Intelligent digital rail transit simulation test system and method

Abstract

The invention discloses an intelligent digital rail transit simulation test system based on a digital twinborn technology, which comprises a digital track simulation platform, a digital track internal test platform and a central digital track server, and a visual digital track 3D scene model is constructed by joint design based on the rubber wheel train dynamic model data and the digital track urban construction coordinate graph. The digital track internal test platform comprises a test navigation control unit, and a digital track mathematical model, an electronic map, magnetic coding data and environment perception simulation information are preset in the test navigation control unit. The train equipment state can be monitored, the train equipment state transition can be controlled in real time, the artificial intelligence machine learning technology is introduced, and the test effectiveness is improved; and a real vehicle and a simulation machine can be accessed at the same time to be converted into a test platform before internal product release. The invention further discloses an intelligent digital rail transit simulation test method based on the digital twinning technology.

Description

Technical Field

[0001] The invention relates to an intelligent digital rail transit simulation test system and method. Background Art

[0002] In the traditional intelligent digital rail transit simulation test system, the system simply monitors the equipment status and cannot control the equipment status change in real time. If the operation of rubber-wheeled trains is remotely controlled in combination with remote sensing technology, it will pose a great challenge to the safety of the operation process. The premise for ensuring all this is to achieve environmental perception fusion technology based on digital tracks and ultra-low latency real-time communication technology of 5G, and according to the level of safety hazards, a dual warning mechanism of on-board human-machine + remote is required.

[0003] In the traditional intelligent digital rail transit simulation test system, the system simply collects massive data for model prediction. If the data is not pre-processed and the error data with large deviations is cleaned, the results are often unreliable. In view of the fact that the operation data of the intelligent digital rail system itself is subject to many interference factors, it is necessary to introduce artificial intelligence machine learning technology, which has a good processing mode for noisy data and prediction confidence intervals, to provide assistance for the design of reverse algorithm parameter adjustment.

[0004] Moreover, the application of traditional digital twin technology in the intelligent transportation industry only focuses on the visual display of customer demand, and fails to focus on the simulation and testing of internal enterprise demand systems. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the prior art and provide an intelligent digital rail transit simulation test system based on digital twin technology, which can not only monitor the status of train equipment, but also control the status change of train equipment in real time. It introduces artificial intelligence machine learning technology to improve the effectiveness of the test; it can be connected to the real vehicle + simulation machine at the same time, and the system is converted into an internal product pre-release test platform.

[0006] Another object of the present invention is to provide an intelligent digital rail transit simulation test method based on digital twin technology.

[0007] A technical solution to achieve the above purpose is: an intelligent digital rail transit simulation test system based on digital twin technology, including a digital track simulation platform, a digital track internal test platform and a central digital track server, wherein:

[0008] The digital track simulation platform is a visualized digital track 3D scene model designed and constructed in combination with the rubber-wheeled train vehicle dynamics model, and relying on the rubber-wheeled train vehicle dynamics model data and the digital track urban construction coordinate diagram;

[0009] The digital track internal test platform includes a test navigation control unit, in which a digital track mathematical model, an electronic map, magnetic coding data and environmental perception simulation information are pre-installed;

[0010] The central digital track server is a digital track system management server deployed in the center, and the central digital track server has a built-in real-time database and a digital track rule library;

[0011] The on-board navigation control unit of the rubber-wheeled train obtains the status of the roadside equipment and driving data through the in-vehicle communication network, and transmits the corresponding data back to the central digital rail server through the 5G vehicle-ground communication network, and stores it in the real-time database of the central digital rail server;

[0012] The digital track simulation platform monitors whether there is any abnormality in the data in the real-time database of the central digital track server, and decides whether to remotely operate the equipment or take over the automatic operation control of the vehicle through the platform dispatcher. At the same time, the digital track simulation platform calls the data in the real-time database of the central digital track server to drive the operation of the digital track 3D scene model, mapping the operating status of the vehicle and roadside equipment in the real environment;

[0013] The communication between the digital track simulation platform and the test navigation control unit adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform sends the vehicle position, speed, posture and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model to the digital track mathematical model preset by the test navigation control unit in the form of network data packets, and then quickly retrieves and solves the magnetic nail information closest to the current vehicle according to the electronic map and magnetic coding data loaded in advance, and then adds environmental perception simulation information, and integrates the magnetic nail information solution, and injects the solution result into the vehicle's lateral and longitudinal control algorithm. When the vehicle's lateral and longitudinal control algorithm gives a control instruction, the lateral and longitudinal control instruction is encapsulated and sent to the vehicle dynamics model through a network data packet, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

[0014] The above-mentioned intelligent digital rail transit simulation test system, wherein, after the digital track simulation platform obtains the digital track test data, the digital track test data is imported into the machine self-learning model for rubber-wheel train control, and through repeated training and test iterations, the optimized parameters are substituted into the digital track internal test platform, while improving the test effectiveness, reversely improving the algorithm design.

