acceleration exceeding a preset threshold. Combining the judgments for each behavior yields a set of atomic judgment results, which can be used to create higher-level judgments, i.e., atomic results, such as "acceleration > 5 m / s² during lane change". 2 The final evaluation of autonomous driving tests can then be based on the atomic results.

[0080] Furthermore, to enrich the test cases, in this specific embodiment, the user can configure or manage the map and scene used for testing. Accordingly, the simulation cloud platform 01 also includes:

[0081] The map and scene library management module is used to configure the static high-precision maps required for simulation scenes; configure and manage the scenes required for simulation testing; and transmit the configured static high-precision maps and scenes required for simulation scenes to the map and static scene simulation module 202.

[0082] In the specific embodiment, the controller 03 is an ADAS / AD controller, which has built-in various core algorithms related to autonomous driving, including environmental perception algorithms, tracking and prediction algorithms, decision and planning algorithms, vehicle control algorithms, etc. It is used to receive real-time sensor data and real-time V2X simulation data transmitted by the local real-time machine 02 through the corresponding input protocol, analyze the real-time sensor data and real-time V2X simulation data using the preset autonomous driving algorithm, obtain control commands for controlling vehicle driving, and send the control commands to the local real-time machine 02.

[0083] In a specific embodiment, if the controller 03 is located in the real vehicle, the local real-time machine 02 communicates with the simulation cloud platform 01 via the 5G air interface; if the controller 03 is not located in the real vehicle, the local real-time machine 02 communicates with the simulation cloud platform 01 via Ethernet.

[0084] Specifically, real-time sensor data and real-time V2X simulation data are transmitted to the local real-time machine 02 via TCP / UDP protocol. Due to the low latency (<5ms) of Ethernet communication and 5G air interface communication, the joint simulation of cloud simulation and local controller 03 can support frequencies above 200Hz, thereby meeting the needs of various simulation modules in the cloud.

[0085] In practice, the local real-time machine 02 acts as a link between the simulation cloud platform 01 and the controller 03. It needs to acquire cloud resources locally for overall planning and also needs to transmit the control commands of the controller 03 to the cloud. This is mainly achieved through a local real-time machine with a standard configuration.

[0086] This invention also provides a local real-time machine 02 that can test real controller hardware while reducing testing costs and improving testing efficiency. The local real-time machine 02 is applied in an autonomous driving test simulation system, such as... Figure 3 As shown, the structure of the local real-time machine 02 includes:

[0087] The dynamics simulation module 301 is used to receive dynamic parameters sent by the data interface layer 302 and construct a vehicle dynamics model. During hardware-in-the-loop testing, it drives the constructed vehicle dynamics model according to the control commands sent by the controller 03 received by the data interface layer to obtain vehicle operation data.

[0088] The data interface layer 302 is used to obtain dynamic parameters from the simulation cloud platform 01 and send them to the dynamic simulation module 301; receive real-time sensor data and real-time V2X simulation data sent by the simulation cloud platform 01, convert the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller 03, and transmit them to the controller 03; receive control commands sent by the controller 03, and send the control commands to the dynamic simulation module 301 during hardware-in-the-loop testing, receive the vehicle operation data determined by the dynamic simulation module 301, and send it to the simulation cloud platform 01; during vehicle-in-the-loop testing, send the control commands to the real vehicle, obtain the vehicle operation data of the real vehicle, and send it to the simulation cloud platform 01.

[0089] In this embodiment of the invention, a vehicle dynamics model is constructed by setting up a dynamics simulation module 301 to receive dynamic parameters sent by a data interface layer 302. During hardware-in-the-loop testing, the constructed vehicle dynamics model is driven according to the control commands sent by the controller received by the data interface layer 302 to obtain vehicle operation data. The dynamic parameters are obtained from the simulation cloud platform 01 by setting up the data interface layer 302 and sent to the dynamics simulation module 301. Real-time sensor data and real-time V2X simulation data sent by the simulation cloud platform 01 are received, and the real-time sensor data and real-time V2X simulation data are converted into the input protocol corresponding to the controller and transmitted to the controller 03. Control commands are received from the controller 03. During hardware-in-the-loop testing, the control commands are sent to the dynamics simulation module 301, and the vehicle operation data determined by the dynamics simulation module 301 is received and sent to the simulation cloud platform 01. During whole-vehicle-in-the-loop testing, the control commands are sent to the real vehicle to obtain the vehicle operation data of the real vehicle and sent to the simulation cloud platform 01. By utilizing the dynamics simulation module 301 and the data interface layer 302, the simulation cloud platform 01 and the controller 03 are connected. This not only enables the testing of real controller hardware, but also allows hardware-in-the-loop simulation to reuse the high-performance server and massive test scenario library of the cloud simulation, greatly improving the utilization rate of hardware resources, reducing testing costs, eliminating the need for reconfiguration, reducing the workload of testers, and thus improving testing efficiency.

