DT-based self-adaptive updating method for intelligent chassis model of unmanned mine

Abstract

The invention belongs to the technical field of coal mine transportation assistance, and aims to solve the problems of long time, high cost and high safety risk of a test method based on an actual environment. The DT-based self-adaptive updating method for the intelligent chassis model of the unmanned mine comprises the following steps: acquiring environmental data around a vehicle, and constructing a three-dimensional environmental map of an underground coal mine haulage roadway based on the environmental data; obtaining vehicle state data, and constructing a whole vehicle mechanism model based on the vehicle state data; building a digital twinborn model based on the three-dimensional environment map and the whole vehicle mechanism model, and carrying out joint simulation; and improving the operation control strategy of the vehicle physical entity based on the simulation result. According to the method, an unmanned mine intelligent chassis test simulation platform is constructed based on a digital twin technology, and the effectiveness of an unmanned algorithm and the reliability of drive-by-wire chassis structure design and high response are efficiently and accurately verified through the method.

Description

Technical Field

[0001] The present invention belongs to the technical field of coal mine transportation assistance, and in particular relates to a DT-based self-adaptive updating method for an unmanned mine intelligent chassis model. Background Art

[0002] At present, the mine transportation system is also gradually developing towards unmanned operation. As the main means of transporting materials and workers other than coal in the production process of underground coal mines, the mine auxiliary transportation system is a key link in the entire coal production system. Its technical level and operating efficiency are directly related to the achievement of the goal of reducing manpower and increasing efficiency in coal mine production. At present, traditional mine auxiliary transportation is mainly based on driver driving and manual dispatching, with low automation and backward information level, which is prone to safety accidents; at the same time, it is mainly based on mechanical and hydraulic transmission, with low transmission efficiency and poor reliability.

[0003] However, due to the poor underground environment and the fact that coal transportation is a high-risk industry, on the one hand, if the path planning of the unmanned driving system is unreasonable, it will cause serious coal accidents. Therefore, the safety and reliability of the unmanned driving system must be guaranteed; on the other hand, since the unmanned driving system is designed based on the characteristics of the wire-controlled chassis, the wire-controlled characteristics of the steering system, braking system and transmission system of the unmanned mine intelligent chassis must highly meet the linear requirements of the unmanned driving system. If the unmanned mine intelligent chassis is in an abnormal state, there will also be safety hazards.

[0004] The current unmanned driving system and wire-controlled chassis performance tests must be carried out in corresponding actual physical scenarios, after a lot of training and experiments, in order to obtain the appropriate structural design parameters and decision-making control algorithms. This method requires a lot of debugging time and cost, and there are also certain safety risks, which is not conducive to the rapid development of products. Summary of the invention

[0005] In order to solve at least one of the above technical problems existing in the prior art, the present invention provides a DT-based unmanned mine intelligent chassis model adaptive updating method.

[0006] The present invention is implemented by the following technical scheme: a DT-based unmanned mine intelligent chassis model adaptive updating method, comprising the following steps: obtaining environmental data around the vehicle, and constructing a three-dimensional environmental map of the underground transportation tunnel of the coal mine based on the environmental data; obtaining vehicle status data, and constructing a whole vehicle mechanism model based on the vehicle status data; the whole vehicle mechanism model includes a mechanical subsystem module, a decision-making subsystem module, and a control subsystem module; constructing a digital twin model based on the three-dimensional environmental map and the whole vehicle mechanism model, and performing joint simulation; improving the operation control strategy of the vehicle physical entity based on the simulation results.

[0007] Preferably, the environmental data includes fused perception data obtained by combining a lidar, a combined inertial navigation system and an RGB-D front camera, wherein the lidar is used to collect point cloud data, the combined inertial navigation system is used to measure the acceleration and angular velocity of the object using an accelerometer and a gyroscope, and obtain speed and position information through integration, and the RGB-D front camera is used to classify and identify traffic signs and signal light status to obtain traffic signal information.