[0015] The above-mentioned intelligent digital rail transit simulation test system, wherein the digital rail internal test platform also includes a driving simulator, and a CAN / serial port communication method is used to establish communication between the driving simulator and the digital rail simulation platform, and by simulating the input of third-party control instructions, the smoothness of the driving mode switching of the digital rail mathematical model and the anti-disturbance ability of the simulation test system are verified.

[0016] An intelligent digital rail transit simulation test method based on digital twin technology, characterized in that it includes the following steps:

[0017] S1, forming a 3D scene model based on the digital magnetic source high-precision map and building a digital track simulation platform, specifically including the following steps:

[0018] S11, generate a digital track 2D scene model: high-precision mapping of magnetic nail coordinate points, draw a magnetic nail urban construction coordinate map through CAD, fix the RTK antenna to the front of the rubber-wheeled train, and manually drive to collect high-precision GPS train running tracks; after fitting the GPS and magnetic nail map tracks, draw a digital track 2D scene model;

[0019] S12, digital-analog conversion to generate digital track simulation platform: Relying on the 2D scene model based on digital twin technology and the 3D mechanical drawings of roadside facilities and equipment, 3D visualization 1:1 modeling restoration is performed, and the full equipment status information is displayed through the echarts front end to obtain the digital track 3D scene model and build a digital track simulation platform framework;

[0020] S2, the scenario model establishes a real-time communication mechanism with the rubber-wheeled train, which specifically includes the following steps:

[0021] S21, digital rail vehicle-ground data communication: When the rubber-wheeled train is powered on, the on-board controller, sensor and processor transmit the driving data to the on-board navigation control unit in real time through the TRDP protocol. The roadside equipment is first collected by the roadside unit, processed once by the edge computing unit, and then the roadside data is transmitted to the on-board navigation control unit through the 5G vehicle-ground communication network. The on-board navigation control unit fuses and solves the driving data and roadside data, and then sends the obtained optimal control instructions to the vehicle actuator, so that the rubber-wheeled train takes corresponding actions; at the same time, the above data information is also forwarded to the central digital rail server through the 5G vehicle-ground communication network and stored in the real-time database of the central digital rail server; the digital rail simulation platform calls the data in the real-time database to drive the operation of the digital rail 3D scene model, so as to map the operating status of the vehicle and roadside equipment in the real environment;

[0022] S22, digital rail system fault warning: Based on the fault code feedback stored in the real-time database and provided by each device, the corresponding status code is assigned according to the fault level, and displayed in different colors in the front-end visualization area of ​​the central digital rail server; the fault level is divided into operational safety fault and operational non-safety fault;

[0023] S23, remote control of digital rail system trains: under the condition of satisfying the automatic driving operation, the operation status of the rubber-wheeled train is remotely controlled by controlling the computer-specific peripheral 5G, and the operation status of the rubber-wheeled train includes the start and stop, acceleration and deceleration, and the entry / switching / exit of the horizontal and vertical automatic control mode of the rubber-wheeled train; at this time, combined with the fault feedback status in step S22, when an operational safety fault is reported, the system is forced to stop operation; and when an operational non-safety fault is reported, the personnel remotely operate the train to slow down or stop according to the situation;

[0024] S24, digital rail system operation history review: the digital rail simulation platform calls the video stream data in the real-time database of the central digital rail server, imports and drives the model, and synchronizes the timestamp to visualize and reproduce the historical scenes;

[0025] S3, a real-time communication mechanism is established between the digital track simulation platform and the digital track internal test platform: the digital track internal test platform includes a test navigation control unit, and the communication between the digital track simulation platform and the test navigation control unit adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform sends the vehicle position, speed, attitude and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model in the form of network data packets to the digital track mathematical model preset by the test navigation control unit, and then quickly retrieves and solves the magnetic nail information closest to the current vehicle according to the electronic map and magnetic coding data loaded in advance, and then adds environmental perception simulation information, and integrates the magnetic nail information solution, and injects the solution result into the vehicle's lateral and longitudinal control algorithm. When the vehicle's lateral and longitudinal control algorithm gives a control instruction, the lateral and longitudinal control instruction is encapsulated and sent to the vehicle dynamics model through a network data packet, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