[0090] In practice, the dynamic parameters are typically configured in advance by the testers on the simulation cloud platform 01 according to the test requirements. These parameters are then used to configure the virtual master vehicle in the cloud. The data interface layer 302 obtains the dynamic parameters from the simulation cloud platform 01 and sends them to the dynamic simulation module 301. The dynamic simulation module 301 constructs a vehicle dynamics model based on these configured parameters, ensuring that the parameters configured in the local real-time machine 02 are completely consistent with the cloud configuration, thus guaranteeing the accuracy of the test. For vehicle-in-the-loop testing, as long as the testers ensure that the dynamic parameters configured on the simulation cloud platform 01 are consistent with the actual vehicle used in the test, the test is successful.

[0091] The vehicle operation data includes: the vehicle's Global Navigation Satellite System (GNSS) signal, Inertial Measurement Unit (IMU) signal, and vehicle chassis status data. The vehicle chassis status data includes, but is not limited to: position, orientation, speed, acceleration, angular velocity, throttle, brake, steering wheel angle, gear, and four-wheel status.

[0092] In the entire simulation system, the accuracy of dynamics has the highest requirements for real-time performance and frame rate (e.g., 1000Hz). Therefore, the dynamics simulation module 301 is located in the local real-time machine 02 in the system architecture. It obtains the latest control commands such as throttle, braking, and steering from the controller 03 in real time and drives the vehicle dynamics model. For the vehicle-in-the-loop (VIL) test in an open field, the dynamics simulation module 301 can be skipped, and no simulation is required. The real-time GNSS, IMU, and vehicle chassis status data of the actual vehicle can be obtained directly through the CAN protocol.

[0093] The data interface layer 302 in the local real-time machine 02 needs to convert real-time sensor data and real-time V2X simulation data from the simulation cloud platform 01 into the input protocol corresponding to the controller 03, typically CAN or Ethernet protocol. Simultaneously, the data interface layer 302 needs to transmit the calculation results of the dynamics simulation module 301 to the simulation cloud platform 01 to drive the update of the main vehicle status in the cloud platform. Specifically, for vehicle-in-the-loop (VIL) testing in an open field, data can be uploaded via the 5G air interface.

[0094] Since the modules of the simulation cloud platform 01 have relatively low frame rates (e.g., 60Hz for dynamic scenes, 30Hz for cameras, and 20Hz for millimeter-wave radar), the communication between the simulation cloud platform 01 and the local real-time machine 02 can run at a low frame rate, such as 60Hz, thereby reducing the bandwidth consumption of the entire simulation system and thus increasing the overall concurrency of the simulation system.

[0095] When testing the perception algorithm in controller 03, the data interface layer 302 in the local real-time machine 02 also acquires raw sensor data from the simulation cloud platform 01. Taking camera simulation as an example, it is necessary to acquire network video stream information from the cloud and convert it into the signals required by the camera injection board (such as...). Figure 4 The implementation of the local real-time machine 2 shown) or the signal required for the camera dark box in the loop (such as...) Figure 4 The implementation of the local real-time machine 3 shown in the figure.

[0096] In practical implementation, considering scalability, the computing resources of simulation cloud platform 01 generally adopt a hardware-independent containerized deployment approach. That is, the hardware service layer of simulation cloud platform 01 includes a pool of hardware resources for storage servers and compute servers, as well as a container orchestration and scheduling management platform used to manage these hardware resources. This layer is used to support the operation of the entire cloud simulation system. Each container in the cloud platform has a daemon process, which is used to register nodes with specific simulation capabilities to the cloud platform.

[0097] Because of this containerized deployment architecture, when running a case in the cloud, for the same case, dynamic scene simulation may run on one physical machine, millimeter-wave radar simulation may run on one physical machine, and camera simulation may run on another physical machine that supports GPU.

[0098] However, all hardware resources in local hardware-in-the-loop simulation are deterministic. For example, once the physical wiring between the local real-time machine 02 and the controller under test 03 is completed, the two are completely bound together. This embodiment of the invention enables the local real-time machine 02 and the controller 03 to be bound to a virtual master vehicle in the cloud through a registration and binding mechanism for the hardware in the loop. In this way, as long as a virtual master vehicle bound to a certain local hardware needs to be tested is selected before running a test case in the cloud, the programs in the simulation cloud platform 01, the local real-time machine 02, and the controller 03 can be started synchronously to form a simulation closed loop. At this time, the test case is selected to run this master vehicle, while the simulation of other nodes is still based on the corresponding container in the cloud. However, for the dynamic simulation module 301 and the data interface layer 302 module required by these two master vehicles, since only the daemon process of the local real-time machine 02 has registered this capability, only these two nodes can respond to the test case to be run, thus forming a binding between cloud simulation and local hardware-in-the-loop. Therefore, when multiple controllers 03 need to be simulated concurrently, each controller 03's corresponding real-time machine 02 has a globally unique dynamic simulation module 301 and data interface layer 302 module registered. As long as a virtual master vehicle is configured for each controller under test 03 in the cloud, concurrent simulation of multiple controllers in the loop 03 can be supported in the cloud.