[0008] Preferably, the mechanical subsystem module is a whole vehicle dynamics model built based on Simulink / Simscape, including the whole vehicle size and shape, driving system, power transmission system, braking system and steering system, wherein the whole vehicle size and shape include parameter settings for body information and aerodynamics, the driving system includes parameter configuration for tires and suspension, and the power transmission system includes parameter settings for transmission and differential; and the mass, inertia matrix and friction coefficient of the vehicle are set in Simulink / Simscape.

[0009] Preferably, based on a recursive least squares algorithm with a forgetting factor, iterative optimization is performed using vehicle status data collected in real time from sensors and encoders to achieve online adaptive updating and parameter synchronization of physical parameter values ​​of tire cornering stiffness values ​​of the whole vehicle mechanism model.

[0010] Preferably, the decision-making subsystem module and the control subsystem module are both built based on MATLAB / Simulink software, and the decision-making subsystem module includes normal driving state, StopToEnd state, RiskLevelA state, RiskLevelB state, QuitStateOne state, QuitStateTwo state and parking state; the control subsystem module is built based on the decision-making subsystem module, and the corresponding control strategy is constructed based on different risk levels and vehicle states; the Simulink / Simscape mechanism model and the MATLAB / Simulink decision-making control module are jointly simulated through the built-in energy conversion and transmission module.

[0011] Preferably, a global map of underground transport tunnels in coal mines is constructed based on point cloud data and the Gazebo physical simulation platform. At the same time, in the Gazebo physical simulation environment, a three-dimensional environmental map is generated by adding vehicles, workers, lighting, dust, water mist and obstacles, and the map is continuously updated by the point cloud data obtained by real-time scanning of the new environment by lidar; based on MATLAB / Simulink, Simulink / Simscape, and Gazebo, a final joint simulation platform is constructed, and the Ray-Col collision detection method and the Muti-PPO algorithm are integrated to perform predictive simulation of the entire vehicle, and the operation control strategy of the vehicle physical entity is updated based on the simulation results.

[0012] Preferably, the input variables of the digital twin model are brake pressure and steering angle, and the output parameters include vehicle speed, body roll angle, pitch angle and yaw angular velocity.

[0013] Preferably, the digital twin model includes a physical data module, a digital twin space, a physical space, a virtual-reality interactive mapping space and a twin data module; wherein the physical space includes the physical entity of the unmanned mine intelligent chassis vehicle, the laser radar, the RGB-D front camera, the combined inertial navigation and its actual working environment arranged on the vehicle, and the core component is the wheeled wire-controlled intelligent chassis. The physical data module includes environmental data and vehicle status data. The digital twin space is the core part of the digital twin model, including a three-dimensional environmental map and a vehicle mechanism model. The virtual-reality interactive mapping space monitors the vehicle movement process in real time through one-dimensional data, two-dimensional charts and three-dimensional models. The twin data module includes twin data obtained based on simulation.

[0014] Preferably, virtual debugging is performed based on the digital twin model, and decision feedback of the unmanned mine intelligent chassis is realized based on the obtained twin data, forming a closed-loop mapping mechanism, and its five-dimensional expression is shown in the formula:

[0015] DV DT =(PDV, VDV, DVDTD, DVSS, CN)

[0016] Among them, DV DT It represents the digital twin framework of unmanned mine intelligent chassis, PDV represents the physical space of unmanned mine intelligent chassis, VDV represents the digital twin space of unmanned mine intelligent chassis, DVDTD represents the twin data obtained by simulating the digital twin model, DVSS represents the virtual-reality interactive mapping space, and CN represents connection.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention proposes to use digital twin technology to construct a test simulation platform for unmanned mine intelligent chassis, and to construct a multi-domain, multi-variable, multi-parameter, multi-coupled, and nonlinear complex mechatronics model of the unmanned mine intelligent chassis. This solves the problem that most current modeling still remains in the single model construction stage, making it difficult to achieve coupled modeling of multi-domain models and the problem of low model simulation accuracy.