[0026] In the above-mentioned intelligent digital rail transit simulation test method based on digital twin technology, in step S3, a machine self-learning model training link is added to the digital rail simulation platform to filter abnormal data and enhance the prediction ability of the digital rail mathematical model. First, the data is divided into a training set and a test set in proportion. Secondly, the model is trained using the training set data, the parameters are adjusted to reduce the training error, and the model prediction effect is evaluated on the test set; the above process is repeated several times and the average is taken to reduce the data division error; finally, the features and models are continuously optimized according to the prediction error to improve the prediction accuracy, and the key parameters of the calibrated vehicle control model state machine are output, which in turn assists the algorithm optimization design.

[0027] In the above-mentioned intelligent digital rail transit simulation test method based on digital twin technology, in step S3, a driving simulator is added to the internal test platform of the digital rail, and the communication between the driving simulator and the digital rail simulation platform is established by using the CAN / serial port communication method. By simulating the input of third-party control commands, the smoothness of the driving mode switching of the digital rail mathematical model and the anti-disturbance ability of the simulation test system are verified.

[0028] The technical solution of the intelligent digital rail transit simulation test system and method of the present invention has the following beneficial effects:

[0029] (1) It solves the problem that the application of digital twin technology in the intelligent transportation industry not only focuses on data interaction (monitoring but not control). In the intelligent digital rail transit simulation test system, the system no longer simply monitors the status of the equipment, but can also control the status change of the equipment in real time. Combined with remote sensing technology, the operation of rubber-wheeled trains can be remotely controlled. This poses a great challenge to the safety of the operation process. The premise for ensuring all this is to achieve the environmental perception fusion technology based on digital rails and the ultra-low latency real-time communication technology of 5G, and to implement the on-board human-machine + remote dual warning mechanism according to the level of safety hazards.

[0030] (2) The application of digital twin technology in the intelligent transportation industry is not only about establishing a big data analysis model. In the intelligent digital rail transit simulation test system, the system no longer simply collects massive data for model prediction. If the data is not preprocessed and the error data with large deviations is cleaned, the results obtained are often unreliable. In view of the fact that the operating data of the intelligent digital rail system itself is subject to many interference factors, artificial intelligence machine learning technology is introduced here, which has a good processing mode for noisy data and prediction confidence intervals. It can provide assistance for the design of reverse algorithm parameter adjustment.

[0031] (3) The application of digital twin technology in the intelligent transportation industry not only focuses on the visual display of customer demand. As the name suggests, the intelligent digital rail transit simulation test system has added the simulation and testing of the internal demand system of the enterprise. A hardware-in-the-loop simulation (HIL) test platform has been established, which can be connected to the real vehicle + simulator at the same time. At this time, the system is converted into an internal test platform before the product is released. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is the schematic diagram of the intelligent digital rail transit simulation test system based on digital twin technology;

[0033] Figure 2 It is a schematic diagram of the communication between the digital track simulation platform and the central digital track server;

[0034] Figure 3It is a schematic diagram of the data flow between the digital track simulation platform and the digital track internal test platform;

[0035] Figure 4 A simulation test flow chart of the intelligent digital rail transit simulation test system of the present invention for a model vehicle. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the technical solution of the present invention, the specific implementation methods thereof are described in detail below in conjunction with the accompanying drawings:

[0037] See also Figures 1 to 3 The intelligent digital rail transit simulation test system of the present invention combines the rubber-wheeled train vehicle dynamics model, and relies on the model data and the digital rail urban construction coordinate diagram to jointly design and build a digital rail 3D visualization simulation platform (hereinafter referred to as the "digital rail simulation platform"), which is deployed on the central digital rail system management server (hereinafter referred to as the "central digital rail server").

[0038] On the one hand, the real-time communication interface between the train and the ground is opened. The onboard navigation control unit (NCU) of the rubber-wheeled train obtains the status of key driving equipment and driving data through the in-vehicle communication network, and then transmits the corresponding data back to the central digital track server through the digital track vehicle-ground dedicated network. After the digital track simulation test platform monitors whether there are any abnormalities in these data, the platform dispatcher decides whether to remotely operate the equipment or take over the automatic operation control of the vehicle. At this time, the 3D model will also reproduce the scene in real time.