[0099] Accordingly, in order to support concurrent simulation of the multi-in-the-loop controller 03 and to prevent mutual interference when the virtual master vehicle is simultaneously simulated on the simulation cloud platform 01, Figure 5 The local real-time machine 02 shown is in Figure 3 In addition to this, it also includes:

[0100] The cloud registration module 501 is used to set unique identifiers for the dynamics simulation module 301 and the data interface layer 302 using the daemon process. Using the unique identifiers, the dynamics simulation module 301 and the data interface layer 302 are registered into the simulation cloud platform 01 to bind the local real-time machine 02 and the virtual master vehicle and the corresponding sensors.

[0101] In a specific embodiment, a daemon process is added to the local real-time machine 02 to remotely authenticate and connect to the simulation cloud platform 01. Other modules with similar functions on the simulation cloud platform 01 generally share the same identifier to support elastic cloud expansion; for example, all dynamic scene simulation modules share the same identifier. However, the modules registered by the daemon process of the local real-time machine 02 possess completely unique identifiers on the cloud.

[0102] Since the data interface layer 302 in the local real-time machine 02 is directly connected to the ADAS / AD controller 03, when the simulation program starts, the data interface layer 202 can send a simulation start signal to the controller 03, thereby indirectly binding the simulation cloud platform 01 and the controller under test 03. After running, multiple simulation modules (dynamic and static scene simulation, sensor simulation and dynamic simulation, etc.) and autonomous driving algorithms (perception, decision-making, and control algorithms, etc.) of the simulation cloud platform 01 and the local real-time machine 02 will start simultaneously, forming a simulation closed loop in the cloud.

[0103] The following is a specific example illustrating how the autonomous driving test simulation system, which includes a local real-time machine, built according to an embodiment of the present invention, works. The working logic of the autonomous driving test simulation system built in this specific example is as follows: Figure 4 As shown:

[0104] First, there's the classic simulation cloud platform architecture, including a test management front-end, map and static scene simulation modules, dynamic scene simulation modules, sensor and V2X simulation modules, dynamics simulation modules, and autonomous driving algorithm modules (including perception, decision-making, and control algorithms). The test management front-end includes a static map and scene library management module, a vehicle and sensor configuration module, and an evaluation and test report module. All scene data is uniformly stored in a database and data storage service.

[0105] The specific module descriptions are as follows:

[0106] Test Management Front-end - Map and Scene Library Management Module: Used to configure the static high-precision maps required for simulation scenes. The OpenDRIVE map format is the standard simulation map format, which supports static map settings such as lanes, intersections, road markings, traffic signs, and traffic lights. It is also used to configure and manage the simulation test scene library. The scene library generally adopts the OpenSCENARIO standard scene format and can be configured with various complex cases based on the OpenDRIVE high-precision map. It supports the configuration of dynamic vehicles, pedestrians, bicycles, motorcycles, static obstacles or special obstacles on the road, etc.

[0107] Test Management Front-End - Vehicle and Sensor Configuration Module: This module creates a virtual vehicle and configures the required dynamic parameters according to user needs, including dynamic configurations for transmission, braking, steering, suspension, and tires. Various sensors, including cameras, LiDAR, millimeter-wave radar, ultrasonic radar, V2X, GNSS, and IMU, can be configured on the virtual vehicle with intrinsic and extrinsic parameters consistent with real sensors.

[0108] Test Management Front-end - Evaluation and Test Report Module: Used for atomic judgments based on preset atomic judgment conditions in the scenario library (such as crossing the line, changing lanes, collision, timeout, rapid acceleration, etc.), and based on the results of atomic judgments, to conduct autonomous driving test evaluations such as safety, comfort, and efficiency, and generate reports.

[0109] Map and Static Scene Simulation Module: During actual operation of the case, when the road network is based on a high-precision map, it can be used to obtain information such as the location of the main vehicle and traffic participants on the map at any time, including lane lines, stop lines, traffic signs and traffic lights that the autonomous driving algorithm is concerned with; it is also used for perception simulation of 3D digital twin scenes, including roadside buildings, trees, flower beds and other auxiliary facilities.

[0110] Dynamic simulation module: The dynamic simulation module consists of two parts. One part is a simulation unit based on intelligent agent traffic flow; the other part is a scene simulation unit based on custom cases. It mainly uses the OpenSCENARIO scene execution engine to calculate the position, orientation, speed and other information of the opponent vehicle and obstacles in real time during the simulation.

[0111] Sensor simulation module: used for simulating various types of sensors installed on autonomous vehicles, such as camera models including simulation of camera extrinsic parameters, intrinsic parameters, physical parameters, and camera defect parameters; millimeter-wave radar models based on millimeter-wave radar principles using ray tracing and digital signal processing of echoes; lidar models emitting simulated lasers and outputting noisy point clouds based on different material reflection intensity models; ultrasonic radar models returning the distance to detected obstacles based on UPA and APA mathematical models; GNSS / IMU simulation including the cumulative errors of the vehicle's position, speed, and heading when GPS signals are lost.