[0019] The present invention integrates the multi-domain coupled unmanned mine intelligent chassis digital twin model and the laser radar SLAM 3D reconstruction map to form a virtual test environment, and further integrates autonomous collision detection with local obstacle avoidance and global path planning algorithms on this basis to verify the effectiveness of the unmanned driving algorithm and the reliability of the wire-controlled chassis structure design and high response, solving the problems of long time, high cost and high safety risks of current actual environment testing methods, which is conducive to the rapid iteration and development of products in the later stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 It is a flow chart of constructing the digital twin system of the unmanned mine intelligent chassis of the present invention;

[0022] Figure 2 It is the digital twin test system framework of the unmanned mine intelligent chassis of the present invention;

[0023] Figure 3 It is a composition diagram of the whole vehicle mechanism model of the present invention. DETAILED DESCRIPTION

[0024] In conjunction with the drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so that people familiar with this technology can understand and read them. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention. It should be noted that in this specification, relational terms such as first and second are only used to distinguish one entity from several other entities, and do not necessarily require or imply any actual relationship or order between these entities.

[0026] The present invention provides an embodiment:

[0027] like Figures 1 to 3 As shown, a DT-based unmanned mine intelligent chassis model adaptive updating method includes the following steps: acquiring environmental data around the vehicle, and constructing a three-dimensional environmental map of the underground transportation tunnel of the coal mine based on the environmental data; acquiring vehicle status data, and constructing a whole vehicle mechanism model based on the vehicle status data; the whole vehicle mechanism model includes a mechanical subsystem module, a decision subsystem module, and a control subsystem module; constructing a digital twin model based on the three-dimensional environmental map and the whole vehicle mechanism model, and performing joint simulation; improving the operation control strategy of the vehicle physical entity based on the simulation results.

[0028] In this embodiment, the test framework of the digital twin model includes a physical data module, a digital twin space, a physical space, a virtual-reality interactive mapping space, and a twin data module;

[0029] The physical space includes the physical entity of the unmanned mine intelligent chassis vehicle, the laser radar, RGB-D front camera, combined inertial navigation and its actual working environment arranged in the vehicle. The core component is the wheeled wire-controlled intelligent chassis, which is both a driving device and a load-bearing structure, and also integrates various unit modules with autonomous driving functions.

[0030] The physical data module includes environmental data and vehicle status data. The environmental data includes the fused perception data obtained by the combination of lidar, combined inertial navigation and RGB-D front camera. The lidar uses a 128-line long-distance lidar with a maximum detection distance of 200m, a wavelength of 865nm, a vertical field of view of 45°, a horizontal field of view of 360°, and a scanning frequency of 20HZ. It is mainly responsible for collecting a large amount of point cloud data; the combined inertial navigation includes GNSS and IMU, which use accelerometers and gyroscopes to measure the acceleration and angular velocity of objects, and obtain speed and position information through integration to achieve accurate navigation and positioning; the RGB-D front camera is mainly responsible for the classification and identification of traffic signs and traffic light status to obtain traffic signal information.

[0031] The environmental data obtained by the fusion perception of LiDAR, combined inertial navigation and RGB-D front camera can be divided into two categories according to their specific uses: the first category is used for mapping and positioning testing. Combining SLAM (simultaneous localization and mapping) mapping algorithm and UWB (ultra-wideband positioning) technology, a multi-modal data-driven fusion positioning strategy is formed to build an environmental map in real time and determine its own absolute coordinates, which can improve the accuracy and robustness of the positioning of the unmanned mine intelligent chassis in the underground environment; the second category is used for perception testing, which is responsible for perceiving the traffic environment near the vehicle, such as pedestrians and cars.

[0032] The vehicle status data is collected through various sensors, such as encoders, speed sensors, temperature sensors, pressure sensors, etc., which comprehensively collect data of the unmanned mine intelligent chassis during operation, including driving signals such as steering angle and brake pressure. After data processing and data cleaning, the above-mentioned historical data and real-time data are transmitted and stored in a multidimensional database using the high-speed industrial Ethernet LAN bus and system safety CAN bus, which facilitates the subsequent mapping of data to the digital twin model.