[0039] On the other hand, a real-time communication interface between the digital track simulation platform and the digital track hardware-in-the-loop simulation (HIL) internal test platform (hereinafter referred to as the "digital track internal test platform") was also opened. After the digital track simulation platform obtains the digital track test data, it puts the data into the machine learning model of train control. Through repeated training and test iterations, the optimized parameters are substituted into the digital track internal test platform. While improving the effectiveness of the test, the algorithm design is improved by reverse deduction.

[0040] The intelligent digital rail transit simulation test system based on digital twin technology of the present invention includes a digital rail simulation platform 1, a digital rail internal test platform 2 and a central digital rail server 3.

[0041] The digital track simulation platform 1 is a visualized digital track 3D scene model designed and built in combination with the rubber-wheeled train vehicle dynamics model, relying on the rubber-wheeled train vehicle dynamics model data and the digital track urban construction coordinate map; the digital track simulation platform 1 can be used for equipment monitoring, scene management, historical backtracking and machine learning.

[0042] The digital track internal test platform 2 comprises a test navigation control unit 21 , in which a digital track mathematical model, an electronic map, magnetic coding data and environmental perception simulation information are preset.

[0043] The central digital track server 3 is a digital track system management server deployed in the center. The central digital track server 3 has a built-in real-time database 31 and a digital track rule library 32 .

[0044] The on-board navigation control unit of the rubber-wheeled train obtains the roadside equipment status and driving data through the in-vehicle communication network, and transmits the corresponding data back to the central digital track server 3 through the 5G vehicle-to-ground communication dedicated network 4, and stores it in the real-time database 31 of the central digital track server 3.

[0045] The digital track simulation platform 1 monitors whether there are any abnormalities in the real-time database 31 of the central digital track server 3, and decides whether to remotely operate the equipment or take over the automatic operation control of the vehicle through the platform dispatcher. At the same time, the digital track simulation platform 1 calls the data in the real-time database of the central digital track server to drive the operation of the digital track 3D scene model, mapping the operating status of vehicles and roadside equipment in the real environment.

[0046] The communication between the digital track simulation platform 1 and the test navigation control unit 21 adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform 1 sends the vehicle position, speed, posture and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model to the digital track mathematical model preset by the test navigation control unit 21 in the form of network data packets. Then, according to the electronic map and magnetic coding data loaded in advance, the magnetic nail information closest to the current vehicle is quickly retrieved and solved, and then the environmental perception simulation information is added, and the magnetic nail information is integrated and solved, and the solution results are injected into the vehicle's lateral and longitudinal control algorithms. When the vehicle's lateral and longitudinal control algorithms give control instructions, the lateral and longitudinal control instructions are encapsulated and sent to the vehicle dynamics model through network data packets, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

[0047] After the digital track simulation platform 1 obtains the digital track test data, the digital track test data is imported into the machine self-learning model for rubber-wheeled train control. Through repeated training and test iterations, the optimized parameters are substituted into the digital track internal test platform, thereby improving the test effectiveness and improving the algorithm design.

[0048] The digital track internal test platform 2 also includes a driving simulator 22, which uses CAN / serial port communication to establish communication between the driving simulator 22 and the digital track simulation platform 1. By simulating the input of third-party control commands, the smoothness of the driving mode switching of the digital track mathematical model is verified (the system sets the brake pedal to exit the automatic driving mode) and the anti-disturbance ability of the simulation test system (when the vehicle is running automatically, if the steering wheel is accidentally turned and the heading angle exceeds the preset threshold, the direction will be automatically corrected).

[0049] An intelligent digital rail transit simulation test method based on digital twin technology, characterized in that it includes the following steps:

[0050] S1, forming a 3D scene model based on the digital magnetic source high-precision map and building a digital track simulation platform, specifically including the following steps:

[0051] S11, generate a digital track 2D scene model: high-precision mapping of magnetic nail coordinate points, draw a magnetic nail urban construction coordinate map through CAD, fix the RTK antenna to the front of the rubber-wheeled train, and manually drive to collect high-precision GPS train running tracks; after fitting the GPS and magnetic nail map tracks, draw a digital track 2D scene model 5 (see Figure 2 );