[0112] V2X Simulation Module: Supports simulation of V2X communication functions based on hierarchical abstract modeling mechanism, while considering the impact of different traffic environments on wireless signal transmission efficiency and packet loss rate.

[0113] Figure 4 The lower half of the document primarily demonstrates several methods for connecting the local controller to the cloud, mainly through a standard local real-time machine. Specific module functions include:

[0114] Dynamics Simulation Module: Constructs and drives vehicle dynamics models based on dynamic parameters, including suspension models, tire models, transmission system models, braking system models, steering system models, etc. The software adopts multibody vehicle dynamics and supports three degrees of freedom of movement in three directions, three degrees of freedom of rotation in three directions, four degrees of freedom of rotation of the tires, four degrees of freedom of unsprung mass and eight degrees of freedom of transient characteristics of the tires, which can fully simulate the overall and internal attitude of the vehicle during operation.

[0115] Data Interface Layer: This layer consists of the API modules for the entire system, mainly including process control API, vehicle communication API, sensor output API, and map query API.

[0116] Once the simulation is running, the dynamic and static scene simulation of the simulation cloud platform, the sensor and V2X simulation, the dynamic simulation and data interface layer of the real-time machine, and the autonomous driving algorithm in the controller form a complete closed loop.

[0117] The autonomous driving test simulation system built in this specific example reduces configuration complexity by uniformly managing and scheduling local in-loop simulation scenario resources; it greatly improves hardware resource utilization by sharing cloud simulation computing resources; and based on the low latency of Ethernet and 5G air interface, the system performance, as tested, fully meets the in-loop requirements of decision-making and control algorithms in ADAS / AD controllers. Furthermore, since camera data is video stream data (approximately 8MB / s after H264 compression of the onboard 1080p camera) and millimeter-wave radar data is structured data with a data volume of <1MB / s), the single-camera + single-millimeter-wave solution consumes less bandwidth overall and can use Ethernet or 5G air interface for low-latency transmission. Even when the number of sensors under test in the controller is small (such as the mainstream single-camera + millimeter-wave fusion solution in ADAS assisted driving), it can still meet the complete closed-loop testing requirements of perception, decision-making, and control algorithms.

[0118] The above specific applications are merely examples; other implementation methods will not be described in detail.

[0119] Based on the same inventive concept, this invention also provides an autonomous driving test simulation method. Since the principle behind the problem solved by this autonomous driving test simulation method is similar to that of the autonomous driving test simulation system, the implementation of the autonomous driving test simulation method can refer to the implementation of the autonomous driving test simulation system. Repeated details will not be elaborated further. This autonomous driving test simulation method is applied to the autonomous driving test simulation system and is implemented based on a simulation cloud platform 01, a local real-time machine 02, and a controller 03, as detailed below. Figure 6 As shown, it includes:

[0120] Step 601: Simulation cloud platform 01 configures the virtual master vehicle and corresponding sensors in the cloud according to the preset dynamic parameters;

[0121] Step 602: Based on the virtual master vehicle state, the simulation cloud platform 01 performs simulation of the virtual master vehicle operation, obtains real-time sensor data and real-time V2X simulation data, and sends the real-time sensor data and real-time V2X simulation data to the local real-time machine 02.

[0122] Step 603: The local real-time machine 02 receives real-time sensor data and real-time V2X simulation data, converts the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller 03, and transmits it to the controller 03.

[0123] Step 604: Controller 03 receives real-time sensor data and real-time V2X simulation data, analyzes the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtains control commands for controlling vehicle driving, and sends the control commands to the local real-time machine 02.

[0124] Step 605: The local real-time machine 02 receives the control command sent by the controller 03, obtains the vehicle operation data according to the control command, and sends the vehicle operation data to the simulation cloud platform 01.

[0125] Step 606: Simulation cloud platform 01 receives vehicle operation data, updates the virtual master vehicle status based on the vehicle operation data, and performs simulation of virtual master vehicle operation based on the updated virtual master vehicle status until the end of the test period;

[0126] Step 607: After the test period ends, the simulation cloud platform 01 performs atomic determination on the virtual master vehicle state according to the preset atomic determination conditions to obtain the atomic result; combining the atomic result and the virtual master vehicle state, the autonomous driving test evaluation result is obtained.

[0127] In practice, the simulation cloud platform 01 first configures the virtual master vehicle and corresponding sensors in the cloud according to the preset dynamic parameters. The specific implementation process includes: the vehicle and sensor configuration module 201 configuring the virtual master vehicle and corresponding sensors according to the user-preset dynamic parameters. For hardware-in-the-loop testing, after the user presets the dynamic parameters for the test on the simulation cloud platform 01, the simulation cloud platform 01 sends these dynamic parameters to the local real-time machine 02. For whole-vehicle-in-the-loop testing, the preset dynamic parameters need to be consistent with the real vehicle used in the test.