[0033] The digital twin space is the core part of the digital twin model, including a three-dimensional environmental map and a vehicle mechanism model. The mechanical subsystem module is a visualized three-dimensional dynamic model constructed based on the geometric structure of the unmanned mine intelligent chassis and the constraints of each component. The mechanical subsystem module is a vehicle dynamics model built based on Simulink / Simscape, including the vehicle size and shape, driving system, power transmission system, braking system and steering system. The vehicle size and shape include parameter settings for body information and aerodynamics, the driving system includes parameter configurations for tires and suspension, and the power transmission system includes parameter settings for the engine, clutch, transmission and differential. The vehicle's mass, inertia matrix and friction coefficient are set in Simulink / Simscape to further improve the accuracy of the model.

[0034] At the same time, considering that the performance status of the mechanical parts of the vehicle will change due to wear and tear during the long-term actual operation of the unmanned mine intelligent chassis, the physical parameters of the system will also change, and the unchanging model cannot accurately reflect its actual operating conditions, and the error between the simulation accuracy and the actual operating results will gradually increase. Therefore, based on the recursive least squares algorithm with a forgetting factor, the vehicle status data collected in real time from sensors and encoders is used for iterative optimization to achieve online adaptive update and parameter synchronization of the physical parameter values ​​of the tire cornering stiffness in the vehicle mechanism model. Through the adaptive update of the unmanned mine intelligent chassis, the model can be better adapted to different operating and environmental conditions, and the simulation accuracy of the model can be improved.

[0035] The decision-making subsystem module and the control subsystem module are both built based on MATLAB / Simulink software. The decision-making subsystem module formulates prevention and control decisions and path planning for various risks during driving based on the finite state machine. The decision-making subsystem module includes normal driving state, StopToEnd state, RiskLevelA state, RiskLevelB state, QuitStateOne state, QuitStateTwo state and parking state; the control subsystem module is built based on the decision-making subsystem module, and the corresponding control strategy is built based on different risk levels and vehicle states, so as to achieve smooth and safe braking and parking of the unmanned mine intelligent chassis when facing collision risks during driving. The Simulink / Simscape mechanism model and the MATLAB / Simulink control decision module are jointly simulated through the built-in energy conversion and transmission module.

[0036] Based on point cloud data and the Gazebo physical simulation platform, a global map of underground transportation tunnels in coal mines is constructed. At the same time, in the Gazebo physical simulation environment, a three-dimensional environmental map is generated by adding vehicles, workers, lighting, dust, water mist and obstacles to meet the actual underground scenarios. This further verifies the stability, sensitivity and reliability of the unmanned driving system in obstacle avoidance and emergency handling. This method can not only establish a high-precision prior map of underground tunnels in coal mines, but also use the point cloud data obtained by real-time scanning of new environments by lidar to continuously update the map, ensuring that the digital twin model has adaptive update capabilities.

[0037] Based on MATLAB / Simulink, Simulink / Simscape, and Gazebo, the final joint simulation platform is built, and the Ray-Col collision detection method and Multi-PPO algorithm are integrated. Through simulation, predictive estimation is made for various motion states under the unmanned driving system and twin data is obtained. The control strategy of the physical entity and the unmanned driving algorithm are further improved based on the twin data.

[0038] The virtual-real interactive mapping space monitors the vehicle movement process in real time through one-dimensional data, two-dimensional charts and three-dimensional models, which improves the monitoring immersion and accuracy during the operation of the unmanned mine intelligent chassis. The state monitoring function can quickly locate problems in the operation of the unmanned mine intelligent chassis, and through the virtual debugging function of the digital twin model, predict possible problems in the future, so as to adjust the model parameters in advance to keep it in the best working state. This not only reduces the time for troubleshooting and shortens the solution process, but also improves the safety of the unmanned mine intelligent chassis during driving, while also reducing maintenance costs.