[0052] S12, digital-analog conversion to generate digital track simulation platform: Relying on the 2D scene model based on digital twin technology and the 3D mechanical drawings of roadside facilities and equipment, 3D visualization 1:1 modeling restoration is performed, and the full equipment status information is displayed through the echarts front end to obtain the digital track 3D scene model and build a digital track simulation platform framework;

[0053] S2, the scenario model establishes a real-time communication mechanism with the rubber-wheeled train, which specifically includes the following steps:

[0054] S21, digital rail vehicle-ground data communication: When the rubber-wheeled train is powered on, the on-board controller, sensor and processor transmit the driving data to the on-board navigation control unit NCU in real time through the TRDP protocol, and the roadside equipment is first collected by the roadside unit RSU, processed once by the edge computing unit MEC, and then the roadside data is transmitted to the on-board navigation control unit NCU through the 5G vehicle-ground communication network. The on-board navigation control unit NCU integrates and solves the driving data and roadside data, and then sends the obtained optimal control instructions to the vehicle actuator, so that the rubber-wheeled train takes corresponding actions; at the same time, the above data information is also forwarded to the central digital rail server 3 through the 5G vehicle-ground communication network 4, and stored in the real-time database 31 of the central digital rail server 3; the digital rail simulation platform 1 calls the data in the real-time database 31 to drive the operation of the digital rail 3D scene model, so as to map the operating status of the vehicle and roadside equipment in the real environment;

[0055] S22, digital rail system fault warning: based on the fault code feedback stored in the real-time database and provided by each device, a corresponding status code is assigned according to the fault level, and displayed in different colors in the front-end visualization area of ​​the central digital rail server 3; the fault level is divided into operational safety fault and operational non-safety fault;

[0056] S23, remote control of digital rail system trains: under the condition of satisfying the automatic driving operation, the operation status of the rubber-wheeled train is remotely controlled by controlling the computer-specific peripheral 5G, and the operation status of the rubber-wheeled train includes the start and stop, acceleration and deceleration, and the entry / switching / exit of the horizontal and vertical automatic control mode of the rubber-wheeled train; at this time, combined with the fault feedback status in step S22, when an operational safety fault is reported, the system is forced to stop operation; and when an operational non-safety fault is reported, the personnel remotely operate the train to slow down or stop according to the situation;

[0057] S24, digital rail system operation history review: the digital rail simulation platform calls the video stream data in the real-time database of the central digital rail server, imports and drives the model, and synchronizes the timestamp to visualize and reproduce the historical scenes;

[0058] S3, the digital track simulation platform and the digital track internal test platform establish a real-time communication mechanism:

[0059] S31, data communication between the digital track simulation platform and the internal data platform: the digital track internal test platform includes a test navigation control unit, and the communication between the digital track simulation platform and the test navigation control unit adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform sends the vehicle position, speed, attitude and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model in the form of network data packets to the digital track mathematical model preset by the test navigation control unit, and then quickly retrieves and solves the magnetic nail information (magnetic nail polarity, magnetic nail deviation, etc.) closest to the current vehicle according to the electronic map and magnetic coding data loaded in advance, and then adds environmental perception simulation information, and integrates the magnetic nail information solution, and injects the solution result into the vehicle's lateral and longitudinal control algorithm. When the vehicle's lateral and longitudinal control algorithm gives a control instruction, the lateral and longitudinal control instruction is encapsulated and sent to the vehicle dynamics model through a network data packet, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

[0060] S32, digital track model self-learning reinforcement training mechanism: A machine self-learning model training link is added to the digital track simulation platform to filter abnormal data and enhance the prediction ability of the digital track mathematical model. First, the data is divided into a training set and a test set in proportion. Secondly, the model is trained using the training set data, and the parameters are adjusted to reduce the training error. The model prediction effect is evaluated on the test set. The above process is repeated several times and the average is taken to reduce the data division error. Finally, the features and models are continuously optimized according to the prediction error to improve the prediction accuracy. The key parameters of the calibrated vehicle control model state machine are output, and this is used to reversely assist the algorithm optimization design.

[0061] A driving simulator 22 is added to the digital track internal test platform 2, and the communication between the driving simulator 22 and the digital track simulation platform 1 is established using CAN / serial port communication. By simulating the input of third-party control commands, the smoothness of the driving mode switching of the digital track mathematical model and the anti-disturbance ability of the simulation test system are verified.