[0128] Next, the simulation cloud platform 01 simulates the operation of the virtual master vehicle based on its state, obtaining real-time sensor data and real-time V2X simulation data. The specific implementation process is as follows: Figure 7 As shown, it includes:

[0129] Step 701: The map and static scene simulation module 202 simulates the road data and static scene during the operation of the virtual master vehicle based on the static high-precision map and scene required for the preset simulation scene, as well as the virtual master vehicle status, to obtain the real-time road data and real-time static scene data during the operation of the virtual master vehicle, and sends them to the dynamic scene simulation module 203.

[0130] Step 702: The dynamic scene simulation module 203 performs traffic flow simulation and scene simulation based on the real-time road data and real-time static scene data during the operation of the virtual master vehicle, and obtains the dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module 204.

[0131] Step 703: Sensor and V2X simulation module 204 simulates the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, and obtains real-time sensor data and real-time V2X simulation data.

[0132] In a specific embodiment of the present invention, in order to improve the adaptability of the test simulation by increasing the simulation scenario, the simulation cloud platform 01 simulates the operation of the virtual master vehicle based on the virtual master vehicle state, and obtains real-time sensor data and real-time V2X simulation data. The specific implementation process is as follows: Figure 8 As shown, in Figure 7 In addition to this, it also includes:

[0133] Step 801: The map and scene library management module configures the static high-precision map required for the simulation scene; configures and manages the scene required for simulation testing, and transmits the configured static high-precision map and scene required for the simulation scene to the map and static scene simulation module 202.

[0134] After obtaining real-time sensor data and real-time V2X simulation data, the system sends these data to the local real-time machine 02. The local real-time machine 02 receives the data, converts it into the input protocol corresponding to the controller 03, and transmits it to the controller 03. The controller 03 receives the data, analyzes it using a preset autonomous driving algorithm, obtains control commands for driving the vehicle, and sends these commands to the local real-time machine 02.

[0135] The local real-time machine 02 receives control commands from the controller 03, obtains vehicle operation data based on the control commands, and then uploads the vehicle operation data to the simulation cloud platform 01.

[0136] After the vehicle operation data is sent to the simulation cloud platform 01, the simulation cloud platform 01 receives the vehicle operation data, updates the virtual master vehicle status based on the vehicle operation data, and performs simulation of the virtual master vehicle operation based on the updated virtual master vehicle status until the end of the test period. Specifically, the vehicle and sensor configuration module 201 receives the vehicle operation data sent by the local real-time machine 02, updates the virtual master vehicle status based on the vehicle operation data, and then repeats steps 602-606 until the end of the test period.

[0137] After the test period ends, the simulation cloud platform 01 performs atomic determinations on the virtual master vehicle state according to preset atomic determination conditions, obtaining atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation results are obtained. The specific process includes:

[0138] The test evaluation module 205 performs atomic determinations on the virtual master vehicle state during the test period according to the preset atomic determination conditions, and obtains atomic results; it analyzes the virtual master vehicle state during the test period to obtain the vehicle state; and it combines the atomic results and the vehicle state to obtain the autonomous driving test evaluation results.

[0139] Among them, atomic results refer to all the subtle behaviors of the autonomous vehicle after it starts from the starting point: such as whether it runs a red light, crosses a solid line, has a collision, or has reached the destination; vehicle state refers to the accelerator, brake, and steering states of the autonomous vehicle throughout the entire journey from the starting point to the destination.

[0140] Evaluations generally include:

[0141] Driving safety refers to a vehicle's driving decisions and behaviors on the road. Autonomous vehicles also need to abide by traffic rules and must provide navigation for users in various driving scenarios (whether expected or unexpected) to ensure driving safety.

[0142] The first step is to determine the reliability of the autonomous driving module, such as whether the module will experience fatal software errors, memory leaks, and data delays.

[0143] Secondly, it involves evaluating the basic functions of autonomous driving, such as whether the vehicle follows road signs, whether it hits pedestrians, and whether it causes traffic accidents.

[0144] Driving comfort: refers to the driving experience of the driver or passengers while the vehicle is on the road. It is based on the accelerator, brake and steering status recorded during the driving process to evaluate whether the vehicle is stable and whether the turning is smooth.

[0145] Using a multi-degree-of-freedom driving simulator, the driver's physical perception and psychological feelings are evaluated through a driver-in-the-loop assessment. Physical perception includes assessment systems for yaw angle and jerking sensation, while psychological feelings include feelings of security and sluggishness.

[0146] Traffic coordination refers to the traffic movement performance of a vehicle relative to other traffic participants while driving on a road. It is analyzed and evaluated from the perspective of external traffic participants or the overall situation. Improvement in coordination relies on the continuous self-reinforcing learning of algorithms.