[0039] Based on the digital twin model, virtual debugging is carried out, and the decision feedback of the unmanned mine intelligent chassis is realized according to the obtained twin data, forming a closed-loop mapping mechanism. Its five-dimensional expression is shown in the formula:

[0040] DV DT =(PDV, VDV, DVDTD, DVSS, CN)

[0041] Among them, DV DT Represents the digital twin framework of the unmanned mine intelligent chassis, PDV represents the physical space of the unmanned mine intelligent chassis, VDV represents the digital twin space of the unmanned mine intelligent chassis, DVDTD represents the twin data obtained by the simulation of the digital twin model, DVSS represents the virtual-real interactive mapping space, and CN represents the connection. The input variables of the digital twin model are brake pressure and steering angle, and the output parameters include vehicle speed, body roll angle, pitch angle, and yaw angular velocity.

[0042] Implementation plan:

[0043] The multi-source perception information collected by various sensors such as LiDAR, RGB-D front camera and vehicle-mounted terminal is transmitted to the digital twin space in real time through communication modules, including Ethernet LAN bus and system safety CAN bus. In this space, the mechanism model of the unmanned mine intelligent chassis and the high-precision 3D reconstruction map are constructed, and the subsequent model is updated in real time.

[0044] In the virtual digital twin space, Simulink / Simscape is used to model the vehicle dynamics model of the unmanned mine intelligent chassis from five aspects: vehicle appearance, transmission system, braking system, steering system and driving system, providing an accurate reference for subsequent simulation and control algorithms.

[0045] In order to realize the global path planning of the unmanned mine intelligent chassis, the PPO algorithm is improved in MATLAB / Simulink under the Critic-Actor learning framework based on the Markov decision process theory. The reward and punishment mechanism is used to enable it to have decision-making ability, and the compilation of the autonomous global path planning algorithm of the digital twin model of the unmanned mine intelligent chassis is realized, so that the vehicle has decision-making ability.

[0046] A digital twin model of an unmanned mine intelligent chassis with multi-domain coupling of electromechanical control is constructed in MATLAB / Simulink. The model takes brake pressure and steering angle as input variables, and outputs key parameters including vehicle speed, body roll angle, pitch angle, and yaw rate. The input variable is the real-time drive signal collected by the sensor, which can realize real-time simulation of the digital twin model.

[0047] Furthermore, under the Gazebo physical simulation platform, combined with high-precision digital three-dimensional maps, the driving conditions of vehicles in a real underground coal mine environment were simulated, and obstacles such as vehicles and people were added. On the one hand, the feasibility and stability of the Ray-Col collision detection method and the Muti-PPO algorithm were verified through simulation. On the other hand, the reliability of each subsystem during operation, such as driving, steering and braking, and the response accuracy of the control strategy were verified.

[0048] By using the predictive simulation results of the digital twin model to develop and improve the unmanned mine intelligent chassis, the actual vehicle can reach the optimal state during actual operation, greatly improving the reliability of the vehicle during actual operation. At the same time, it also solves the problem that the actual vehicle testing method in the open underground coal mine environment is time-consuming, costly, and has high safety risks.

[0049] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A DT-based adaptive updating method for unmanned mine intelligent chassis model, characterized in that: The following steps are involved: Obtain environmental data around the vehicle and build a three-dimensional environmental map of the underground transportation tunnel in the coal mine based on the environmental data; Acquire vehicle status data, and build a whole vehicle mechanism model based on the vehicle status data; the whole vehicle mechanism model includes a mechanical subsystem module, a decision subsystem module, and a control subsystem module; Build a digital twin model based on the 3D environment map and vehicle mechanism model, and perform joint simulation; Improve the operation control strategy of the vehicle physical entity based on the simulation results.