[0062] The intelligent digital rail transit simulation test system based on digital twin technology (hereinafter referred to as "digital rail transit simulation test system") of the present invention is different from the conventional intelligent transportation simulation system, mainly in the following aspects:

[0063] 1. The fundamental difference between rail transit and road bus simulation systems is:

[0064] (1) Different vehicle models and transport capacity. Due to the differences in chassis, the complexity of establishing simulation dynamic models and the parameter setting rules of rail trains and buses are completely different. The digital track simulation test system uses a new type of rubber-wheeled train as the carrier of the rubber-wheeled tram on the road. Its chassis design retains the core of the rail train and combines the road advantages of the bus steering, traction and braking systems. When establishing the simulation kinematic model, it takes into account environmental restrictions such as road bearing capacity, and ensures predictable road operation conditions such as insufficient or saturated transport capacity.

[0065] (2) Different driving routes and operation organizations. Rail transit has fixed infrastructure such as operation lines and stations, and has better capacity shaving and valley filling effects. However, the operation routes and schedules of road buses are complex and are greatly affected by traffic conditions. The digital rail simulation test system combines the characteristics of fixed rail transit infrastructure (dedicated lanes built on the basis of digital magnetic tags) and relatively simple line network operation organization with the advantages of flexible dispatching of road buses, green wave traffic setting, and signal priority control. When conducting traffic flow simulation, a variety of working conditions and strategies can be set to improve the accuracy of the output results.

[0066] 2. Different levels of virtual-reality mapping and information interaction:

[0067] Conventional intelligent traffic simulation systems are nothing more than 3D visualization environment rendering effects, rubber-tyred tram dynamics / kinematics models, behavior modeling, traffic flow simulation management and control, and result analysis. However, these simulation models are often operated through historical data or delayed communication, and the timeliness of the results obtained is greatly reduced.

[0068] The digital track simulation test system relies on the dedicated communication network of the actual vehicle site to transmit the comprehensive operating status of the rubber-wheeled tram back to the simulation system in real time. Input the control command on the simulation system, and the actual vehicle will take corresponding actions after receiving the information. At the same time, each equipment module is a separate information source, and mapping it on the simulation system is not only information synchronization, but also the physical state of the equipment operation can be reproduced to a certain extent. In terms of information synchronization, for example, if the current rubber-wheeled tram does not detect a certain magnetic tag, the corresponding magnetic tag model icon in the simulation system will light up in different colors, and selecting the magnetic tag will automatically pop up its related information; in terms of equipment physical state monitoring, for example, when the steering wheel of the actual rubber-wheeled tram turns a certain angle when turning, the steering wheel inside the rubber-wheeled tram model on the current simulation system will also turn the same angle. The same is true for the status mapping of other facilities such as traffic lights. The above is based on digital twin technology and its new application in intelligent digital track systems.

[0069] In addition, the digital track simulation test system also has the ability to test, verify and optimize algorithms. In principle, the digital track simulation test system can simulate and reproduce the system operation status through static historical data and dynamic real-time communication. Noise factors are added to conduct anti-disturbance analysis on the system operation results. The obtained deviation data can be grouped in proportion and substituted into the training set and test set of the algorithm model for model self-learning. According to the prediction error, the features and models are continuously optimized to improve the prediction accuracy.

[0070] The intelligent digital rail transit simulation test system and method of the present invention rely on the dedicated communication network of the site where the actual vehicle is located to transmit the comprehensive operating status of the rubber-wheeled tram back to the simulation system in real time. Input the control command on the simulation system, and the actual vehicle takes corresponding actions after receiving the information. At the same time, each device module is a separate information source, and mapping it on the simulation system is not only information synchronization, but also the physical state of the equipment operation can be reproduced to a certain extent. In terms of information synchronization, for example, if the current rubber-wheeled tram does not detect a certain magnetic tag, the corresponding magnetic tag model icon in the simulation system will light up in different colors, and selecting the magnetic tag will automatically pop up its related information; in terms of equipment physical state monitoring, for example, when the steering wheel of the actual rubber-wheeled tram turns a certain angle when turning, the steering wheel inside the rubber-wheeled tram model currently on the simulation system will also turn the same angle.