[0147] Standard compatibility: refers to evaluating autonomous driving behavior according to the requirements of laws and regulations in different countries.

[0148] Within the simulation platform, because the output is atomic results, different permutations and combinations and conditional filtering can be used to evaluate autonomous driving algorithms from different dimensions and according to different industry standards.

[0149] Furthermore, this embodiment of the invention also provides an autonomous driving test simulation method, applied to a local real-time machine, such as... Figure 9 As shown, it includes:

[0150] Step 901: Data interface layer 302 receives control commands sent by controller 03;

[0151] Step 902: During hardware-in-the-loop testing, the data interface layer 302 obtains the aforementioned dynamic parameters from the simulation cloud platform 01, and sends the control commands and dynamic parameters to the dynamic simulation module 301. The dynamic simulation module 301 constructs a vehicle dynamics model based on the dynamic parameters; it drives the constructed vehicle dynamics model according to the control commands to obtain vehicle operation data. The data interface layer 302 receives the vehicle operation data determined by the dynamic simulation module 301 and sends it to the simulation cloud platform 01.

[0152] Step 903: During the vehicle-in-the-loop test, the data interface layer 302 sends control commands to the real vehicle, obtains the vehicle operation data of the real vehicle, and sends it to the simulation cloud platform 01.

[0153] The vehicle operation data includes: the vehicle's global navigation satellite system signals, inertial measurement signals, and vehicle chassis status data.

[0154] In a specific embodiment of the present invention, an autonomous driving test simulation method is provided to achieve concurrent simulation of multiple in-the-loop controllers. Figure 9 In addition to this, it also includes:

[0155] The cloud registration module 501 uses a daemon process to set a unique identifier for the dynamics simulation module 301 and the data interface layer 302. Using the unique identifier, the dynamics simulation module 301 and the data interface layer 302 are registered into the simulation cloud platform 01 to bind the local real-time machine 02 and the virtual master vehicle and the corresponding sensors.

[0156] This invention also provides a computer device. Figure 10 This is a schematic diagram of a computer device in an embodiment of the present invention. This computer device is capable of implementing all steps in the autonomous driving test simulation method described above. Specifically, the computer device includes the following components:

[0157] Processor 1001, memory 1002, communications interface 1003, and communication bus 1004;

[0158] The processor 1001, memory 1002, and communication interface 1003 communicate with each other through the communication bus 1004; the communication interface 1003 is used to realize information transmission between related devices.

[0159] The processor 1001 is used to call the computer program in the memory 1002, and when the processor executes the computer program, it implements the autonomous driving test simulation method in the above embodiment.

[0160] This invention also provides a computer-readable storage medium storing a computer program that executes the above-described autonomous driving test simulation method.

[0161] In summary, the autonomous driving test simulation system, method, and local real-time machine provided in the embodiments of the present invention have the following advantages:

[0162] By setting up a simulation cloud platform, a virtual master vehicle and corresponding sensors are configured in the cloud according to preset dynamic parameters. Based on the virtual master vehicle's state, the operation of the virtual master vehicle is simulated to obtain real-time sensor data and real-time V2X simulation data, which are then sent to the local real-time machine. Vehicle operation data sent from the local real-time machine is received, and the virtual master vehicle's state is updated accordingly. After the test period ends, atomic judgments are performed on the virtual master vehicle's state according to preset atomic judgment conditions to obtain atomic results. Combining the atomic results and the virtual master vehicle's state, the autonomous driving test evaluation results are obtained. The local real-time machine is then configured. The system receives real-time sensor data and real-time V2X simulation data from the simulation cloud platform, converts this data into the corresponding input protocol for the controller, and transmits it to the controller. It also receives control commands from the controller, obtains vehicle operation data based on these commands, and uploads this data to the simulation cloud platform. Furthermore, the controller receives real-time sensor data and real-time V2X simulation data transmitted from the local real-time machine via the corresponding input protocol, analyzes this data using a pre-defined autonomous driving algorithm, obtains control commands for vehicle driving, and sends these commands to the local real-time machine. Compared to existing technologies that separate cloud simulation and on-loop simulation testing, this system connects the simulation cloud platform and the controller using a local real-time machine. This allows for testing of the actual controller hardware, while enabling hardware-in-the-loop simulation to reuse the high-performance servers and massive test scenario library of the cloud simulation. This significantly improves hardware resource utilization, reduces testing costs, eliminates the need for reconfiguration, reduces the workload of testing personnel, and thus increases testing efficiency.