2. According to claim 1, a DT-based unmanned mine intelligent chassis model adaptive update method is characterized by: Environmental data includes fused perception data obtained by combining lidar, combined inertial navigation and RGB-D front camera. Lidar is used to collect point cloud data, combined inertial navigation is used to measure the acceleration and angular velocity of objects using accelerometers and gyroscopes, and obtain speed and position information through integration, and the RGB-D front camera is used to classify and identify traffic signs and traffic light status to obtain traffic signal information.

3. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 1 is characterized in that: The mechanical subsystem module is a vehicle dynamics model built based on Simulink / Simscape, including vehicle size and shape, driving system, powertrain system, braking system and steering system. The vehicle size and shape include parameter settings for body information and aerodynamics, the driving system includes parameter configurations for tires and suspension, and the powertrain system includes parameter settings for transmission and differential. The vehicle's mass, inertia matrix and friction coefficient are set in Simulink / Simscape.

4. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 3 is characterized in that: Based on the recursive least squares algorithm with a forgetting factor, iterative optimization is performed using the vehicle status data collected in real time from sensors and encoders to achieve online adaptive update and parameter synchronization of the tire cornering stiffness value in the vehicle mechanism model.

5. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 3 is characterized in that: The decision-making subsystem module and the control subsystem module are both built based on MATLAB / Simul ink software. The decision-making subsystem module includes normal driving state, StopToEnd state, RiskLevelA state, RiskLevelB state, QuitStateOne state, QuitStateTwo state and parking state; The control subsystem module is built based on the decision-making subsystem module, and the corresponding control strategy is constructed based on different risk levels and vehicle status.

6. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 5 is characterized in that: Based on point cloud data and the Gazebo physical simulation platform, a global map of underground transportation tunnels in coal mines is constructed. In the Gazebo physical simulation environment, a 3D environment map is generated by adding vehicles, workers, lighting, dust, water mist, and obstacles. The map is continuously updated by using point cloud data obtained by real-time scanning of new environments by lidar. Based on MATLAB / Simulink, Simulink / Simscape, and Gazebo, the final joint simulation platform is built, and the Ray-Col collision detection method and Muti-PPO algorithm are integrated to perform predictive simulation of the entire vehicle, and the operation control strategy of the vehicle physical entity is updated based on the simulation results.

7. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 5 is characterized in that: The input variables of the digital twin model are brake pressure and steering angle, and the output parameters include vehicle speed, body roll angle, pitch angle and yaw rate.

8. The method for adaptively updating the model of an unmanned mine intelligent chassis based on DT according to claim 1 is characterized in that: The digital twin model includes physical data module, digital twin space, physical space, virtual-reality interactive mapping space and twin data module; The physical space includes the physical entity of the unmanned mine intelligent chassis, the laser radar, RGB-D front camera, combined inertial navigation and its actual working environment arranged on the whole vehicle. The core component is the wheeled wire-controlled intelligent chassis. The physical data module includes environmental data and vehicle status data. The digital twin space is the core part of the digital twin model, including a three-dimensional environmental map and a whole vehicle mechanism model. The virtual-reality interactive mapping space monitors the vehicle movement process in real time through one-dimensional data, two-dimensional charts and three-dimensional models. The twin data module includes twin data obtained based on simulation.

9. The method for adaptively updating the DT-based unmanned mine intelligent chassis model according to claim 8 is characterized in that: Based on the digital twin model, virtual debugging is carried out, and the decision feedback of the unmanned mine intelligent chassis is realized according to the obtained twin data, forming a closed-loop mapping mechanism. Its five-dimensional expression is shown in the formula: DV DT =(PDV,VDV,DVDTD,DVSS,CN) Among them, DV DT It represents the digital twin framework of unmanned mine intelligent chassis, PDV represents the physical space of unmanned mine intelligent chassis, VDV represents the digital twin space of unmanned mine intelligent chassis, DVDTD represents the twin data obtained by simulating the digital twin model, DVSS represents the virtual-reality interactive mapping space, and CN represents connection.

Images