[0071] See also Figure 4When the intelligent digital rail transit simulation test system of the present invention is used to test a model vehicle, the model vehicle is started, the electronic map of the operating line is loaded, and the initial heading angle value of the simulated vehicle is displayed. In the manual mode, the vehicle slows down and matches the track entry mark. When the model vehicle meets the track entry state, the driving simulator 22 command key is pressed once, and the model vehicle then enters the automatic driving mode. If it is necessary to exit the automatic driving mode, the brake can be stepped on with a foot. During the automatic driving of the model vehicle, if the steering wheel is misoperated during driving, the model vehicle first determines whether the changed heading angle value is within the preset threshold interval. If it is within the preset threshold interval, the model vehicle feeds back the heading angle information to the driving simulator 22, and the steering wheel automatically corrects the deviation; if it is not within the preset threshold interval, the vehicle automatically performs emergency braking and exits the automatic driving mode.

[0072] In summary, the intelligent digital rail transit simulation test system and method of the present invention can not only monitor the status of train equipment, but also control the status change of train equipment in real time. It introduces artificial intelligence machine learning technology to improve the effectiveness of the test. It can be connected to the real vehicle + simulation machine at the same time. At this time, the system is converted into an internal product pre-release test platform.

[0073] Those skilled in the art should recognize that the above embodiments are only used to illustrate the present invention, and are not intended to limit the present invention. As long as they are within the spirit of the present invention, any changes or modifications to the above embodiments will fall within the scope of the claims of the present invention.

Claims

1. An intelligent digital rail transit simulation test system based on digital twin technology, characterized in that: It includes digital track simulation platform, digital track internal test platform and central digital track server, among which: The digital track simulation platform is a visualized digital track 3D scene model designed and constructed in combination with the rubber-wheeled train vehicle dynamics model, and relying on the rubber-wheeled train vehicle dynamics model data and the digital track urban construction coordinate diagram; The digital track internal test platform includes a test navigation control unit, in which a digital track mathematical model, an electronic map, magnetic coding data and environmental perception simulation information are pre-installed; The central digital track server is a digital track system management server deployed in the center, and the central digital track server has a built-in real-time database and a digital track rule library; The on-board navigation control unit of the rubber-wheeled train obtains the status of the roadside equipment and driving data through the in-vehicle communication network, and transmits the corresponding data back to the central digital rail server through the 5G vehicle-ground communication network, and stores it in the real-time database of the central digital rail server; The digital track simulation platform monitors whether there is any abnormality in the data in the real-time database of the central digital track server, and decides whether to remotely operate the equipment or take over the automatic operation control of the vehicle through the platform dispatcher. At the same time, the digital track simulation platform calls the data in the real-time database of the central digital track server to drive the operation of the digital track 3D scene model, mapping the operating status of the vehicle and roadside equipment in the real environment; The communication between the digital track simulation platform and the test navigation control unit adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform sends the vehicle position, speed, posture and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model to the digital track mathematical model preset by the test navigation control unit in the form of network data packets, and then quickly retrieves and solves the magnetic nail information closest to the current vehicle according to the electronic map and magnetic coding data loaded in advance, and then adds environmental perception simulation information, and integrates the magnetic nail information solution, and injects the solution result into the vehicle's lateral and longitudinal control algorithm. When the vehicle's lateral and longitudinal control algorithm gives a control instruction, the lateral and longitudinal control instruction is encapsulated and sent to the vehicle dynamics model through a network data packet, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

2. An intelligent digital rail transit simulation test system as claimed in claim 1, characterized in that: After the digital track simulation platform obtains the digital track test data, the digital track test data is imported into the machine self-learning model for rubber-wheeled train control. Through repeated training and test iterations, the optimized parameters are substituted into the digital track internal test platform, thereby improving the test effectiveness and improving the algorithm design.

3. The intelligent digital rail transit simulation test system according to claim 1, characterized in that: The digital track internal test platform also includes a driving simulator, which uses CAN / serial port communication to establish communication between the driving simulator and the digital track simulation platform, and verifies the smoothness of the driving mode switching of the digital track mathematical model and the anti-disturbance ability of the simulation test system by simulating third-party control command input.