[0163] While this invention provides the method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-inventive labor. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual device or client product execution, the methods shown in the embodiments or drawings can be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0164] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, apparatus (systems), or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0165] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0166] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0167] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0168] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are relatively simple in description because they are fundamentally similar to the method embodiments; relevant parts can be referred to the descriptions in the method embodiments. In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. Unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art can understand the specific meaning of these terms in this invention according to the specific circumstances. It should be noted that, without conflict, the embodiments and features in the embodiments of this invention can be combined with each other. This invention is not limited to any single aspect, nor to any single embodiment, nor to any combination and / or substitution of these aspects and / or embodiments. Moreover, each aspect and / or embodiment of this invention can be used alone or in combination with one or more other aspects and / or embodiments.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. An autonomous driving test simulation system, characterized in that, include: The simulation cloud platform is used to configure a virtual master vehicle and corresponding sensors in the cloud according to preset dynamic parameters; Based on the virtual master vehicle state, the virtual master vehicle operation is simulated to obtain real-time sensor data and real-time V2X simulation data, and the real-time sensor data and real-time V2X simulation data are sent to the local real-time machine. Receive vehicle operation data sent from the local real-time machine, and update the virtual master vehicle status based on the vehicle operation data; After the test period ends, the virtual master vehicle state is determined according to the preset atomic determination conditions to obtain the atomic result; the autonomous driving test evaluation result is obtained by combining the atomic result and the virtual master vehicle state. The local real-time machine is used to obtain dynamic parameters from the simulation cloud platform, construct the vehicle dynamics model, and receive real-time sensor data and real-time V2X simulation data sent by the simulation cloud platform. It converts the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller and transmits it to the controller. The system receives control commands from the controller, drives the constructed vehicle dynamics model to obtain vehicle operation data according to the control commands, and uploads the vehicle operation data to the simulation cloud platform. The controller is used to receive real-time sensor data and real-time V2X simulation data transmitted by the local real-time machine through the corresponding input protocol, analyze the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtain control commands for controlling vehicle driving, and send the control commands to the local real-time machine.

2. The autonomous driving test simulation system according to claim 1, characterized in that, The simulation cloud platform includes: The vehicle and sensor configuration module is used to configure the virtual master vehicle and corresponding sensors according to preset dynamic parameters; receive vehicle operation data sent from the local real-time machine; and update the virtual master vehicle status according to the vehicle operation data. The map and static scene simulation module is used to simulate the road data and static scene of the virtual master vehicle during operation based on the static high-precision map and scene required for the preset simulation scene, as well as the virtual master vehicle state, to obtain the real-time road data and real-time static scene data of the virtual master vehicle during operation, and send them to the dynamic scene simulation module. The dynamic scene simulation module is used to perform traffic flow simulation and scene simulation based on real-time road data and real-time static scene data during the operation of the virtual master vehicle, and to obtain dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module. The sensor and V2X simulation module is used to simulate the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, so as to obtain real-time sensor data and real-time V2X simulation data. The test evaluation module is used to perform atomic determination on the virtual master vehicle state during the test period according to preset atomic determination conditions to obtain atomic results; analyze the virtual master vehicle state during the test period to obtain vehicle state; and combine the atomic results and the vehicle state to obtain autonomous driving test evaluation results.

3. The autonomous driving test simulation system according to claim 2, characterized in that, The simulation cloud platform also includes: The map and scene library management module is used to configure the static high-precision maps required for simulation scenes; configure and manage the scenes required for simulation testing; and transmit the configured static high-precision maps and scenes required for simulation scenes to the map and static scene simulation module.

4. The autonomous driving test simulation system according to claim 1, characterized in that, If the controller is located in a real vehicle, the local real-time machine communicates with the simulation cloud platform via a 5G air interface.

5. The autonomous driving test simulation system according to claim 1, characterized in that, If the controller is not located in the actual vehicle, the local real-time machine communicates with the simulation cloud platform via Ethernet.

6. A local real-time machine, characterized in that, The local real-time machine is used in the autonomous driving test simulation system, and includes: The dynamics simulation module is used to receive dynamic parameters sent by the data interface layer and construct a vehicle dynamics model. During hardware-in-the-loop testing, it drives the constructed vehicle dynamics model according to the control commands sent by the controller received by the data interface layer to obtain vehicle operation data. The vehicle operation data includes: the vehicle's global navigation satellite system signal, inertial measurement signal, and vehicle chassis state data. The data interface layer is used to obtain dynamic parameters from the simulation cloud platform and send them to the dynamics simulation module; receive real-time sensor data and real-time V2X simulation data from the simulation cloud platform, convert the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller, and transmit them to the controller; receive control commands sent by the controller, and during hardware-in-the-loop testing, send the control commands to the dynamics simulation module, receive vehicle operation data determined by the dynamics simulation module, and send it to the simulation cloud platform; during vehicle-in-the-loop testing, send the control commands to the real vehicle, obtain the vehicle operation data of the real vehicle, and send it to the simulation cloud platform.

7. The local real-time machine according to claim 6, characterized in that, Also includes: The cloud registration module is used to set unique identifiers for the dynamics simulation module and the data interface layer using a daemon process. Using these unique identifiers, the dynamics simulation module and the data interface layer are registered into the simulation cloud platform to bind the local real-time machine, the virtual master vehicle, and the corresponding sensors.

8. An autonomous driving test simulation method, characterized in that, The autonomous driving test simulation method, applied to an autonomous driving test simulation system, includes: The simulation cloud platform configures the virtual master vehicle and corresponding sensors in the cloud according to the preset dynamic parameters; The simulation cloud platform simulates the operation of the virtual master vehicle based on the virtual master vehicle state, obtains real-time sensor data and real-time V2X simulation data, and sends the real-time sensor data and real-time V2X simulation data to the local real-time machine. The local real-time machine receives real-time sensor data and real-time V2X simulation data, converts the real-time sensor data and real-time V2X simulation data into the input protocol corresponding to the controller, and transmits it to the controller. The controller receives the real-time sensor data and real-time V2X simulation data, analyzes the real-time sensor data and real-time V2X simulation data using a preset autonomous driving algorithm, obtains control commands for controlling vehicle driving, and sends the control commands to the local real-time machine. The local real-time machine obtains dynamic parameters from the simulation cloud platform, constructs a vehicle dynamics model, and receives control commands sent by the controller. Based on the control commands, it drives the constructed vehicle dynamics model to obtain vehicle operation data and uploads the vehicle operation data to the simulation cloud platform. The simulation cloud platform receives vehicle operation data, updates the virtual master vehicle status based on the vehicle operation data, and performs simulation of virtual master vehicle operation based on the updated virtual master vehicle status until the end of the test period. After the test period ends, the simulation cloud platform performs atomic determination on the virtual master vehicle state according to the preset atomic determination conditions to obtain atomic results; combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation results are obtained.

9. The autonomous driving test simulation method according to claim 8, characterized in that, The simulation cloud platform simulates the operation of a virtual master vehicle based on its state, obtaining real-time sensor data and real-time V2X simulation data, including: The map and static scene simulation module simulates the road data and static scene during the operation of the virtual master vehicle based on the static high-precision map and scene required by the preset simulation scene, as well as the virtual master vehicle status, to obtain the real-time road data and real-time static scene data during the operation of the virtual master vehicle, and sends them to the dynamic scene simulation module. The dynamic scene simulation module performs traffic flow simulation and scene simulation based on real-time road data and real-time static scene data during the operation of the virtual master vehicle, and obtains dynamic scene data during the operation of the virtual master vehicle, which is then sent to the sensor and V2X simulation module. The sensor and V2X simulation module simulates the sensor detection and V2X communication functions of the virtual master vehicle based on the dynamic scene data during the operation of the virtual master vehicle, and obtains real-time sensor data and real-time V2X simulation data.

10. The autonomous driving test simulation method according to claim 9, characterized in that, The simulation cloud platform simulates the operation of a virtual master vehicle based on its state, obtaining real-time sensor data and real-time V2X simulation data, and also includes: The map and scene library management module configures the static high-precision maps required for the simulation scene; it configures and manages the scenes required for simulation testing, and transmits the configured static high-precision maps and scenes required for the simulation scene to the map and static scene simulation module.

11. The autonomous driving test simulation method according to claim 8, characterized in that, After the test period ends, the simulation cloud platform performs atomic determination on the virtual master vehicle status according to the preset atomic determination conditions and obtains the atomic results; Combining the atomic results and the virtual master vehicle state, the autonomous driving test evaluation results are obtained, including: The test evaluation module performs atomic determinations on the virtual master vehicle state during the test period based on preset atomic determination conditions to obtain atomic results; it analyzes the virtual master vehicle state during the test period to obtain the vehicle state; and it combines the atomic results and the vehicle state to obtain the autonomous driving test evaluation results.

12. An autonomous driving test simulation method, characterized in that, The autonomous driving test simulation method, applied to a local real-time machine, includes: The data interface layer receives control commands sent by the controller; During hardware-in-the-loop testing, the data interface layer obtains dynamic parameters from the simulation cloud platform and sends control commands and dynamic parameters to the dynamic simulation module. The dynamic simulation module constructs a vehicle dynamics model based on the dynamic parameters. The vehicle dynamics model is driven by the control commands to obtain vehicle operation data. The data interface layer receives the vehicle operation data determined by the dynamic simulation module and sends it to the simulation cloud platform. During the vehicle-in-the-loop test, the data interface layer sends the control commands to the real vehicle, obtains the vehicle operation data of the real vehicle, and uploads it to the simulation cloud platform. The vehicle operation data includes: the vehicle's global navigation satellite system signal, inertial measurement signal, and vehicle chassis status data.

13. The autonomous driving test simulation method according to claim 12, characterized in that, Also includes: The cloud registration module uses a daemon process to set a unique identifier for the dynamics simulation module and the data interface layer. Using this unique identifier, the dynamics simulation module and the data interface layer are registered into the simulation cloud platform to bind the local real-time machine, the virtual master vehicle, and the corresponding sensors.

14. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 8 to 13.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that performs the method of any one of claims 8 to 13.

Images