4. An intelligent digital rail transit simulation test method based on digital twin technology, characterized in that: The following steps are involved: S1, forming a 3D scene model based on the digital magnetic source high-precision map and building a digital track simulation platform, specifically including the following steps: S11, generate a digital track 2D scene model: high-precision mapping of magnetic nail coordinate points, draw a magnetic nail urban construction coordinate map through CAD, fix the RTK antenna to the front of the rubber-wheeled train, and manually drive to collect high-precision GPS train running tracks; after fitting the GPS and magnetic nail map tracks, draw a digital track 2D scene model; S12, digital-analog conversion to generate digital track simulation platform: Relying on the 2D scene model based on digital twin technology and the 3D mechanical drawings of roadside facilities and equipment, 3D visualization 1:1 modeling restoration is performed, and the full equipment status information is displayed through the echarts front end to obtain the digital track 3D scene model and build a digital track simulation platform framework; S2, the scenario model establishes a real-time communication mechanism with the rubber-wheeled train, which specifically includes the following steps: S21, digital rail vehicle-ground data communication: When the rubber-wheeled train is powered on, the on-board controller, sensor and processor transmit the driving data to the on-board navigation control unit in real time through the TRDP protocol. The roadside equipment is first collected by the roadside unit, processed once by the edge computing unit, and then the roadside data is transmitted to the on-board navigation control unit through the 5G vehicle-ground communication network. The on-board navigation control unit fuses and solves the driving data and roadside data, and then sends the obtained optimal control instructions to the vehicle actuator, so that the rubber-wheeled train takes corresponding actions; at the same time, the above data information is also forwarded to the central digital rail server through the 5G vehicle-ground communication network and stored in the real-time database of the central digital rail server; the digital rail simulation platform calls the data in the real-time database to drive the operation of the digital rail 3D scene model, so as to map the operating status of the vehicle and roadside equipment in the real environment; S22, digital track system fault warning: Based on the fault code feedback stored in the real-time database and provided by each device, the corresponding status code is assigned according to the fault level, and displayed in different colors in the front-end visualization area of ​​the central digital track server; the fault level is divided into operational safety fault and operational non-safety fault; S23, remote control of digital rail system trains: under the condition of satisfying the automatic driving operation, the operation status of the rubber-wheeled train is remotely controlled by controlling the computer-specific peripheral 5G, and the operation status of the rubber-wheeled train includes the start and stop, acceleration and deceleration, and the entry / switching / exit of the horizontal and vertical automatic control mode of the rubber-wheeled train; at this time, combined with the fault feedback status in step S22, when an operational safety fault is reported, the system is forced to stop operation; and when an operational non-safety fault is reported, the personnel remotely operate the train to slow down or stop according to the situation; S24, digital rail system operation history review: the digital rail simulation platform calls the video stream data in the real-time database of the central digital rail server, imports and drives the model, and synchronizes the timestamp to visualize and reproduce the historical scenes; S3, a real-time communication mechanism is established between the digital track simulation platform and the digital track internal test platform: the digital track internal test platform includes a test navigation control unit, and the communication between the digital track simulation platform and the test navigation control unit adopts UDP / TRDP-based data packets. During the simulation process, the digital track simulation platform sends the vehicle position, speed, attitude and angular velocity information calculated in real time for each frame of the rubber-wheeled train vehicle dynamics model in the form of network data packets to the digital track mathematical model preset by the test navigation control unit, and then quickly retrieves and solves the magnetic nail information closest to the current vehicle according to the electronic map and magnetic coding data loaded in advance, and then adds environmental perception simulation information, and integrates the magnetic nail information solution, and injects the solution result into the vehicle's lateral and longitudinal control algorithm. When the vehicle's lateral and longitudinal control algorithm gives a control instruction, the lateral and longitudinal control instruction is encapsulated and sent to the vehicle dynamics model through a network data packet, thereby driving the vehicle dynamics model to perform the next simulation calculation and operation.

5. The intelligent digital rail transit simulation test method based on digital twin technology as claimed in claim 4, characterized in that: In step S3, a machine self-learning model training link is added to the digital track simulation platform to filter abnormal data and enhance the prediction ability of the digital track mathematical model. First, the data is divided into a training set and a test set in proportion. Secondly, the model is trained using the training set data, and the parameters are adjusted to reduce the training error. The model prediction effect is evaluated on the test set. The above process is repeated several times and the average is taken to reduce the data division error. Finally, the features and models are continuously optimized according to the prediction error to improve the prediction accuracy, and the key parameters of the calibrated vehicle control model state machine are output, which in turn assists the algorithm optimization design.

6. The intelligent digital rail transit simulation test method based on digital twin technology according to claim 4, characterized in that: In step S3, a driving simulator is added to the internal test platform of the digital track, and the communication between the driving simulator and the digital track simulation platform is established by using the CAN / serial port communication method. By simulating the input of third-party control commands, the smoothness of the driving mode switching of the digital track mathematical model and the anti-disturbance ability of the simulation test system are verified.

Images